# Fine-tuning

Fine-tuning is a machine-learning technique that further trains a pre-trained model for a specific task or use case.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## FlowMap-OPD: On-Policy Distillation for Few-Step Flow-Map Generators

DevFeed: [FlowMap-OPD: On-Policy Distillation for Few-Step Flow-Map Generators](<https://devfeed.tech/articles/flowmap-opd-on-policy-distillation-for-few-step-flow-map-generators-144946.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.37851>)

Author: Zhiqi Li, Cheng Wan, Bo Zhu

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Post-training optimization](<https://devfeed.tech/topics/post-training-optimization.md>), [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [flux](<https://devfeed.tech/topics/flux.md>)

Tags: [consistency](<https://devfeed.tech/tags/consistency.md>), [consistency-models](<https://devfeed.tech/tags/consistency-models.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [distillation](<https://devfeed.tech/tags/distillation.md>), [flow](<https://devfeed.tech/tags/flow.md>), [flux](<https://devfeed.tech/tags/flux.md>), [generators](<https://devfeed.tech/tags/generators.md>), [policy](<https://devfeed.tech/tags/policy.md>), [sd3-5](<https://devfeed.tech/tags/sd3-5.md>), [supervision](<https://devfeed.tech/tags/supervision.md>), [velocity](<https://devfeed.tech/tags/velocity.md>)

### AI overview

FlowMap-OPD trains few-step flow-map students using states collected by the student's native sampler, while separating that collection from teacher supervision. The study finds that supervising instantaneous velocity with a small consistency penalty transfers teacher gains best; directly supervising the deployed flow map performs worst when teacher and student use different pretrained initializations. The authors summarize the approach as "sample with the flow map, learn the velocity." It also supports teachers that define a velocity, including multi-step teachers, and students initialized from multi-step models. On FLUX and SD3.5, one few-step student consolidated three specialist teachers within 300 updates, matched or exceeded each on its own benchmark, and outperformed multi-reward reinforcement learning and LoRA merging. The supplied text is an abstract, so it does not provide full methodological or evaluation details.

### Source excerpt

arXiv:2609.37851v2 Announce Type: replace-cross Abstract: Few-step flow-map generators, such as MeanFlow and consistency models, sample in a few long-range jumps. On-policy distillation (OPD) trains a student on states drawn from its own sampler, but existing methods compare teacher and student through the local stochastic transitions of a multi-step sampler, which a flow-map student never deploys. We present FlowMap-OPD, the first systematic study of OPD for flow-map students. It separates state collection from supervision: the student collects states with its native few-step sampler, while the teacher supervises the flow map, velocities induced from it, or the instantaneous velocity linked to the map by a separately weighted consistency penalty. We find that instantaneous-velocity supervision with a small consistency weight, our chosen form, transfers the teacher's gains best, whereas supervising the deployed map, although the natural choice, transfers them worst when teacher and student start from different pretrained models. This gives a counter-intuitive rule: \emph{sample with the flow map, learn the velocity}. As a further benefit, this form applies to any teacher that defines a velocity, including multi-step ones, and to students initialized from multi-step models. On FLUX and SD3.5, a single few-step student consolidates three specialist teachers within 300 updates, matches or exceeds each on its own benchmark (e.g., OCR 0.883 versus 0.850 on FLUX), and outperforms both multi-reward reinforcement learning and LoRA merging.

## Can LLMs Resolve Entity Conflicts in Domain Specific NLU Pipelines?

DevFeed: [Can LLMs Resolve Entity Conflicts in Domain Specific NLU Pipelines?](<https://devfeed.tech/articles/can-llms-resolve-entity-conflicts-in-domain-specific-nlu-pipelines-145095.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25733>)

Author: Akansha Bhardwaj, Mingbo Cui, Victor Guimaraes, Dominik Kohler, Diana Nicoleta Popa, Meghana M. Reddy, Paolo Rosso

Published: 2026-11-20T05:00:00Z

Content type: comparison

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Entity resolution](<https://devfeed.tech/topics/entity-resolution.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [lora](<https://devfeed.tech/topics/lora.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [cases](<https://devfeed.tech/tags/cases.md>), [comparison](<https://devfeed.tech/tags/comparison.md>), [complex](<https://devfeed.tech/tags/complex.md>), [conflict-resolution](<https://devfeed.tech/tags/conflict-resolution.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [flexibility](<https://devfeed.tech/tags/flexibility.md>), [llama3](<https://devfeed.tech/tags/llama3.md>), [llms](<https://devfeed.tech/tags/llms.md>), [lora](<https://devfeed.tech/tags/lora.md>), [pipelines](<https://devfeed.tech/tags/pipelines.md>), [prompting](<https://devfeed.tech/tags/prompting.md>), [strategies](<https://devfeed.tech/tags/strategies.md>)

### AI overview

The study compares rule-based entity conflict resolution with two Llama3.1-8B-Instruct approaches in domain-specific NLU pipelines: multi-shot prompting and LoRA fine-tuning. It evaluates how well the models match existing pipeline outputs and, separately, their accuracy on human-labeled data. The abstract reports that prompting is easy to adapt and allows rapid behavioral changes, while fine-tuning performs better on complex, ambiguous cases. It suggests combining the approaches to pair prompting's flexibility with fine-tuning's broader accuracy gains.

### Source excerpt

arXiv:2611.25733v1 Announce Type: cross Abstract: Resolving entity conflicts is a critical step in Pipeline-based NLU architectures, especially when multiple extractors provide overlapping or conflicting outputs. Traditional rule-based entity conflict resolution methods often lack flexibility and are unscalable. In this work, we aim to study if the broad knowledge of LLMs can be leveraged to resolve entity conflicts in domain specific NLU pipelines. We present an experimental comparison between a traditional rule-based conflict resolution method, and two LLM-based approaches: a multi-shot prompting method using the Llama3.1-8B-Instruct model and a LoRA fine-tuned Llama3.1-8B-Instruct model. The evaluation was conducted under two setups: (1) a broad assessment of LLM's capability to match existing NLU pipeline output, and (2) a more controlled scenario assessing the LLM accuracy on human-labeled data. Results show that while multi-shot prompting offers ease of adaptation and rapid behavioral changes, the fine-tuned model delivers superior performance on complex, ambiguous cases. Ultimately, the best approach appears to be a combination of both strategies, leveraging multi-shot prompting's flexibility alongside the broad accuracy gains of LoRA fine-tuning for a more comprehensive performance.

## The Post-Training Dilemma: Why We Should Rethink the Sequential SFT-RL Paradigm

DevFeed: [The Post-Training Dilemma: Why We Should Rethink the Sequential SFT-RL Paradigm](<https://devfeed.tech/articles/the-post-training-dilemma-why-we-should-rethink-the-sequential-sft-rl-paradigm-144871.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2601.07389>)

Author: Xueyan Niu, Bo Bai, Wei Han, Weixi Zhang

Published: 2026-11-20T05:00:00Z

Content type: opinion

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [post-training](<https://devfeed.tech/topics/post-training.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [capabilities](<https://devfeed.tech/tags/capabilities.md>), [condition](<https://devfeed.tech/tags/condition.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-it](<https://devfeed.tech/tags/cs-it.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [hh](<https://devfeed.tech/tags/hh.md>), [language](<https://devfeed.tech/tags/language.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [large-language-model](<https://devfeed.tech/tags/large-language-model.md>), [learning](<https://devfeed.tech/tags/learning.md>), [llm](<https://devfeed.tech/tags/llm.md>), [math-500](<https://devfeed.tech/tags/math-500.md>), [math-it](<https://devfeed.tech/tags/math-it.md>), [mechanism](<https://devfeed.tech/tags/mechanism.md>), [misalignment](<https://devfeed.tech/tags/misalignment.md>), [post-training](<https://devfeed.tech/tags/post-training.md>), [rl](<https://devfeed.tech/tags/rl.md>), [sft](<https://devfeed.tech/tags/sft.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

The paper argues that sequential SFT-then-RL training can raise the SFT loss because the two objectives have nearly orthogonal gradients, a conflict the authors say hyperparameter tuning or a KL penalty cannot resolve. Under the Polyak-Lojasiewicz condition, it gives a lower bound on this increase based on distance from the SFT-optimal parameter set. Experiments with Qwen3-0.6B report higher SFT loss after RL on MATH-500, Alpaca, HH-RLHF, and SST-2; the authors also describe SFT as flattening and RL as re-sharpening the loss landscape. They propose joint optimization with gradient surgery, landscape-aware RL stopping, and studying the model-scale threshold where SFT may be unnecessary.

### Source excerpt

arXiv:2601.07389v3 Announce Type: replace Abstract: This position paper argues that the sequential supervised fine-tuning followed by reinforcement learning (SFT-then-RL) pipeline---the de facto paradigm for large language model (LLM) post-training---has a fundamental property that cannot be fixed by hyperparameter tuning: SFT and RL optimize over nearly orthogonal parameter subspaces, so RL inevitably degrades the capabilities acquired during the preceding SFT phase. Under the Polyak-Lojasiewicz (PL) condition, which is empirically observed for over-parameterized networks, we show that RL increases the SFT loss by at least $\frac{\mu}{2}\mathrm{dist}(\theta_{\mathrm{RL}}, \Theta^*_{\mathrm{SFT}})^2$ for some $\mu>0$. Although the KL penalty can constrain distributional shift, it cannot correct the underlying gradient misalignment between the two objectives. Experiments on Qwen3-0.6B across four tasks confirm the theoretical predictions: on MATH-500, Alpaca, HH-RLHF, and SST-2, the SFT loss increases after the RL phase. We identify the underlying mechanism as persistent near-orthogonality between the gradients of the two objectives, paired with opposing landscape dynamics: SFT flattens the loss landscape while RL re-sharpens it. Based on our analysis, we propose three promising research directions: (1) joint optimization of the two objectives with gradient surgery to resolve conflicting updates; (2) landscape-aware scheduling that terminates RL training at the theoretically optimal point; and (3) characterizing the model scale threshold above which SFT becomes unnecessary.

