# Enhancing AI model performance

Techniques and strategies for improving AI system performance, including prompt engineering, retrieval-augmented generation, and fine-tuning.

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

## 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.

## 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.

## Democratizing Cross-Project Knowledge for Transforming Water Governance and Climate Resilience: an Agentic AI Ecosystem

DevFeed: [Democratizing Cross-Project Knowledge for Transforming Water Governance and Climate Resilience: an Agentic AI Ecosystem](<https://devfeed.tech/articles/democratizing-cross-project-knowledge-for-transforming-water-governance-and-climate-resilience-an-agentic-ai-ecosystem-145666.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26398>)

Author: Oriol Al\\\`as, Ian Palac\\'in, Iv\\'an Cester, Edgar Rubi\\'on

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

Content type: article

Language: en

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

Topics: [information retrieval](<https://devfeed.tech/topics/information-retrieval.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Agentic AI](<https://devfeed.tech/topics/what-is-agentic-ai.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [knowledge-engineering](<https://devfeed.tech/topics/knowledge-engineering.md>), [knowledge-management](<https://devfeed.tech/topics/knowledge-management.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agentic-retrieval](<https://devfeed.tech/tags/agentic-retrieval.md>), [ai](<https://devfeed.tech/tags/ai.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [budget](<https://devfeed.tech/tags/budget.md>), [climate](<https://devfeed.tech/tags/climate.md>), [collections](<https://devfeed.tech/tags/collections.md>), [controller](<https://devfeed.tech/tags/controller.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [data](<https://devfeed.tech/tags/data.md>), [decision](<https://devfeed.tech/tags/decision.md>), [decision-making](<https://devfeed.tech/tags/decision-making.md>), [documentation](<https://devfeed.tech/tags/documentation.md>), [ecosystem](<https://devfeed.tech/tags/ecosystem.md>), [governance](<https://devfeed.tech/tags/governance.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [resilience](<https://devfeed.tech/tags/resilience.md>), [water](<https://devfeed.tech/tags/water.md>)

### AI overview

The paper presents an agentic retrieval system for finding evidence across EU-funded water and climate projects. It organizes documents from IMPETUS, NATALIE, and InnWater into hierarchical section trees with structured metadata, then uses separate water- and climate-domain embedding models. A ReAct-style controller allocates retrieval budget across domain collections and gathers evidence until a stopping criterion is met. In a benchmark on this corpus, the approach improved Mean Reciprocal Rank by 26.7% relative to a single-domain embedding baseline. The excerpt does not describe evaluation beyond that retrieval benchmark.

### Source excerpt

arXiv:2611.26398v1 Announce Type: cross Abstract: EU-funded projects produce extensive evidence in the form of deliverables, reports, and case studies. However, this knowledge is typically scattered across projects and document structures, making it difficult to locate relevant, contextual information when making decisions. This paper introduces an Agentic AI Ecosystem that consolidates heterogeneous project documentation into a domain-aware knowledge base and provides adaptive access through an agentic Retrieval-Augmented Generation pipeline. Documents from three EU-funded projects (IMPETUS, NATALIE, and InnWater) are reconstructed into hierarchical section trees and indexed with structured metadata. Retrieval is supported by two domain-specific embedding models fine-tuned on domain data (water and climate). On top of this knowledge base, a ReAct-style controller iteratively orchestrates retrieval actions by allocating retrieval budget across domain collections and accumulating evidence until a stopping criterion is met. In a retrieval benchmark over the corpus, the proposed approach improves Mean Reciprocal Rank by 26.7% (relative) compared to a single-domain embedding baseline, indicating more consistent retrieval of relevant evidence to support cross-project knowledge synthesis and climate-resilient decision-making.