## VoxForge: A Unified Foundation Model for Dense 3D Features Across MRI, CT and PET

DevFeed: [VoxForge: A Unified Foundation Model for Dense 3D Features Across MRI, CT and PET](<https://devfeed.tech/articles/voxforge-a-unified-foundation-model-for-dense-3d-features-across-mri-ct-and-pet-145039.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25309>)

Author: Rushil Anirudh, Jayaraman J. Thiagarajan

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [article](<https://devfeed.tech/tags/article.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [distillation](<https://devfeed.tech/tags/distillation.md>), [encoder](<https://devfeed.tech/tags/encoder.md>), [features](<https://devfeed.tech/tags/features.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [foundation-model](<https://devfeed.tech/tags/foundation-model.md>), [generalist](<https://devfeed.tech/tags/generalist.md>), [global](<https://devfeed.tech/tags/global.md>), [matching](<https://devfeed.tech/tags/matching.md>), [medical](<https://devfeed.tech/tags/medical.md>), [models](<https://devfeed.tech/tags/models.md>), [pet](<https://devfeed.tech/tags/pet.md>), [tasks](<https://devfeed.tech/tags/tasks.md>), [test](<https://devfeed.tech/tags/test.md>), [top](<https://devfeed.tech/tags/top.md>), [volumes](<https://devfeed.tech/tags/volumes.md>), [wide](<https://devfeed.tech/tags/wide.md>)

### AI overview

VoxForge is a single 3D encoder trained on permissively licensed MRI and CT volumes using 3D architecture advances, self-distillation, and cross-modal curricula. Evaluated with its features frozen, it achieves the best or tied-best average rank in every task group across segmentation, classification, and registration, including on PET, a modality absent from training. The abstract reports performance on localized and global tasks at the level of state-of-the-art specialists and says it matches top-performing models on TotalSegmentator CT using one tenth as many labels. The supplied text is an abstract, so it does not provide implementation details or per-task results.

### Source excerpt

arXiv:2611.25309v1 Announce Type: cross Abstract: A unified 3D medical foundation model promises reusable, frozen features that serve a wide variety of tasks such as segmentation, classification and registration on MRI, CT and PET as well as specialist models do. Generalist models still need separate backbones for localized and global tasks, specialist models are confined to a single modality, with the concern that spanning modalities under-emphasizes fine anatomy, and both typically rely on end-to-end fine-tuning to be competitive. We present VoxForge, a single encoder that integrates recent advances in 3D architecture, self-distillation and cross-modal curricula to learn transferable features from permissively licensed MRI and CT volumes. We evaluate it strictly frozen, the test a reusable foundation model must pass, against state-of-the-art generalist and specialist models on segmentation, classification and registration across MRI and CT, and on PET, which it never sees. VoxForge has the best or tied-best average rank in every task group, serves localized and global tasks across modalities at the level of state-of-the-art specialists, and is extremely label-efficient, matching top-performing models on TotalSegmentator CT with a tenth of the labels.

## GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed

DevFeed: [GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed](<https://devfeed.tech/articles/groundanything-reconciling-parallel-decoding-with-precise-visual-grounding-at-flash-speed-144950.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.39600>)

Author: Qize Yu, Lianrui Fan, Bowen Ping, Xini Ding, Zetian Song, Junbo Niu, Kaixuan Wang, Tianxing Chen, Yue Chen, Minghua He, Yuran Wang, Jie Huang, Haojun Zhang, Min Chen, Hao Li, Wenxuan Song, Ruihai Wu,

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Decoding](<https://devfeed.tech/topics/decoding.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [grpo](<https://devfeed.tech/topics/grpo.md>), [post-training](<https://devfeed.tech/topics/post-training.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [also](<https://devfeed.tech/tags/also.md>), [ar](<https://devfeed.tech/tags/ar.md>), [art](<https://devfeed.tech/tags/art.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [astra](<https://devfeed.tech/tags/astra.md>), [autoregressive](<https://devfeed.tech/tags/autoregressive.md>), [based](<https://devfeed.tech/tags/based.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [coco](<https://devfeed.tech/tags/coco.md>), [combines](<https://devfeed.tech/tags/combines.md>), [conversion](<https://devfeed.tech/tags/conversion.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [data](<https://devfeed.tech/tags/data.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [decoding](<https://devfeed.tech/tags/decoding.md>), [flash](<https://devfeed.tech/tags/flash.md>), [grounding](<https://devfeed.tech/tags/grounding.md>), [parallel](<https://devfeed.tech/tags/parallel.md>), [speed](<https://devfeed.tech/tags/speed.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

GroundAnything is a 4B-parameter visual grounding model that uses blockwise diffusion denoising to generate and refine spatial hypotheses in parallel, addressing the sequential latency of autoregressive decoding. Its training combines public-data and dedicated-data-engine pretraining, conversion from autoregressive models with joint objectives, supervised fine-tuning, and GRPO reinforcement post-training. Across 30 benchmarks, its autoregressive variant scores 72.42%, compared with 71.35% for GPT-6 Astra among similarly sized models. With entropy-guided decoding, GroundAnything averages 61.75%, versus 53.32% for the fast MTP-based LocateAnything. An optional self-speculative mode reports a 4.51x speedup over the autoregressive counterpart, with a 0.74-point drop in COCO F1mIoU. The excerpt also says inference optimizations yield practical speedups, without detailing them.

### Source excerpt

arXiv:2609.39600v3 Announce Type: replace-cross Abstract: Autoregressive (AR) grounding models serialize spatial predictions, introducing sequential latency and imposing a causal order on output tokens. We view grounding as visual evidence extraction: objects, locations, and spatial relations are jointly constrained by the image and query, yet their dependencies do not imply an intrinsic left-to-right generation order. This distinction makes bidirectional diffusion a natural fit, allowing spatial hypotheses to emerge in parallel and be jointly refined through iterative denoising. We introduce GroundAnything, a 4B-parameter grounding foundation model that reconciles fast parallel decoding with precise localization through blockwise denoising. Training combines grounding pretraining from public datasets and dedicated data engines, direct AR-to-diffusion conversion with joint AR and diffusion objectives, supervised fine-tuning, and GRPO-based reinforcement post-training. Across 30 grounding benchmarks, our autoregressive variant, GroundAnything-VLM, establishes a new overall state of the art among similarly sized models at 72.42%, remaining competitive with GPT-6 Astra (71.35%). With entropy-guided decoding, GroundAnything also surpasses the prior state of the art at this scale, averaging 61.75% versus 53.32% for the fast MTP-based LocateAnything model. We further explore decoding strategies, showing that an optional self-speculative mode achieves a 4.51$\times$ speedup over the AR counterpart with a 0.74 percentage-point drop in COCO F1mIoU. Infrastructure experiments show that progressive inference optimizations translate parallel decoding into practical speedups. These gains support efficient visual grounding in latency-sensitive real-world systems.

## Interference Bounds and Merge-Aware Training Dynamics for Composable Fine-Tuned Models

DevFeed: [Interference Bounds and Merge-Aware Training Dynamics for Composable Fine-Tuned Models](<https://devfeed.tech/articles/interference-bounds-and-merge-aware-training-dynamics-for-composable-fine-tuned-models-145056.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25430>)

Author: Alireza Moayedikia

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [LLM Techniques](<https://devfeed.tech/topics/llm-techniques.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [alignment](<https://devfeed.tech/tags/alignment.md>), [composability](<https://devfeed.tech/tags/composability.md>), [control](<https://devfeed.tech/tags/control.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fine](<https://devfeed.tech/tags/fine.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [functional](<https://devfeed.tech/tags/functional.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [math-oc](<https://devfeed.tech/tags/math-oc.md>), [merge](<https://devfeed.tech/tags/merge.md>), [metric](<https://devfeed.tech/tags/metric.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [rank](<https://devfeed.tech/tags/rank.md>), [study](<https://devfeed.tech/tags/study.md>), [training](<https://devfeed.tech/tags/training.md>), [transformer](<https://devfeed.tech/tags/transformer.md>)

### AI overview

The paper bounds the excess loss from merging fine-tuned task updates using an interference functional that captures cross-task curvature overlap and signed update alignment. In the quadratic regime, this functional equals excess loss and measures alignment in the curvature metric. A curvature-guided regularizer bounds displacement in the base curvature metric during training; under a Loewner comparison with task curvatures, it also bounds the curvature component of interference inversely with penalty strength. It does not similarly reduce the first-order term, which arises because regularized solutions are not stationary for the task loss. Experiments confirm the bounds in an exact non-quadratic model, where the first-order term dominates under strong penalties and the curvature-guided penalty performs slightly better than an isotropic one. In a three-seed transformer study, interference is rank-correlated with merge degradation, but no more than Euclidean displacement.