## 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.

## SPARK: Adaptive Parametric Knowledge Participation for Faithful Retrieval-Augmented Generation

DevFeed: [SPARK: Adaptive Parametric Knowledge Participation for Faithful Retrieval-Augmented Generation](<https://devfeed.tech/articles/spark-adaptive-parametric-knowledge-participation-for-faithful-retrieval-augmented-generation-145353.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.27786>)

Author: Ruochang Li, Pengcheng Huang, Zhenghao Liu, Yukun Yan, Huiyuan Xie, Yu Gu, Ge Yu, Maosong Sun

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

Content type: article

Language: en

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

Topics: [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [peft](<https://devfeed.tech/topics/peft.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [faithfulness](<https://devfeed.tech/tags/faithfulness.md>), [generation](<https://devfeed.tech/tags/generation.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [large-language-models-llms](<https://devfeed.tech/tags/large-language-models-llms.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [retrieval-augmented-generation](<https://devfeed.tech/tags/retrieval-augmented-generation.md>), [spark](<https://devfeed.tech/tags/spark.md>)

### AI overview

SPARK addresses conflicts between retrieved evidence and an LLM's internal knowledge by learning a lightweight policy that dynamically adjusts the contribution of feed-forward-network-mediated parametric knowledge during generation. It keeps the backbone model frozen and optimizes the policy with a faithfulness-oriented reward, avoiding direct edits to the model's stored knowledge. The abstract reports experiments on six datasets, with 0.01% additional trainable parameters, and says SPARK achieved more faithful generation than competitive baselines while preserving the backbone's general capabilities.

### Source excerpt

arXiv:2606.27786v3 Announce Type: replace-cross Abstract: Retrieval-augmented generation (RAG) enhances the faithfulness of large language models (LLMs) by incorporating external knowledge into the generation process. However, conflicts between retrieved context and parametric knowledge can cause LLMs to rely on inconsistent internal knowledge, hindering their ability to faithfully leverage retrieved evidence. Existing approaches mitigate this issue by editing parametric knowledge to resolve such conflicts. However, locating and modifying knowledge encoded in model parameters is challenging and may inadvertently compromise the backbone LLM's capabilities. In this paper, we introduce SPARK, a novel framework that formulates parametric knowledge control as a latent participation policy learning problem, enabling LLMs to adaptively regulate the contribution of parametric knowledge during generation. Specifically, SPARK introduces a lightweight participation policy that dynamically modulates the contribution of FFN-mediated parametric knowledge while keeping the backbone LLM frozen. The participation policy is optimized with a faithfulness-oriented reward, enabling fine-grained control over parametric knowledge participation during generation. Extensive experiments on six datasets validate the effectiveness of SPARK, which introduces and optimizes only 0.01\% additional trainable parameters while keeping the backbone LLM fully frozen, thereby preserving its general capabilities and achieving more faithful generation than competitive baselines. All datasets and code are available at https://github.com/OpenBMB/SPARK.

## ALIoTh: Orchestrating Multi-Agent LLMs for Autonomous IoT Penetration Testing

DevFeed: [ALIoTh: Orchestrating Multi-Agent LLMs for Autonomous IoT Penetration Testing](<https://devfeed.tech/articles/alioth-orchestrating-multi-agent-llms-for-autonomous-iot-penetration-testing-146047.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25809>)

Author: Pironti Francesco Aurelio, Arena Luigi, Blefari Francesco, Lupinacci Matteo, Romeo Francesco, Maresca Alessio

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: [Hacking](<https://devfeed.tech/topics/hacking.md>), [Internet of things](<https://devfeed.tech/topics/iot.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [smartdevices](<https://devfeed.tech/topics/smartdevices.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [assessment](<https://devfeed.tech/tags/assessment.md>), [attack](<https://devfeed.tech/tags/attack.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [cameras](<https://devfeed.tech/tags/cameras.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [devices](<https://devfeed.tech/tags/devices.md>), [enumeration](<https://devfeed.tech/tags/enumeration.md>), [exploit](<https://devfeed.tech/tags/exploit.md>), [exploitation](<https://devfeed.tech/tags/exploitation.md>), [framework](<https://devfeed.tech/tags/framework.md>), [gathering](<https://devfeed.tech/tags/gathering.md>), [generation](<https://devfeed.tech/tags/generation.md>), [group-policy](<https://devfeed.tech/tags/group-policy.md>), [iiot](<https://devfeed.tech/tags/iiot.md>), [iot](<https://devfeed.tech/tags/iot.md>), [llms](<https://devfeed.tech/tags/llms.md>), [multi-agent](<https://devfeed.tech/tags/multi-agent.md>), [penetration-testing](<https://devfeed.tech/tags/penetration-testing.md>)

### AI overview

ALIoTh is a workflow-driven framework that coordinates specialized LLM agents through LangGraph and a Self-RAG architecture to automate IoT and IIoT penetration testing. Its described pipeline covers network, device, and service enumeration, vulnerability assessment, payload generation, reverse-shell deployment, and privilege escalation. Self-RAG gives agents contextual access to IoT security documentation, tool manuals, and known attack techniques, supporting scenarios involving multiple devices and group-policy manipulation. The supplied text describes the approach but provides no evaluation results or comparative performance.

### Source excerpt

arXiv:2611.25809v1 Announce Type: new Abstract: Penetration testing in Internet of Things (IoT) environments, especially those composed of various different devices, e.g., smart plugs, smart cameras, and smart doors, presents great challenges across the phases of information gathering, attack surface enumeration, privilege escalation, and lateral movement. This paper introduces ALIoTh, a multi-agent, workflow-driven penetration testing framework that integrates Large Language Models agents within a Self-Retrieval-Augmented Generation (Self-RAG) architecture to autonomously identify and exploit vulnerabilities in IoT and Industrial-IoT (IIoT) environments. By leveraging state-of-the-art language models and orchestrating specialized agents via the LangGraph framework, ALIoTh dynamically plans and executes a full-spectrum offensive pipeline. This includes network, devices, and service enumeration, vulnerability assessment, payload generation, reverse shell deployment, and privilege escalation on each device. The Self-RAG mechanism equips expert agents with contextual access to IoT security knowledge, e.g., related documentation, tool manuals, and known attack techniques, enhancing their reasoning capabilities in scenarios involving multiple device exploitation and group policy manipulation.

## Rate-distortion Assessment for Hierarchical Content Summarization

DevFeed: [Rate-distortion Assessment for Hierarchical Content Summarization](<https://devfeed.tech/articles/rate-distortion-assessment-for-hierarchical-content-summarization-145014.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25151>)

Author: Yunye Gong, Indranil Sur, Jared Claypoole, Pritish Sahu, Ajay Divakaran

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 evaluation](<https://devfeed.tech/topics/ai-evaluation.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>)

Tags: [assessment](<https://devfeed.tech/tags/assessment.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [budget](<https://devfeed.tech/tags/budget.md>), [compression](<https://devfeed.tech/tags/compression.md>), [content](<https://devfeed.tech/tags/content.md>), [context](<https://devfeed.tech/tags/context.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [design](<https://devfeed.tech/tags/design.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [encoding](<https://devfeed.tech/tags/encoding.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [framework](<https://devfeed.tech/tags/framework.md>), [generation](<https://devfeed.tech/tags/generation.md>), [generative](<https://devfeed.tech/tags/generative.md>), [generative-models](<https://devfeed.tech/tags/generative-models.md>), [guidance](<https://devfeed.tech/tags/guidance.md>), [hierarchy](<https://devfeed.tech/tags/hierarchy.md>), [information](<https://devfeed.tech/tags/information.md>), [insights](<https://devfeed.tech/tags/insights.md>), [language](<https://devfeed.tech/tags/language.md>), [language-models](<https://devfeed.tech/tags/language-models.md>)

### AI overview

The paper frames hierarchical summarization as information encoding and proposes extracting a source's semantic hierarchy to measure the trade-off between compression rate and information distortion. This enables model comparisons at a chosen rate budget. The authors report experiments on long-context question-answering summaries using state-of-the-art large language models, aimed at assessing efficiency and comprehension, and say that using the hierarchy to guide retrieval-augmented generation improves rate-distortion characteristics over current baselines. The supplied text gives no quantitative results or details of the baselines.