### Source excerpt

arXiv:2611.25430v1 Announce Type: new Abstract: Combining fine-tuned models by composing their parameter updates has become a standard alternative to multi-task training. We formalize composability as a property of a collection of task vectors and prove that the excess loss of the merged model is controlled by an interference functional of the cross-task curvature overlap and the signed alignment of the updates. In the quadratic regime this functional equals the excess loss, giving a criterion that is necessary as well as sufficient and that measures the angle between task vectors in the curvature metric rather than the Euclidean one. Fine-tuning under a curvature-guided regularizer -- constructed from curvature statistics of the kind memory-efficient optimizers maintain -- keeps displacement in the base curvature metric below a ceiling set by the penalty coefficient at every iterate, at the standard $O(1/\sqrt{T})$ non-convex rate. Under a single Loewner comparison between the base statistic and the task curvatures, this control converts into a ceiling on the interference functional, inversely proportional to the penalty coefficient, so that the curvature term of the bound is controlled during training. A first-order term, present because regularized solutions are not stationary for the task loss, is not reduced in the same way. Experiments on an exact non-quadratic model confirm the budget and the ceiling, show the first-order term dominating for strong penalties, and find the curvature-guided penalty slightly better than an isotropic one. In a three-seed transformer study the functional is rank-correlated with merging degradation, but no more strongly than the Euclidean displacement norm.

## OceanCoder: A Verification-Guided Large Language Model for Automated Simulation Modelling in Ocean Engineering

DevFeed: [OceanCoder: A Verification-Guided Large Language Model for Automated Simulation Modelling in Ocean Engineering](<https://devfeed.tech/articles/oceancoder-a-verification-guided-large-language-model-for-automated-simulation-modelling-in-ocean-engineering-145165.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26333>)

Author: Can Xue, Changhong He, Ziyi Sun, Anliang Ge, Shimin Yu, Bingchen Liang

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>), [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [grpo](<https://devfeed.tech/topics/grpo.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Climate / Weather / Ocean Modeling](<https://devfeed.tech/topics/climate-weather-ocean-modeling.md>), [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>), [qwen](<https://devfeed.tech/topics/qwen.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [data-pipeline](<https://devfeed.tech/tags/data-pipeline.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [grpo](<https://devfeed.tech/tags/grpo.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [large-language-model](<https://devfeed.tech/tags/large-language-model.md>), [modelling](<https://devfeed.tech/tags/modelling.md>), [programming](<https://devfeed.tech/tags/programming.md>), [qwen3](<https://devfeed.tech/tags/qwen3.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [verification](<https://devfeed.tech/tags/verification.md>)

### AI overview

OceanCoder uses engineering verification as a training signal for automated ocean simulation modelling: successful code execution alone does not show that a model or its results meet engineering requirements. Built on Qwen3-8B, it is trained on 12,000 verified examples across 12 task categories, then guided by graded rewards from checks of inputs and constraints, completion of required analyses, and plausibility of computed outputs. The paper reports Pass@1 rising from 0% for the base model to 46.77% after supervised fine-tuning and 51.43% after group relative policy optimisation, outperforming the compared base models and closed-source system. Its benchmark includes out-of-distribution tasks and real-world engineering cases; the supplied text does not specify the comparison system or broader deployment limitations.

### Source excerpt

arXiv:2611.26333v1 Announce Type: new Abstract: Numerical simulation is important in ocean engineering, but model construction requires expert knowledge. Large language models support automated modelling, but code syntax and successful execution alone do not establish whether generated models and results meet engineering requirements. We present OceanCoder, a Qwen3-8B-based large language model for automated simulation modelling in ocean engineering. To address specialised programming requirements and scarce training data, we develop an extensible data construction pipeline producing 12,000 verified supervised fine-tuning (SFT) examples across 12 engineering task categories. A three-layer verification framework checks model inputs against engineering parameters and task constraints, verifies completion of required simulation analyses, and assesses the engineering plausibility of computed outputs. Its outcomes provide graded rewards for group relative policy optimisation (GRPO), guiding the model towards engineering task completion. We also construct the first benchmark for automated modelling in ocean engineering, covering out-of-distribution tasks and real-world engineering cases. Experiments and ablation studies demonstrate the effectiveness of the data pipeline, verification framework and reward design. OceanCoder outperforms both base models and the closed-source comparison, with Pass@1 increasing from 0% for the base model to 46.77% after SFT and 51.43% after GRPO. These results support engineering verification as a training signal for automated simulation modelling.

## FEA solver integration of a composite World Model with a Spiking Neural Network memory Block for computationally efficient modelling of defects and variabilities

DevFeed: [FEA solver integration of a composite World Model with a Spiking Neural Network memory Block for computationally efficient modelling of defects and variabilities](<https://devfeed.tech/articles/fea-solver-integration-of-a-composite-world-model-with-a-spiking-neural-network-memory-block-for-computationally-efficient-modelling-of-defects-and-variabilities-145179.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26419>)

Author: Bassam El Said, Adam Sobey

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [model architecture](<https://devfeed.tech/topics/model-architecture.md>), [Transformer](<https://devfeed.tech/topics/transformer.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [model-deployment](<https://devfeed.tech/topics/model-deployment.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [defects](<https://devfeed.tech/tags/defects.md>), [embedded](<https://devfeed.tech/tags/embedded.md>), [encoder](<https://devfeed.tech/tags/encoder.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [inference](<https://devfeed.tech/tags/inference.md>), [integration](<https://devfeed.tech/tags/integration.md>), [material](<https://devfeed.tech/tags/material.md>), [memory](<https://devfeed.tech/tags/memory.md>), [model](<https://devfeed.tech/tags/model.md>), [modelling](<https://devfeed.tech/tags/modelling.md>), [models](<https://devfeed.tech/tags/models.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [snn](<https://devfeed.tech/tags/snn.md>), [state](<https://devfeed.tech/tags/state.md>), [structure](<https://devfeed.tech/tags/structure.md>), [world-model](<https://devfeed.tech/tags/world-model.md>)

### AI overview

The proposed Composites World Model aims to make structural-scale composite simulation more computationally practical by replacing costly material-scale models with a surrogate that can represent local architectures and defects, track irreversible state, and run recursively. It combines a material encoder, a Spiking Neural Network memory block for long-term material state, and transformer blocks for short-term applied strain. Training proceeds through architecture encoding, short-window base training, and recursive fine-tuning with scheduled sampling to limit cumulative error drift. At inference, the model is disassembled and coupled to a nonlinear FE solver. Benchmarks on IM7/8552 laminates under tension, compression, and transverse shear, including pristine and wrinkled cases, report prediction errors of 0.4%-4.8%; the abstract attributes successful tracking of material state to the SNN memory block.

### Source excerpt

arXiv:2611.26419v1 Announce Type: new Abstract: Each fibre-reinforced composite structure has a unique material architecture. The manufacturing methods used and the presence of defects lead to local variations which require material-scale models that can resolve the response in real time at every location in the structure. Concurrent multi-scale schemes provide this capability, but at a prohibitive computational cost for large structures as the number of integration points and time increments scales with size. Surrogates have been studied extensively as replacements for material-scale models to mitigate these computational limitations. A surrogate must represent an arbitrary local architecture including defects, track the irreversible material state over an unknown number of increments, remain stable when driven recursively by its own predictions, and fit within the memory available at every integration point. Here, a Composites World Model based on a Multi-Modal Spiking Transformer architecture is designed to meet these requirements. The proposed architecture combines a material encoder, a Spiking Neural Network memory block that tracks the long-term internal state of the material, and transformer blocks that learn the short-term applied strain state. The system is trained in three stages: material architecture encoding, base training on short time windows, and recursive fine-tuning with scheduled sampling to suppress cumulative error drift. The trained model is disassembled at inference and coupled to a non-linear FE solver, enabling structural scale simulation of composites containing embedded defects. Benchmarks against experimental results for IM7/8552 laminates under tension, compression and transverse shear, for pristine and wrinkled composites, show that the SNN memory block successfully tracks the material internal state and results in prediction errors between 0.4% and 4.8%.

## Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation

DevFeed: [Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation](<https://devfeed.tech/articles/kandinsky-6-0-video-foundation-models-for-synchronized-video-and-audio-generation-144957.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.05608>)

Author: Team Kandinsky, Julia Agafonova, Bulat Akhmatov, Mikhail Aksyutin, Grigorii Alekseenko, Anastasia Aliaskina, Olga Androsova, Vladimir Arkhipkin, Anna Averchenkova, Alexander Belykh, Serafima Bocharova

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [multimodal-ai](<https://devfeed.tech/topics/multimodal-ai.md>), [Generative AI](<https://devfeed.tech/topics/generative-ai.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [diffusers](<https://devfeed.tech/topics/diffusers.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [attention](<https://devfeed.tech/tags/attention.md>), [audio](<https://devfeed.tech/tags/audio.md>), [audio-video](<https://devfeed.tech/tags/audio-video.md>), [code](<https://devfeed.tech/tags/code.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-mm](<https://devfeed.tech/tags/cs-mm.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [diffusion-models](<https://devfeed.tech/tags/diffusion-models.md>), [distillation](<https://devfeed.tech/tags/distillation.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [generate](<https://devfeed.tech/tags/generate.md>), [generation](<https://devfeed.tech/tags/generation.md>), [image](<https://devfeed.tech/tags/image.md>), [integration](<https://devfeed.tech/tags/integration.md>), [kandinsky](<https://devfeed.tech/tags/kandinsky.md>), [kandinsky-6-0-video](<https://devfeed.tech/tags/kandinsky-6-0-video.md>), [models](<https://devfeed.tech/tags/models.md>), [video](<https://devfeed.tech/tags/video.md>)

### AI overview

Kandinsky 6.0 Video offers 3B-parameter Lite and 29B-parameter Pro diffusion models for text- or image-conditioned generation of five-second clips with synchronized 44 kHz audio and lip-sync; built-in super-resolution targets Full HD. Its dual-stream CrossDiT links a pretrained video stream and a newly trained audio stream with bidirectional cross-attention. Training progresses from audio-only pretraining to joint paired audio-video training, followed by supervised fine-tuning, reinforcement-learning post-training, and distillation. In side-by-side human evaluation, Pro outperformed Kandinsky 5.0 Video Pro and was competitive with leading models; the authors report better speech quality than LTX 2.5 and Seedance 2.0. Code, checkpoints, and diffusers integration are released under the MIT license.

### Source excerpt

arXiv:2610.05608v2 Announce Type: replace-cross Abstract: We present Kandinsky 6.0 Video, a family of foundation diffusion models for synchronized text-to-audio-video generation, comprising Kandinsky 6.0 Video Lite (3B parameters) and Kandinsky 6.0 Video Pro (29B parameters). Both models generate 5-second video clips with synchronized 44 kHz audio, including lip-sync, in text-to-audio-video (T2AV) and image-to-audio-video (I2AV) modes; a built-in super-resolution model raises the output resolution to Full-HD (1920x1080). Building on the video generation capabilities of Kandinsky 5.0, Kandinsky 6.0 Video employs a dual-stream CrossDiT architecture that connects a pretrained video stream and a newly trained audio stream through bidirectional cross-attention for temporal and semantic alignment. Our continuous pretraining strategy first trains the audio stream from scratch on large-scale audio corpora and then trains both streams jointly on paired audio-video data while preserving unimodal fidelity; pretraining is followed by supervised fine-tuning, reinforcement-learning-based post-training, and distillation. In side-by-side human evaluation, Kandinsky 6.0 Video Pro clearly outperforms its predecessor, Kandinsky 5.0 Video Pro, and remains competitive with leading audio-video generation models, outperforming LTX 2.5 and Seedance 2.0 in speech quality. To accelerate open research and deployment in multimedia generation, we release the code, model checkpoints, and diffusers integration under the MIT license.

## NL2AGBench: Benchmarking LLM Auto-Formalization for AlphaGeometry

DevFeed: [NL2AGBench: Benchmarking LLM Auto-Formalization for AlphaGeometry](<https://devfeed.tech/articles/nl2agbench-benchmarking-llm-auto-formalization-for-alphageometry-145365.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2608.28481>)

Author: Samuel Xiao, Judy Song, Rory Hu, Ziliang Zong

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Language Engineering](<https://devfeed.tech/topics/language-engineering.md>), [domain-specific language](<https://devfeed.tech/topics/domain-specific-language.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [machine-translation](<https://devfeed.tech/topics/machine-translation.md>)

Tags: [arxiv](<https://devfeed.tech/tags/arxiv.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [domain-specific-language](<https://devfeed.tech/tags/domain-specific-language.md>), [error](<https://devfeed.tech/tags/error.md>), [execution](<https://devfeed.tech/tags/execution.md>), [few-shot](<https://devfeed.tech/tags/few-shot.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [llm](<https://devfeed.tech/tags/llm.md>), [open-source-models](<https://devfeed.tech/tags/open-source-models.md>), [performance](<https://devfeed.tech/tags/performance.md>), [source](<https://devfeed.tech/tags/source.md>), [taxonomy](<https://devfeed.tech/tags/taxonomy.md>), [translation](<https://devfeed.tech/tags/translation.md>)

### AI overview

NL2AGBench evaluates how well LLMs translate English geometry problems into AlphaGeometry's specialized DSL, using execution within AlphaGeometry to assess whether translations work rather than relying only on textual similarity. Across ten open and closed models at multiple scales, leading closed models exceed 80% executable translation, while even the largest open models struggle to preserve geometric constraints and produce valid formalizations. The benchmark distinguishes syntax from logic errors and reports measurable improvements from few-shot prompting, fine-tuning, and human-guided hints across multiple model families.

### Source excerpt

arXiv:2608.28481v2 Announce Type: replace-cross Abstract: Recent advances in large language models (LLMs) have demonstrated strong capabilities in natural language understanding and mathematical reasoning. However, their ability to translate informal mathematical problems into formal representations remains underexplored. This limitation is particularly important for neuro-symbolic geometry systems such as AlphaGeometry, whose theorem-proving engine requires inputs in a specialized domain-specific language (DSL). Although AlphaGeometry achieves near-IMO gold-medalist performance, manually converting natural-language problems into its formal syntax remains a significant usability bottleneck. To address this challenge, we introduce the Natural Language to AlphaGeometry Benchmark (NL2AGBench), which evaluates LLMs in translating English geometry problems into AlphaGeometry-compatible formal representations. NL2AGBench uses execution-based verification within AlphaGeometry to assess translation quality rather than relying solely on textual similarity. We evaluate ten state-of-the-art open- and closed-source LLMs across multiple parameter scales and analyze executable translation accuracy, syntactic correctness, and error characteristics. Our experiments reveal a substantial performance gap between closed- and open-source models: leading closed-source models achieve executable translation rates above 80%, while even the largest open-source models struggle to consistently preserve geometric constraints and produce valid formalizations. We introduce an error taxonomy distinguishing syntax and logic errors and investigate mitigation strategies, including few-shot prompting, fine-tuning, and human-guided hinting, which yield measurable improvements across multiple model families.

## Sage-Evolve-Preview:Adaptive Observation for Post-Training Self-Improvement

DevFeed: [Sage-Evolve-Preview:Adaptive Observation for Post-Training Self-Improvement](<https://devfeed.tech/articles/sage-evolve-preview-adaptive-observation-for-post-training-self-improvement-145635.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26132>)

Author: Yinhao Tang, Youqing Fang, Yanan Sun, Xintian Shen, Hanyu Sun, Wenran Liu, Yue Wu, Weiming Zhang, Bin Liu, Wenwei Zhang

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [post-training](<https://devfeed.tech/topics/post-training.md>), [Advanced AI workflows](<https://devfeed.tech/topics/advanced-ai-workflows.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [agent](<https://devfeed.tech/tags/agent.md>), [anchor](<https://devfeed.tech/tags/anchor.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [budget](<https://devfeed.tech/tags/budget.md>), [controller](<https://devfeed.tech/tags/controller.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [diagnostic](<https://devfeed.tech/tags/diagnostic.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [evolution](<https://devfeed.tech/tags/evolution.md>), [harness](<https://devfeed.tech/tags/harness.md>), [independent](<https://devfeed.tech/tags/independent.md>), [long-context](<https://devfeed.tech/tags/long-context.md>), [post-training](<https://devfeed.tech/tags/post-training.md>), [progress](<https://devfeed.tech/tags/progress.md>), [rsi](<https://devfeed.tech/tags/rsi.md>), [score](<https://devfeed.tech/tags/score.md>), [self-improvement](<https://devfeed.tech/tags/self-improvement.md>), [tasks](<https://devfeed.tech/tags/tasks.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

AdaObs addresses plateaus in post-training recursive self-improvement by adapting what the training loop observes while keeping its anchor objective fixed. At a plateau, its controller sharpens diagnostics, raises task difficulty, and adds related tasks; validated gains feed into training, and checkpoints remain scored against the anchor. Across eight anchors in long-context reading comprehension and instruction following, mean scores rose from 66.4 to 73.7 under a matched data budget, outperforming the FixedObs anchor-only baseline. The evolved harness also improved the seven-task average on independent PostTrainBench across two agent backends. The evidence is limited to the reported benchmarks and settings.

### Source excerpt

arXiv:2611.26132v1 Announce Type: new Abstract: Post-training recursive self-improvement (RSI) iteratively improves a target capability, operationalized by a fixed anchor benchmark, through repeated intervention, training, and evaluation. Yet most loops use the anchor's aggregate score as their sole observation, so early gains give way to stalled progress and repetitive exploration. When this anchor-only feedback ceases to reveal actionable gaps, a plateau may reflect an exhausted observation rather than an exhausted capability. We introduce adaptive observation (AdaObs), an observation-side mechanism of harness evolution that keeps the anchor objective fixed while adapting the active observation. At a plateau, a controller refines diagnostic granularity, escalates difficulty, and adds related tasks; validated increments are fused back into training, while every checkpoint remains evaluated against the anchor. Across eight independently optimized anchors in long-context reading comprehension and instruction following, AdaObs raises the mean score from $66.4$ to $73.7$ under a matched data budget, outperforming FixedObs, the anchor-only baseline. The evolved harness further improves the seven-task average on the independent PostTrainBench across two agent backends. These results identify observation adaptation as a missing mechanism of harness evolution for sustaining post-training RSI.

## UniCom: A Unified Framework for Community Search and Detection via Transfer Learning

DevFeed: [UniCom: A Unified Framework for Community Search and Detection via Transfer Learning](<https://devfeed.tech/articles/unicom-a-unified-framework-for-community-search-and-detection-via-transfer-learning-145729.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2512.02460>)

Author: Yifan Zhu, Hanchen Wang, Wenjie Zhang, Alexander Zhou, Ying Zhang

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>)

Topics: [Model Development](<https://devfeed.tech/topics/model-development.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [community](<https://devfeed.tech/tags/community.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-si](<https://devfeed.tech/tags/cs-si.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [detection](<https://devfeed.tech/tags/detection.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [framework](<https://devfeed.tech/tags/framework.md>), [graph](<https://devfeed.tech/tags/graph.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [learning](<https://devfeed.tech/tags/learning.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [search](<https://devfeed.tech/tags/search.md>), [transfer-learning](<https://devfeed.tech/tags/transfer-learning.md>)

### AI overview

UniCom combines community search, which finds a community for a query, and community detection, which partitions a graph, in one transfer-learning framework. Its Universal Graph Learning stage distills local and global community knowledge from multiple source domains; Domain-aware Specialization then adapts to unseen graphs and tasks on the fly, avoiding costly retraining while keeping the model compact. The abstract reports experiments on 16 datasets against 22 baselines, claiming UniCom outperforms all state-of-the-art baselines across both tasks while maintaining runtime efficiency. The supplied evidence does not give performance details or explain the adaptation procedure.