### Source excerpt

arXiv:2611.25151v1 Announce Type: new Abstract: Powerful generative models including large language and vision models have demonstrated success across a wide range of content generation, editing and manipulation tasks via end-to-end prompt-based generation. However, current assessment of the quality of generative content manipulation remains subjective, limiting scientific understanding of model capabilities. Inspired by rate-distortion theory from Shannon information compression theory, in this work we formalize hierarchical content summarization as an information encoding problem. We design a principled framework to extract the semantic hierarchy of the source content to support quantitative characterization of the rate-distortion trade-off between compression efficiency (rate) against information loss (distortion) during generative content manipulation, which allows us to benchmark various models with respect to performance conditional on rate budget of interest. We demonstrate the use of the framework with empirical experiments on content summarization for long-context question-answering with state-of-the-art large language models, offering new insights into their efficiency and comprehension capabilities. We further demonstrate the use of the semantic hierarchy as guidance for performing effective retrieval augmented generation, showing improved rate-distortion characteristics against current baselines.

## Meta FAIR Self-Improving Memory Hits 56.0 on Continual Learning Bench

DevFeed: [Meta FAIR Self-Improving Memory Hits 56.0 on Continual Learning Bench](<https://devfeed.tech/articles/meta-fair-self-improving-memory-hits-56-0-on-continual-learning-bench-144512.md>)

Original publisher: [Read original article](<https://www.marktechpost.com/2026/11/19/meta-fair-self-improving-memory-harness-distillation/>)

Author: Michal Sutter

Published: 2026-11-19T21:46:22Z

Content type: article

Language: en

Sources: [MarkTechPost](<https://devfeed.tech/sources/marktechpost.md>)

Topics: [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-memory](<https://devfeed.tech/tags/agent-memory.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-infrastructure](<https://devfeed.tech/tags/ai-infrastructure.md>), [ai-shorts](<https://devfeed.tech/tags/ai-shorts.md>), [applications](<https://devfeed.tech/tags/applications.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [bench](<https://devfeed.tech/tags/bench.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [continual-learning](<https://devfeed.tech/tags/continual-learning.md>), [database](<https://devfeed.tech/tags/database.md>), [distillation](<https://devfeed.tech/tags/distillation.md>), [editors-pick](<https://devfeed.tech/tags/editors-pick.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [failed](<https://devfeed.tech/tags/failed.md>), [files](<https://devfeed.tech/tags/files.md>), [found](<https://devfeed.tech/tags/found.md>), [generation](<https://devfeed.tech/tags/generation.md>), [generic](<https://devfeed.tech/tags/generic.md>), [harness](<https://devfeed.tech/tags/harness.md>), [kimi-k2-6](<https://devfeed.tech/tags/kimi-k2-6.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [learning](<https://devfeed.tech/tags/learning.md>), [memory](<https://devfeed.tech/tags/memory.md>), [meta](<https://devfeed.tech/tags/meta.md>), [staff](<https://devfeed.tech/tags/staff.md>), [tech-news](<https://devfeed.tech/tags/tech-news.md>), [technology](<https://devfeed.tech/tags/technology.md>), [uncategorized](<https://devfeed.tech/tags/uncategorized.md>)

### AI overview

Meta FAIR's Self-Improving Memory lets Kimi-K2.6 discover LOAD and OFFLOAD prompts for file-based memory, then distills the resulting behaviors into model weights. On three Continual Learning Bench tasks, optimized prompts average 56.0 versus 44.6 for the best hand-written harness; Codebase scores 64.3, 20.6 points higher than its best hand-written result. LoRA fine-tuning on demonstrations made under the optimized prompts, with those prompts replaced by generic ones in training logs, yields 43.1 versus 37.6 for SFT on generic trajectories. Search is noisy: 37 of 44 prompt revisions failed to beat the prior best. The reported tests measure generalization to new instances of the same task types, and benchmark results are not directly comparable to other methods because their benchmarks differ.