### Source excerpt

arXiv:2512.02460v2 Announce Type: replace-cross Abstract: Searching and detecting communities in real-world graphs underpins a wide range of applications. Specifically, community search identifies a community for a given query, whereas community detection partitions the entire graph into multiple communities. Although they target different community analysis objectives, both tasks aim to uncover cohesive community structures and are complementary in real-world applications. Despite this connection, current learning-based solutions generally regard them as separate problems, necessitating task- and dataset-specific retraining. This separated learning paradigm limits the applicability and generalization ability of the existing models. To mitigate this limitation, we propose UniCom, a unified framework to solve both community search and detection tasks. UniCom centers on a Universal Graph Learning (UGL) stage that distills local and global community-aware knowledge from multiple source domains. Following UGL, the Domain-aware Specialization (DAS) stage supports on-the-fly adaptation to unseen graphs and tasks through transfer learning, avoiding costly retraining while keeping the model compact. Extensive experiments on both tasks across 16 datasets and 22 baselines have been conducted to ensure a comprehensive, rigorous, and fair evaluation. UniCom consistently outperforms all state-of-the-art baselines across all tasks while maintaining runtime efficiency.

## LingSante MedGuard Technical Report

DevFeed: [LingSante MedGuard Technical Report](<https://devfeed.tech/articles/lingsante-medguard-technical-report-146034.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25509>)

Author: LingSante MedGuard Team

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [ai safety](<https://devfeed.tech/topics/ai-safety.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [ai](<https://devfeed.tech/tags/ai.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [clinical-decision-support](<https://devfeed.tech/tags/clinical-decision-support.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [distillation](<https://devfeed.tech/tags/distillation.md>), [ethics](<https://devfeed.tech/tags/ethics.md>), [f1](<https://devfeed.tech/tags/f1.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [guardrail](<https://devfeed.tech/tags/guardrail.md>), [hallucination-detection](<https://devfeed.tech/tags/hallucination-detection.md>), [medical](<https://devfeed.tech/tags/medical.md>), [medical-ai](<https://devfeed.tech/tags/medical-ai.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [risk-assessment](<https://devfeed.tech/tags/risk-assessment.md>), [technical](<https://devfeed.tech/tags/technical.md>), [traceability](<https://devfeed.tech/tags/traceability.md>)

### AI overview

LingSante MedGuard combines a medical guardrail teacher with six specialized capability experts in one model to assess medical safety, compliance, and ethics. Training uses supervised fine-tuning on teacher-generated demonstrations, then multi-teacher on-policy distillation from feedback on student responses. Its MedGuardBench covers 11 risk-detection subsets; mean F1 scores are 82.3%, 82.4%, and 87.4% for 4B, 9B, and 27B models. The 4B model reaches 73.9% mean accuracy on MedHallu and MedHALT-NOTA with retrieval enabled. The abstract also reports gains in evidence-supported consistency, source traceability, and rule-guided rationale quality, without specifying their magnitude.

### Source excerpt

arXiv:2611.25509v1 Announce Type: new Abstract: As medical AI is increasingly used for patient consultation and clinical decision support, identifying harmful, non-compliant, or ethically problematic responses is essential for responsible deployment. Assessing these risks requires checking clinical claims against relevant evidence, verifying calculations, and applying medical requirements under their intended conditions. We introduce LingSante MedGuard, a capability-enhanced guardrail that integrates supervision from a medical guardrail teacher model and six specialized medical capability experts into a unified model. The guardrail teacher supervises risk decisions and explanations across medical safety, compliance, and ethics, while the capability experts support medical rule understanding, numerical reasoning, multimodal hallucination detection, agentic safety verification, risk localization, and multimodal entity understanding. We consolidate these sources of supervision through offline supervised fine-tuning on teacher-generated demonstrations, followed by multi-teacher on-policy distillation using feedback on student-generated responses. We further introduce LingSante MedGuardBench, a benchmark comprising 11 risk-detection subsets spanning medical safety, compliance, and ethics. At 4B, 9B, and 27B scales, LingSante MedGuard achieves mean benchmark F1 scores of 82.3%, 82.4%, and 87.4%, respectively. The 4B model also achieves 73.9% mean accuracy across MedHallu and MedHALT-NOTA under retrieval-enabled evaluation. Additional evaluations of the verification components show improvements in evidence-supported consistency, source traceability, and rule-guided rationale quality across model scales. These results support the integration of specialized medical capabilities into a unified guardrail for medical risk assessment and evidence-based medical safety verification.

## What If Chain-of-Thought Is Not Enough?

DevFeed: [What If Chain-of-Thought Is Not Enough?](<https://devfeed.tech/articles/what-if-chain-of-thought-is-not-enough-145648.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26282>)

Author: Marc Jansen

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [decision-making](<https://devfeed.tech/topics/decision-making.md>), [Chain-of-thought](<https://devfeed.tech/topics/chain-of-thought.md>)

Tags: [chain-of-thought](<https://devfeed.tech/tags/chain-of-thought.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [decision](<https://devfeed.tech/tags/decision.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [structured](<https://devfeed.tech/tags/structured.md>)

### AI overview

The paper tests whether LLMs' internal reasoning can be deliberately changed by fine-tuning them to use Question-Option-Criteria (QOC): break a task into explicit questions, competing options, and evaluation criteria. Its three-generation pipeline progresses from a basic decision core to richer answer synthesis and statistically enriched reasoning. The reported experiments assess internal consistency and output quality; QOC fine-tuning improves answer quality over non-fine-tuned base models, with especially strong gains when models are large enough to carry out the QOC computations reliably. The abstract presents structured reasoning as a promising alternative to Chain-of-Thought, while providing no specific benchmark results or detailed limitations.

### Source excerpt

arXiv:2611.26282v1 Announce Type: cross Abstract: Reasoning capabilities in large language models have evolved from prompting-based methods to internalized model behaviors. In this paper, we investigate whether the internal thinking model of an LLM can be deliberately changed and whether such a change is beneficial. Instead of relying on Chain-of-Thought, we train models to follow Question--Option--Criteria (QOC), a structured decision process that decomposes a task into explicit questions, competing options, and evaluation criteria. We present a three-generation development pipeline that ranges from a basic QOC decision core to richer answer synthesis and statistically enriched QOC reasoning. We then evaluate the resulting models with respect to both internal consistency and output quality. Across the reported experiments, QOC-based fine-tuning improves answer quality relative to non-fine-tuned base models and yields particularly strong gains once models are large enough to execute the underlying QOC computations reliably. The results suggest that the internal thinking model of an LLM is trainable and that structured alternatives to Chain-of-Thought can provide a promising path toward more reliable, interpretable, and task-adapted decision support.

## Learning to Read Events for Exact State Tracking with Frozen Language Models

DevFeed: [Learning to Read Events for Exact State Tracking with Frozen Language Models](<https://devfeed.tech/articles/learning-to-read-events-for-exact-state-tracking-with-frozen-language-models-145024.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25211>)

Author: Jeonghoon Lee (Attractor Dynamics Inc.)

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [event-streams](<https://devfeed.tech/tags/event-streams.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [learning](<https://devfeed.tech/tags/learning.md>), [state](<https://devfeed.tech/tags/state.md>), [tracking](<https://devfeed.tech/tags/tracking.md>)

### AI overview

For exact state tracking across long event streams, a frozen language model can extract individual transitions while an external integer register applies them in order. A trained selector paired with a frozen 0.8B model read every event in 299 windows of 43,687 events, but a fresh-machine run missed the stated 0.99 lower-confidence-bound target (298/299); a separate follow-up achieved 472/473. On 1,000-event room-swap documents, extraction plus the register solved all five documents for each of seven tested models, without requested chain of thought; only three of ten models solved all five by reasoning alone. Concurrent extraction reduced median time and mean per-document cost for Sonnet 5.5 and GPT-6.1 Sol. A regular-expression executor also solved the 15 fixed-format documents, while on free-form issue comments an adapted 4B reader reached 68.8%, versus 61.0% for a frequency baseline. The abstract reports these benchmark results; they do not establish performance on other tasks or data.

### Source excerpt

arXiv:2611.25211v1 Announce Type: new Abstract: Long event streams require identifying each change and applying the changes in order. We train a selector on a frozen language model to read each transition. With API-only access, we ask the model to list the transitions. An external integer register composes them in order. A confidence rule can flag uncertain updates or retain several candidates. We study products of permutations of five items, a task that fixed-depth, log-precision transformers cannot solve on all input lengths under the standard conjecture that $\mathrm{TC}^0 \ne \mathrm{NC}^1$. On two-event windows, a 27B model answered 3 of 100 correctly without visible reasoning and 96 of 100 when it generated intermediate steps. A frozen 0.8B model with a trained selector read every event in 299 of 299 windows of 43,687 events each. The same package passed 298 of 299 new windows on a fresh machine, missing the target of 0.99 for the one-sided 95% lower confidence bound. A separate follow-up sized to allow one failure yielded 472 of 473. On 1,000-event room-swap documents, three of ten models solved all five documents by reasoning alone. All seven models tested both ways at this length solved all five with extraction and the register, with no chain of thought requested. With concurrent extraction, median wall-clock time fell from 136.1 to 10.4 s for Sonnet 5.5 and from 313.5 to 28.0 s for GPT-6.1 Sol. At fixed per-token rates, mean cost per document fell by 55% for Sonnet 5.5 and 44% for GPT-6.1 Sol. A regular expression with an executor also solved all 15 fixed-format documents. On 500 free-form issue-tracker comments, a separately adapted 4B reader identified the recorded value in 68.8% of cases, compared with 61.0% for a frequency baseline fitted on the same training projects.

## Cyber-Physical RSI for AI-RAN: A Case Study

DevFeed: [Cyber-Physical RSI for AI-RAN: A Case Study](<https://devfeed.tech/articles/cyber-physical-rsi-for-ai-ran-a-case-study-145554.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25619>)

Author: Binghan Wu, Kai Zeng, Zhiqi Li, Zhigang Wang, Yang Liu, Ya-Qin Zhang, Joseph Sifakis, Ye Ouyang

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Omniverse](<https://devfeed.tech/topics/omniverse.md>), [ReAct agent Python](<https://devfeed.tech/topics/react-agent-python.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [case-study](<https://devfeed.tech/tags/case-study.md>), [computer](<https://devfeed.tech/tags/computer.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [experiment](<https://devfeed.tech/tags/experiment.md>), [feedback](<https://devfeed.tech/tags/feedback.md>), [framework](<https://devfeed.tech/tags/framework.md>), [ping-pong](<https://devfeed.tech/tags/ping-pong.md>), [radio](<https://devfeed.tech/tags/radio.md>), [recursive-self-improvement](<https://devfeed.tech/tags/recursive-self-improvement.md>), [rsi](<https://devfeed.tech/tags/rsi.md>), [validation](<https://devfeed.tech/tags/validation.md>)

### AI overview

The paper presents a cyber-physical recursive self-improvement framework for AI-RAN that uses physical measurements to adapt a student model and guide later experiments. It links the student with an experiment-planning coach, a calibratable virtual environment, and a physical host agent. In a ping-pong handover case study, 12 physical executions were organized into six baseline-candidate pairs, and the revised model was evaluated on reserved physical configurations. In short-window tests, raising the time-to-trigger from 320 to 1280 ms reduced the ping-pong count from three to zero. The updated student's grouped-validation MAE was 2.673 round trips/min; on a common three-configuration physical evaluation, pre-feedback and post-feedback MAEs were 1.240 and 1.601 round trips/min, respectively. The results show that measured outcomes inform the feedback loop, while generalization varies with physical configuration.

### Source excerpt

arXiv:2611.25619v1 Announce Type: new Abstract: Learned models for radio access networks (RANs) are often developed using simulation, yet their predictions can shift when exposed to physical propagation conditions, protocol behavior, and equipment implementations. This paper presents a cyber-physical recursive self-improvement (RSI) framework for AI-RAN that connects a task-oriented student model, an experiment-planning coach agent, a calibratable virtual environment, and a physical host agent. Physical measurements are organized as evidence for both model adaptation and subsequent experiment planning. We examine an implementation of the physical-feedback pathway using ping-pong handover as a case study. The system completes 12 physical executions organized into six baseline--candidate pairs, updates the student from measured outcomes, and evaluates the revised model on reserved physical configurations. In the observed short-window tests, increasing the time-to-trigger from 320 to 1280\,ms changes the ping-pong count from three to zero. The updated student achieves a grouped-validation mean absolute error (MAE) of 2.673 round trips/min, while a common three-configuration physical evaluation yields MAEs of 1.240 and 1.601 round trips/min for the pre-feedback and post-feedback models, respectively. These results characterize both the operational feedback loop and the parameter-dependent generalization behavior revealed by new physical experiments.

## SKILL.PNG: Learning Agent Skills in Pixel Space

DevFeed: [SKILL.PNG: Learning Agent Skills in Pixel Space](<https://devfeed.tech/articles/skill-png-learning-agent-skills-in-pixel-space-145160.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26296>)

Author: Maximilian Seeth, Chengzhi Martin Hu, Marie Kempkes, Paul R\\"ottger, Leo Schwinn

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>), [vlm](<https://devfeed.tech/topics/vlm.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [gradient-descent](<https://devfeed.tech/tags/gradient-descent.md>), [hybrid](<https://devfeed.tech/tags/hybrid.md>), [image](<https://devfeed.tech/tags/image.md>), [instructions](<https://devfeed.tech/tags/instructions.md>), [language](<https://devfeed.tech/tags/language.md>), [learning](<https://devfeed.tech/tags/learning.md>), [llms](<https://devfeed.tech/tags/llms.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [optimized](<https://devfeed.tech/tags/optimized.md>), [performance](<https://devfeed.tech/tags/performance.md>), [pixel](<https://devfeed.tech/tags/pixel.md>), [pixels](<https://devfeed.tech/tags/pixels.md>), [skill](<https://devfeed.tech/tags/skill.md>), [vision-language-models](<https://devfeed.tech/tags/vision-language-models.md>), [vision-language-models-vlms](<https://devfeed.tech/tags/vision-language-models-vlms.md>), [vlms](<https://devfeed.tech/tags/vlms.md>)

### AI overview

Skill.png represents agent skills as images whose pixels are optimized by gradient descent through the vision encoder of a frozen vision-language model (VLM). The authors report that these image skills beat frontier-LLM-optimized textual skills in 9 of 16 settings across four VLMs and four benchmarks, using 64-256 image tokens at inference versus about 900-2,000 text tokens. Jointly optimized text and image skills further improve performance. An image optimized across an ensemble can transfer among its VLMs, while the reported compatibility claim is for VLMs. The abstract describes the evaluation results but does not provide per-setting details.

### Source excerpt

arXiv:2611.26296v1 Announce Type: new Abstract: AI agents are commonly specialized for new tasks and domains using skills, i.e. natural language instructions placed in their context. However, the discrete nature of textual skills prevents continuous optimization. Gradient-based approaches such as finetuning, on the other hand, are typically model-specific, preventing skill transfer across different agents. Since many agents today are built on vision-language models (VLMs), we can address these limitations by introducing Skill.png: a method for obtaining skill-like image inputs whose pixels are optimized with gradient descent through the vision encoder of a frozen VLM. Image skills are compatible with any VLM, and a single image can be optimized jointly for an ensemble of models. We show that Skill.png significantly outperforms state-of-the-art textual skills optimized using frontier LLMs in 9 of 16 settings (across 4 VLMs and 4 benchmarks) while requiring only 64 to 256 image tokens instead of roughly 900 to 2000 text tokens at inference time. Hybrid skills, i.e. text and image skills jointly optimized, further improve agent performance. Moreover, an image optimized with an ensemble of VLMs can successfully transfer across the models in the ensemble. Skill.png thus presents the first agent skill that can be optimized with gradient descent while retaining transfer benefits of text skills. Our code is available at: https://github.com/omseeth/skill.png

## Watch Where You Linearize: Predicting and Correcting Model-Merging Interference

DevFeed: [Watch Where You Linearize: Predicting and Correcting Model-Merging Interference](<https://devfeed.tech/articles/watch-where-you-linearize-predicting-and-correcting-model-merging-interference-144999.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25000>)

Author: Ido Nitzan Hidekel, Dan Raviv

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [lora](<https://devfeed.tech/tags/lora.md>), [matching](<https://devfeed.tech/tags/matching.md>), [model](<https://devfeed.tech/tags/model.md>), [predictions](<https://devfeed.tech/tags/predictions.md>), [vector](<https://devfeed.tech/tags/vector.md>)

### AI overview

Evaluating a first-order directional derivative at the receiver model predicts nearby task-merging interference: it halves the accuracy error of a no-change baseline and identifies most examples whose predictions flip. Full multi-task merges require two evaluations along the merge path. Derivatives evaluated at the shared pretrained backbone perform poorly in vision experiments and are only partly predictive in LoRA language-model pilots; controls associate the mismatch with movement along the receiver's own task vector. Freezing receiver-local responses yields an affine surrogate used by LR-Merge to optimize coefficients through convex teacher matching on unlabeled probes, without backpropagating through merged networks. LR-Merge improves the tested closed-form, isotropic, and sparsified mergers, with gains generally increasing as tasks are added. It matches exact teacher matching through a merged network, which takes roughly three times longer to converge, while both perform similarly at equal compute. The surrogate supports new coefficient and task-subset queries in seconds.

### Source excerpt

arXiv:2611.25000v1 Announce Type: new Abstract: Model merging combines separately fine-tuned models by adding their task vectors to a shared backbone. Interference is typically assessed only after constructing and evaluating a candidate merge, making coefficient search expensive. We show that a first-order directional derivative of task A's outputs along task B's vector can predict interference, but only when evaluated at the appropriate model. At the receiver model, this derivative forecasts nearby merges per task and per example, halves the accuracy error of a no-change baseline, and identifies most examples whose predictions flip. For full multi-task merges, a single evaluation is insufficient; two evaluations along the merge path are needed. At the shared pretrained backbone, however, the derivative is nearly orthogonal to realized changes in our vision experiments and only partly predictive in LoRA language-model pilots. Controls link this failure to displacement along the receiver's own task vector, not simply parameter distance; the mismatch appears within the first few dozen fine-tuning steps. Freezing receiver-local responses produces a surrogate affine in the merge coefficients. We use it to construct LR-Merge, a label-free correction that optimizes coefficients through convex teacher matching on unlabeled probes, without backpropagation through merged networks. LR-Merge improves all tested closed-form, isotropic, and sparsified base mergers, with gains generally increasing as more tasks are merged. Using pretrained rather than receiver-local responses instead degrades the base merge. LR-Merge matches exact teacher matching through a merged network, which takes roughly three times longer to converge and performs similarly at an equal compute budget. Once built, the surrogate answers new coefficient and task-subset queries in seconds.