### Source excerpt

Meta FAIR's Self-Improving Memory lets Kimi-K2.6 write its own LOAD and OFFLOAD memory prompts, then distills them into its weights. Optimized prompts score 56.0 versus 44.6 for the best hand-written harness on Continual Learning Bench. The post Meta FAIR Self-Improving Memory Hits 56.0 on Continual Learning Bench appeared first on MarkTechPost.

## SWE-growth: Teacher-Guided Adaptive Self-Play for Software Engineering Agents

DevFeed: [SWE-growth: Teacher-Guided Adaptive Self-Play for Software Engineering Agents](<https://devfeed.tech/articles/swe-growth-teacher-guided-adaptive-self-play-for-software-engineering-agents-142200.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23876>)

Author: Songshi Liang, Zixuan Huang, Mingxuan Song, Haoyu Tang, Shusheng Xu, Tongkai Yang, Zihe Wang, Rui Yan

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Software Engineering](<https://devfeed.tech/topics/software-engineering.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [agents](<https://devfeed.tech/tags/agents.md>), [capability](<https://devfeed.tech/tags/capability.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [growth](<https://devfeed.tech/tags/growth.md>), [play](<https://devfeed.tech/tags/play.md>), [qwen3](<https://devfeed.tech/tags/qwen3.md>), [qwen3-coder](<https://devfeed.tech/tags/qwen3-coder.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [repair](<https://devfeed.tech/tags/repair.md>), [repository](<https://devfeed.tech/tags/repository.md>), [selection](<https://devfeed.tech/tags/selection.md>), [sft](<https://devfeed.tech/tags/sft.md>), [software](<https://devfeed.tech/tags/software.md>), [software-engineering](<https://devfeed.tech/tags/software-engineering.md>), [supervised-fine-tuning](<https://devfeed.tech/tags/supervised-fine-tuning.md>), [swe-bench-pro](<https://devfeed.tech/tags/swe-bench-pro.md>), [tasks](<https://devfeed.tech/tags/tasks.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

SWE-Growth uses verified Teacher repairs to turn Student failures into training opportunities. Tasks the Student consistently fails but the Teacher reliably solves supply demonstrations for supervised fine-tuning; after fine-tuning, mixed-outcome tasks with a positive Teacher-Student success gap are selected for reinforcement learning. Task eligibility also guides updates to the task-generating Injector. Starting from 101 seed issues, the framework raised Qwen3-Coder-30B-A3B-Instruct from 27.7% to 31.7% on SWE-bench Pro and from 11.0% to 17.2% on held-out TerminalBench 2.1 by Round 2. Under approximately matched budgets, gains exceeded Student-only self-play by 2.6 and 4.1 percentage points, and Teacher-augmented self-play by 1.8 and 2.5 points, respectively.

### Source excerpt

arXiv:2611.23876v1 Announce Type: new Abstract: Self-play offers a scalable way to train software engineering agents by generating repository-level repair tasks. However, with binary repair rewards and finite rollout budgets, all-fail groups provide no reward-based group-relative advantage. Such tasks may still offer learning opportunities when a Teacher can provide verified repairs absent from the Student's own attempts. We introduce SWE-Growth, a teacher-guided adaptive self-play framework that uses verified Teacher outcomes to identify learning opportunities for the current Student. Tasks with all students failure but reliable Teacher success provide demonstrations for supervised fine-tuning (SFT), and post-SFT Student rollouts identify mixed-outcome tasks with a positive Teacher--Student gap for reinforcement learning. Task eligibility also guides Injector updates, connecting Student training with subsequent task generation. Starting from 101 seed issues, \textsc{SWE-Growth} improves Qwen3-Coder-30B-A3B-Instruct from 27.7% to 31.7% on SWE-bench Pro and from 11.0% to 17.2% on held-out TerminalBench 2.1 by Round~2. It exceeds Student-only self-play by 2.6 and 4.1 percentage points, respectively, and Teacher-augmented self-play by 1.8 and 2.5 points under approximately matched budgets. These results support adapting both supervision and task selection to the Student's evolving repair capability.

## Improving Cross-Lingual Token Representations by Adding a Pinch of SALT

DevFeed: [Improving Cross-Lingual Token Representations by Adding a Pinch of SALT](<https://devfeed.tech/articles/improving-cross-lingual-token-representations-by-adding-a-pinch-of-salt-142559.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.09953>)

Author: Guillem Ram\\'irez

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Internationalization (i18n)](<https://devfeed.tech/topics/i18n.md>)

Tags: [alignment](<https://devfeed.tech/tags/alignment.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [detection](<https://devfeed.tech/tags/detection.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [hallucination-detection](<https://devfeed.tech/tags/hallucination-detection.md>), [languages](<https://devfeed.tech/tags/languages.md>), [learning](<https://devfeed.tech/tags/learning.md>), [lightweight](<https://devfeed.tech/tags/lightweight.md>), [mining](<https://devfeed.tech/tags/mining.md>), [performance](<https://devfeed.tech/tags/performance.md>), [post-training](<https://devfeed.tech/tags/post-training.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [salt](<https://devfeed.tech/tags/salt.md>), [sequence](<https://devfeed.tech/tags/sequence.md>), [signal](<https://devfeed.tech/tags/signal.md>), [span](<https://devfeed.tech/tags/span.md>), [supervision](<https://devfeed.tech/tags/supervision.md>), [tasks](<https://devfeed.tech/tags/tasks.md>), [token](<https://devfeed.tech/tags/token.md>), [translation](<https://devfeed.tech/tags/translation.md>)

### AI overview

SALT improves token representations in cross-lingual sentence encoders by post-training them with span-level supervision. Across five multilingual token-level benchmarks, it achieves the best overall results on four, outperforming alternative fine-tuning strategies and competitive encoders. The abstract also reports gains on cross-lingual retrieval and classification, suggesting the method can benefit sentence-level performance too.

### Source excerpt

arXiv:2609.09953v2 Announce Type: replace-cross Abstract: Cross-lingual sentence encoders enable scalable transfer across hundreds of languages, powering applications such as translation mining and zero-shot learning in low-resource settings. Although trained for sentence-level alignment, they are increasingly also applied to token-level tasks such as hallucination detection and sequence tagging, exposing a mismatch between training and usage. We propose SALT, a lightweight post-training method that improves token representations by injecting span-level supervision into existing sentence encoders. Across five multilingual token-level benchmarks, SALT achieves the best overall results on four of them, outperforming alternative fine-tuning strategies and competitive encoders. It also improves sentence-level performance on cross-lingual retrieval and classification tasks. These results demonstrate that span-level supervision is an effective signal for improving both token and sentence representations.

## TOGBench: Benchmarking and Improving Tiny-Object Generation Fidelity in Aerial Imagery

DevFeed: [TOGBench: Benchmarking and Improving Tiny-Object Generation Fidelity in Aerial Imagery](<https://devfeed.tech/articles/togbench-benchmarking-and-improving-tiny-object-generation-fidelity-in-aerial-imagery-142708.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23840>)

Author: Chanhwi Jeong, Dongmin You, Seungwook Kim

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Image](<https://devfeed.tech/topics/image.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [generative graphics](<https://devfeed.tech/topics/generative-graphics.md>)

Tags: [aerial-imagery](<https://devfeed.tech/tags/aerial-imagery.md>), [aircraft](<https://devfeed.tech/tags/aircraft.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [generation](<https://devfeed.tech/tags/generation.md>), [generative](<https://devfeed.tech/tags/generative.md>), [inference](<https://devfeed.tech/tags/inference.md>), [layout](<https://devfeed.tech/tags/layout.md>), [techniques](<https://devfeed.tech/tags/techniques.md>), [tiny](<https://devfeed.tech/tags/tiny.md>), [vehicles](<https://devfeed.tech/tags/vehicles.md>)

### AI overview

Aerial image generators often render tiny vehicles, ships, and aircraft less faithfully than larger objects, while common evaluations focus on prompt following or human preference. TOGBench addresses this gap with a quantitative, object-level benchmark for tiny-object fidelity in aerial imagery. The authors also propose two train-time or inference-time techniques for layout-to-image generators and report that they improve performance on TOGBench. The supplied excerpt does not describe the benchmark's metrics or the techniques' implementation.