## ZeroVLN: Learning to Navigate without Supervised Fine-Tuning of the VLM Brain

DevFeed: [ZeroVLN: Learning to Navigate without Supervised Fine-Tuning of the VLM Brain](<https://devfeed.tech/articles/zerovln-learning-to-navigate-without-supervised-fine-tuning-of-the-vlm-brain-145668.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26414>)

Author: Zixuan Huang, Yang Zhou, Guli Zhang, Sunzhu Li, Hongyan Xie, Zehao Chen, Fuzhen Zhuang, Kaixuan Wang, Xiaozhi Chen, Yikun Ban, Deqing Wang

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Robot Navigation](<https://devfeed.tech/topics/robot-navigation.md>), [vlm](<https://devfeed.tech/topics/vlm.md>), [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [capabilities](<https://devfeed.tech/tags/capabilities.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [model](<https://devfeed.tech/tags/model.md>), [natural-language](<https://devfeed.tech/tags/natural-language.md>), [navigation](<https://devfeed.tech/tags/navigation.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [sft](<https://devfeed.tech/tags/sft.md>), [tool-calling](<https://devfeed.tech/tags/tool-calling.md>), [vision-language-model](<https://devfeed.tech/tags/vision-language-model.md>)

### AI overview

ZeroVLN targets a tradeoff in vision-language navigation: fine-tuning the VLM brain on navigation data can change its weights and weaken general visual capabilities. Instead, the framework leaves the brain without navigation-domain supervised fine-tuning and gives it a semantic tool-calling interface to send natural-language movement requests to an action expert. The expert is first adapted to execute those requests; closed-loop reinforcement learning then jointly improves the brain's movement decisions and the expert's continuous control. On standard VLN-CE benchmarks, using the same navigation episodes, the authors report substantial gains over representative SFT-based hierarchical methods, with only minor changes to the brain's weights and largely preserved general capabilities.

### Source excerpt

arXiv:2611.26414v1 Announce Type: new Abstract: Representative vision-language navigation (VLN) methods typically couple a vision-language model (VLM) brain with a learnable action head and fine-tune the VLM brain on navigation-domain data to predict geometric goals. However, such adaptation can substantially alter base VLM weights and degrade general visual capabilities. We introduce ZeroVLN, a slow-fast navigation framework that achieves effective navigation with zero navigation-domain SFT of the VLM brain. Rather than teaching the brain to predict geometric goals, we equip it with a semantic tool-calling interface for conveying movement decisions to an action expert in natural language. We first adapt the action expert to carry out these movement requests, then jointly improve the brain's movement decisions and the action expert's continuous execution through closed-loop reinforcement learning (RL). Trained on the same navigation episodes, ZeroVLN substantially outperforms representative SFT-based hierarchical VLN methods on standard VLN-CE (continuous environments) benchmarks, while inducing only minor changes to the brain's weights and largely preserving its general capabilities. To our knowledge, ZeroVLN is the first slow-fast VLN framework with zero navigation-domain SFT of the VLM brain, and the first to jointly optimize its slow and fast policies through closed-loop agentic RL in VLN-CE.

## FlexTSF: A Flexible Forecasting Model for Time Series with Variable Regularities

DevFeed: [FlexTSF: A Flexible Forecasting Model for Time Series with Variable Regularities](<https://devfeed.tech/articles/flextsf-a-flexible-forecasting-model-for-time-series-with-variable-regularities-144849.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2410.23160>)

Author: Jingge Xiao, Yile Chen, Gao Cong, Wolfgang Nejdl, Simon Gottschalk

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [attention](<https://devfeed.tech/tags/attention.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [domains](<https://devfeed.tech/tags/domains.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [ivp](<https://devfeed.tech/tags/ivp.md>), [missing-values](<https://devfeed.tech/tags/missing-values.md>), [sampling](<https://devfeed.tech/tags/sampling.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [transformer](<https://devfeed.tech/tags/transformer.md>), [zero-shot](<https://devfeed.tech/tags/zero-shot.md>)

### AI overview

FlexTSF targets forecasting across time series with uneven sampling, variable lengths, and missing values. Its continuous-time IVP Patcher handles these irregularities, while a decoder-only Transformer uses normalized timestamps, domain-specific statistical descriptors, and specialized causal self-attention to adapt across domains. The abstract reports experiments on 16 datasets, with FlexTSF outperforming existing models in classic forecasting, zero-shot generalization, and low-resource fine-tuning; it provides no detailed metrics or comparison breakdown.

### Source excerpt

arXiv:2410.23160v3 Announce Type: replace Abstract: Generalizing across time series with irregular temporal structures poses a persistent challenge for universal time series forecasting models. Existing approaches often assume regular sampling or depend heavily on imputation, which limits their applicability in real-world scenarios where irregularities are prevalent due to diverse sensing devices and recording practices. To address this gap, we propose FlexTSF, a flexible and domain-adaptive forecasting model specifically designed for time series with variable temporal regularities. At its foundation lies the IVP Patcher, a continuous-time patching module leveraging Initial Value Problems (IVPs) to inherently handle uneven sampling intervals, variable sequence lengths, and missing values. Built upon a decoder-only Transformer architecture, FlexTSF integrates normalized timestamp inputs and domain-specific statistical descriptors through a specialized causal self-attention mechanism, enabling adaptability across domains. Extensive experiments on 16 datasets demonstrate FlexTSF's effectiveness, significantly outperforming existing models in classic forecasting scenarios, zero-shot generalization, and low-resource fine-tuning conditions.

## The Dichotomy Between Pattern Recognition and Step-by-Step Reasoning

DevFeed: [The Dichotomy Between Pattern Recognition and Step-by-Step Reasoning](<https://devfeed.tech/articles/the-dichotomy-between-pattern-recognition-and-step-by-step-reasoning-144959.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.09186>)

Author: Amrut Nadgir, Vijay Balasubramanian, Pratik Chaudhari

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [arxiv](<https://devfeed.tech/tags/arxiv.md>), [attention](<https://devfeed.tech/tags/attention.md>), [cond-mat-dis-nn](<https://devfeed.tech/tags/cond-mat-dis-nn.md>), [context](<https://devfeed.tech/tags/context.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [graph](<https://devfeed.tech/tags/graph.md>), [inference](<https://devfeed.tech/tags/inference.md>), [language](<https://devfeed.tech/tags/language.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [large-language-model](<https://devfeed.tech/tags/large-language-model.md>), [learning](<https://devfeed.tech/tags/learning.md>), [linear](<https://devfeed.tech/tags/linear.md>), [llm](<https://devfeed.tech/tags/llm.md>), [llms](<https://devfeed.tech/tags/llms.md>), [math-500](<https://devfeed.tech/tags/math-500.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [pattern](<https://devfeed.tech/tags/pattern.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [recognition](<https://devfeed.tech/tags/recognition.md>)

### AI overview

The paper models step-by-step reasoning as composing learned transitions between short token contexts in a de Bruijn graph. It argues and reports evidence that transformers need training samples proportional to the number of graph edges, which is far smaller than the number of possible reasoning traces. In experiments fine-tuning Qwen2.5-1.5B-Instruct on equations and story questions, frequent states (shorter contexts) improve accuracy but make models more fragile to test-time perturbations; sparse states offer a balance between accuracy and robustness, while large contexts perform no better than pattern recognition. For Qwen3-14B and Qwen3-32B, restricting attention to a sliding window under 15% of the full reasoning trace retains over 75% of accuracy on GSM8K, MATH-500, and GPQA-Diamond.

### Source excerpt

arXiv:2610.09186v3 Announce Type: replace Abstract: We argue that pattern recognition and step-by-step reasoning are two ends of a spectrum. A large language model (LLM) learns to reason step-by-step when data is structured such that the next token depends on a small amount of preceding context. Inference in LLMs resembles pattern recognition when the next token depends on a large amount of preceding context. If the next token depends on only the $c$ most recent tokens, reasoning traces are paths on a de Bruijn graph whose nodes are $c$-length contexts and edges are next-token transitions between contexts. The set of reasoning traces of a task forms a directed acyclic subgraph of the de Bruijn graph. An LLM that has learned all edges of this subgraph can compose them to solve longer, unseen tasks, i.e., it reasons step-by-step. We prove that the number of edges is vanishingly small compared to the number of reasoning traces. We also prove and show empirically that the number of training samples a transformer needs is a linear in the number of edges, so learning to reason step-by-step is sample efficient. We can induce de Bruijn structure in any task by maintaining a ``state'' that makes future reasoning independent of the past. The frequency of states in the reasoning trace determines $c$. We show, by fine-tuning Qwen2.5-1.5B-Instruct to solve equations and answer questions about stories, that frequent states (small $c$) result in higher accuracy but greater fragility to perturbations at test time. LLMs trained with a large $c$ are only as good as models that perform pattern recognition without reasoning. A moderate density of states balances accuracy and robustness. Real-world data has de Bruijn structure: Qwen3-14B and Qwen3-32B retain over 75\% of their accuracy on GSM8K, MATH-500 and GPQA-Diamond when attention is restricted to a sliding window less than 15\% as long as the full reasoning trace.