### Source excerpt

arXiv:2611.23840v1 Announce Type: cross Abstract: Advances in deep generative models now enable the synthesis of realistic imagery. However, aerial imagery is dominated by tiny objects, such as vehicles, ships, and aircraft, which occupy only a few pixels owing to the overhead, wide-area nature of aerial imaging. Existing generative models struggle substantially more with tiny objects than with larger ones. This limitation has been overlooked, as prior work on generative evaluation emphasizes prompt following or human preference rather than whether small objects are rendered faithfully. To address this gap, we introduce TOGBench, a quantitative, object-level benchmark for evaluating tiny-object generation fidelity in aerial imagery. We further propose two simple train-/inference-time techniques that improve the ability of layout-to-image generators to render small objects, raising their performance on TOGBench.

## NaviGATE: Noise-Aware Dual-View Generation for Multimodal Knowledge Graph Reasoning

DevFeed: [NaviGATE: Noise-Aware Dual-View Generation for Multimodal Knowledge Graph Reasoning](<https://devfeed.tech/articles/navigate-noise-aware-dual-view-generation-for-multimodal-knowledge-graph-reasoning-142808.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24552>)

Author: Jie Peng, Yongxue Shan, Yongfu Zha, Cui Miao, Xiaodong Wang

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>)

Tags: [activation](<https://devfeed.tech/tags/activation.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [generation](<https://devfeed.tech/tags/generation.md>), [generative](<https://devfeed.tech/tags/generative.md>), [graph](<https://devfeed.tech/tags/graph.md>), [inference](<https://devfeed.tech/tags/inference.md>), [instruction-tuning](<https://devfeed.tech/tags/instruction-tuning.md>), [knowledge-graph](<https://devfeed.tech/tags/knowledge-graph.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [noise](<https://devfeed.tech/tags/noise.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [robust](<https://devfeed.tech/tags/robust.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [semantics](<https://devfeed.tech/tags/semantics.md>), [stage](<https://devfeed.tech/tags/stage.md>), [view](<https://devfeed.tech/tags/view.md>)

### AI overview

NaviGATE addresses noisy retrieval and invalid entity generation in multimodal knowledge graph reasoning. It aligns graph structure with text, fine-tunes a multimodal language model using entity summaries and structural tokens, then trains retrieval-free and retrieval-aware views together. Contrastive logit sharpening reinforces reliable retrieved candidates; parametric consistency distillation suppresses misleading ones. At inference, structure-constrained decoding limits outputs to valid graph entities. The abstract reports that NaviGATE outperforms strong unimodal, multimodal, and generation-based baselines on standard benchmarks and remains robust under noisy retrieval.

### Source excerpt

arXiv:2611.24552v1 Announce Type: new Abstract: Multimodal knowledge graph reasoning (MKGR) aims to infer missing facts by jointly exploiting relational structures, textual descriptions, and visual evidence. Existing embedding-based methods effectively capture graph topology, but their fixed multimodal representations are less effective for query-specific fine-grained semantic reasoning, while generation-based methods are vulnerable to noisy retrieval and may produce invalid entities. To address these challenges, we propose NaviGATE, a noise-aware generative framework for MKGR. NaviGATE first aligns graph-derived structural representations with textual semantics, then performs structure-aware internal knowledge activation by fine-tuning the multimodal language model with multimodal entity summaries and injected structural tokens. In the noise-aware dual-view instruction tuning stage, NaviGATE jointly optimizes retrieval-free and retrieval-aware views: reliable retrieved candidates are reinforced through contrastive logit sharpening, while misleading candidates are suppressed through parametric consistency distillation. In inference, structure-constrained decoding restricts generation to valid entities in the target knowledge graph. Experiments on standard benchmarks show that NaviGATE consistently outperforms strong unimodal, multimodal, and generation-based baselines, and remains robust under noisy retrieval scenarios.