## Goal-Guided Reinforcement Learning via Fine-Tuning Vision-Language Models

DevFeed: [Goal-Guided Reinforcement Learning via Fine-Tuning Vision-Language Models](<https://devfeed.tech/articles/goal-guided-reinforcement-learning-via-fine-tuning-vision-language-models-145097.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25742>)

Author: Yajie Qi, Wei Wei, Lin Li, Lijun Zhang, Jianye Hao

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [vlm](<https://devfeed.tech/topics/vlm.md>), [dpo](<https://devfeed.tech/topics/dpo.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [direct-preference-optimization](<https://devfeed.tech/tags/direct-preference-optimization.md>), [exploration](<https://devfeed.tech/tags/exploration.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [kl-regularization](<https://devfeed.tech/tags/kl-regularization.md>), [large-language-models-llms](<https://devfeed.tech/tags/large-language-models-llms.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [vision-language-models](<https://devfeed.tech/tags/vision-language-models.md>)

### AI overview

The proposed framework fine-tunes a vision-language model on CCGP-2M, a multimodal goal-preference dataset built from Craftax-Classic interaction trajectories, then uses the adapted model to guide online reinforcement learning. Offline fine-tuning uses KLTDPO, combining token-level KL regularization with distributionally robust reweighting. In Craftax-Classic experiments, the method collected diamonds earlier than state-of-the-art baselines, achieved the highest score and reward, and had the highest success rates on most achievements. The source does not provide details about the baselines or quantitative effect sizes.

### Source excerpt

arXiv:2611.25742v1 Announce Type: new Abstract: Exploration remains a major challenge for reinforcement learning (RL) agents in Craftax-style environments due to sparse rewards and long horizons. Large language models (LLMs) can guide exploration by generating high-level goals, but methods using models without explicit task adaptation often rely on carefully designed prompts and may produce inconsistent or contextually inappropriate goals. We introduce CCGP-2M, a multimodal goal-preference dataset constructed from interaction trajectories in Craftax-Classic, and propose a vision-language model (VLM)-guided RL framework that learns environment-specific goal guidance from this dataset. The framework consists of two stages: offline VLM fine-tuning and online goal-guided RL. In the offline stage, we align the VLM with goal preferences using Kullback-Leibler Token-level Direct Preference Optimization (KLTDPO), which combines token-level KL regularization with distributionally robust reweighting. Experiments on Craftax-Classic show that our method collects diamonds earlier than state-of-the-art baselines, achieves the highest score and reward, and attains the highest success rates on most achievements.

## Cross-Behavior Policy Optimization: Unified Learning from Offline Demonstrations and On- and Off-Policy Rollouts via a Shared Regression Target

DevFeed: [Cross-Behavior Policy Optimization: Unified Learning from Offline Demonstrations and On- and Off-Policy Rollouts via a Shared Regression Target](<https://devfeed.tech/articles/cross-behavior-policy-optimization-unified-learning-from-offline-demonstrations-and-on-and-off-policy-rollouts-via-a-shared-regression-target-144960.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.12830>)

Author: Du Cheng, Junyu Lu, Jingfan Zhang, Haoxuan Jia, Shichao Weng

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [post-training](<https://devfeed.tech/topics/post-training.md>), [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>)

Tags: [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [buffer](<https://devfeed.tech/tags/buffer.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [history](<https://devfeed.tech/tags/history.md>), [leading](<https://devfeed.tech/tags/leading.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [offline](<https://devfeed.tech/tags/offline.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [policy](<https://devfeed.tech/tags/policy.md>), [regression](<https://devfeed.tech/tags/regression.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [target](<https://devfeed.tech/tags/target.md>)

### AI overview

CBPO trains on offline demonstrations and current or historical policy rollouts with one regression loss using a fixed-reference target. Its reward-labeled replay buffer can include positive and negative offline completions and reference responses; the target does not depend on the generating policy, so the method does not require stored behavior-policy probabilities, importance weighting, or a learned critic. Source restrictions yield offline, on-policy, and off-policy variants, while CBPO-Unified combines them. Across two model scales and four mathematical reasoning benchmarks, the paper reports leading avg@32 and pass@32 results. On Qwen3-1.7B, CBPO-Unified exceeds SFT DAPO by 2.33 and 4.44 percentage points on benchmark-averaged avg@32 and pass@32; retaining historical rollouts adds 1.87 and 3.72 points over the no-history replay configuration.

### Source excerpt

arXiv:2610.12830v3 Announce Type: replace Abstract: Reasoning post-training draws on offline demonstrations and policy-generated rollouts, yet conventional supervised fine-tuning (SFT) followed by reinforcement learning (RL) separates their objectives and limits continued reuse of collected responses. We propose \textbf{Cross-Behavior Policy Optimization} (CBPO), which makes these sources compatible with one regression loss through a fixed-reference target. Because the generating policy does not enter this target, positive and negative offline completions, reference responses, and current and historical rollouts can share a common reward-labeled record format and replay buffer. As new responses enter the buffer, optimization retains the same regression objective without coordinating separate SFT and RL objectives or requiring stored behavior-policy probabilities, importance weighting, or a learned critic. Restricting the available sources gives offline, on-policy, and off-policy variants, while CBPO-Unified combines them. Across two model scales and four mathematical reasoning benchmarks, CBPO delivers leading \texttt{avg@32} and \texttt{pass@32} performance across benchmarks and data regimes. On Qwen3-1.7B, CBPO-Unified improves benchmark-averaged \texttt{avg@32} and \texttt{pass@32} over SFT DAPO by 2.33 and 4.44 percentage points, respectively. In the replay ablation, retaining historical rollouts improves these metrics by 1.87 and 3.72 percentage points over the no-history configuration.

## GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives

DevFeed: [GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives](<https://devfeed.tech/articles/groundingpi-a-grounding-foundation-model-towards-physical-intelligence-with-visual-primitives-144951.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.39601>)

Author: Qize Yu, Lianrui Fan, Boyu Chen, Jiaqi Liang, Xini Ding, Yue Chen, Zetian Song, Yuran Wang, Yi Zou, Kaixuan Wang, Tianxing Chen, Wenxuan Song, Bohan Zhou, Mingleyang Li, Siqiao Huang, Yuqi Ye, Caigao

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [grpo](<https://devfeed.tech/topics/grpo.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [scaling laws](<https://devfeed.tech/topics/scaling-laws.md>)

Tags: [autonomous-driving](<https://devfeed.tech/tags/autonomous-driving.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [foundation-model](<https://devfeed.tech/tags/foundation-model.md>), [grounding](<https://devfeed.tech/tags/grounding.md>), [grpo](<https://devfeed.tech/tags/grpo.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [ocr](<https://devfeed.tech/tags/ocr.md>), [pretraining](<https://devfeed.tech/tags/pretraining.md>), [robotic-manipulation](<https://devfeed.tech/tags/robotic-manipulation.md>)

### AI overview

GroundingPI is a 4B model that represents points and boxes as quantized coordinates in a shared vocabulary, trained with multimodal and spatial pretraining, supervised fine-tuning, and GRPO reinforcement learning. The abstract reports an average of 73.68% across 34 grounding benchmarks, ahead of the larger GPT-6 Astra at 71.54%. As a visual backbone, it outperformed evaluated alternatives in all four out-of-distribution settings across RoboTwin 2.0 and RoboCasa-GR1, with gains up to 24.8%; on RoboCasa-GR1, using 50% of demonstrations beat baselines using 75%. Its mean open-loop L2 error on nuScenes was 0.296 m. The authors report that greater pretraining scale improved grounding and manipulation, dense grounding benefited both, and OCR may aid perceptual learning.

### Source excerpt

arXiv:2609.39601v3 Announce Type: replace-cross Abstract: Precise grounding matters. It specifies which object is the target and where that object is, even in clutter and for tiny objects, and it has to be fast enough for closed-loop control. Yet vision-language-action (VLA) and world-action models (WAMs) take perception from general-purpose vision-language and video-generation backbones, which still fail in these settings. We introduce GroundingPI, a 4B grounding foundation model that generates points and boxes as quantized coordinates in a shared vocabulary. Training combines multimodal and spatial pretraining, supervised fine-tuning, and reinforcement learning with GRPO, using supervision from public datasets and dedicated data engines. Against 44 baselines across 34 grounding benchmarks spanning 11 perceptual capabilities, GroundingPI establishes a new state of the art, averaging 73.68%, above the larger GPT-6 Astra (71.54%). As a downstream visual backbone, GroundingPI improves performance on robotic manipulation and autonomous driving. Across RoboTwin 2.0 and RoboCasa-GR1, it outperforms all evaluated backbones in all four out-of-distribution settings, with relative gains of up to 24.8% over the strongest baseline. On RoboCasa-GR1, GroundingPI trained with 50% of the demonstrations outperforms those baselines trained with 75%. Its mean open-loop L2 error on nuScenes is 0.296 m. We systematically analyze GroundingPI's pretraining in scale and data composition. Grounding quality and downstream manipulation improve with pretraining scale, with generalization tracking perceptual gains most closely. Analyzing the data recipe across these 11 perceptual capabilities shows dense grounding's substantial benefits for both, and OCR's potential as a catalyst for perceptual learning. These results support grounding as a perceptual foundation, and dedicated perceptual pretraining as a promising direction for foundation models of physical intelligence.

[Next page](<https://devfeed.tech/topics/fine-tuning.md?cursor=WyIyMDI2LTExLTIwVDA1OjAwOjAwKzAwOjAwIiwgIjE3ZWYyODZhLTI0MjUtNDI5MC05ODQ1LWU0OGYxY2UwNzRkOCJd>)