## Pose6DAug: Physically Plausible Multi-view Object Swapping for Robot Data Augmentation

DevFeed: [Pose6DAug: Physically Plausible Multi-view Object Swapping for Robot Data Augmentation](<https://devfeed.tech/articles/pose6daug-physically-plausible-multi-view-object-swapping-for-robot-data-augmentation-143277.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.20118>)

Author: Jonghoon Lee, Seong Hyeon Park, Byungwoo Jeon, Minha Lee, Jinwoo Shin

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [data augmentation](<https://devfeed.tech/topics/data-augmentation.md>), [3D](<https://devfeed.tech/topics/3d.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [2d](<https://devfeed.tech/tags/2d.md>), [3d](<https://devfeed.tech/tags/3d.md>), [appearance](<https://devfeed.tech/tags/appearance.md>), [camera](<https://devfeed.tech/tags/camera.md>), [coherent](<https://devfeed.tech/tags/coherent.md>), [consistency](<https://devfeed.tech/tags/consistency.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>), [data-augmentation](<https://devfeed.tech/tags/data-augmentation.md>), [data-collection](<https://devfeed.tech/tags/data-collection.md>), [distribution](<https://devfeed.tech/tags/distribution.md>), [failure](<https://devfeed.tech/tags/failure.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [framework](<https://devfeed.tech/tags/framework.md>), [manipulation](<https://devfeed.tech/tags/manipulation.md>), [mesh](<https://devfeed.tech/tags/mesh.md>), [object](<https://devfeed.tech/tags/object.md>), [performance](<https://devfeed.tech/tags/performance.md>), [plausible](<https://devfeed.tech/tags/plausible.md>), [policies](<https://devfeed.tech/tags/policies.md>), [robot](<https://devfeed.tech/tags/robot.md>)

### AI overview

Pose6DAug creates targeted demonstrations for VLA policy failure cases by swapping the manipulated object in existing successful episodes while preserving their action trajectories. It anchors the replacement object with a mesh and temporally coherent 6D pose trajectory in 3D, enabling geometrically consistent renderings across calibrated camera views. This addresses multi-view inconsistency and implausible edits that can arise from naive 2D video editing, especially with heavy occlusion and egocentric views. The abstract reports a 16.5% success-rate improvement over the state-of-the-art baseline on novel objects while preserving in-distribution performance; on a real robot, success more than doubled relative to the base policy. These are abstract-level claims; the supplied text gives no further experimental details.

### Source excerpt

arXiv:2606.20118v3 Announce Type: replace-cross Abstract: Vision-language-action (VLA) policies have shown strong potential for general-purpose manipulation, yet they often fail on novel, out-of-distribution objects whose appearance or geometry deviates from the training distribution. The standard remedy is to collect multi-view teleoperation data for every failure case, but this scales poorly in both cost and time. We introduce Pose6DAug, a failure-driven data augmentation framework that turns existing successful episodes into targeted demonstrations for a policy's failure modes, without any new data collection. Our key insight is that each successful episode already encodes a physically valid action trajectory together with calibrated multi-view observations. By swapping only the manipulated object while preserving this trajectory, we obtain new and physically grounded demonstrations. However, the challenge lies in synthesizing object swaps, e.g., naive 2D video editing breaks multi-view consistency and physical plausibility, particularly under heavy occlusion and egocentric viewpoints. Our method instead operates directly in 3D, anchoring the target object with an explicit mesh driven by a temporally coherent 6D pose trajectory. This ensures geometrically consistent renderings across all camera views. Fine-tuning a VLA on data augmented by our method improves success rates by 16.5% relative to the state-of-the-art baseline on novel objects, while preserving in-distribution performance. On a real robot, it more than doubles the success rate of the base policy. These results show that multi-view and physically consistent augmentation is a practical path to scalable VLA generalization.

## Leveraging Segmentation Representations for Efficient Classification: A Self-Configuring Framework Based on nnU-Net

DevFeed: [Leveraging Segmentation Representations for Efficient Classification: A Self-Configuring Framework Based on nnU-Net](<https://devfeed.tech/articles/leveraging-segmentation-representations-for-efficient-classification-a-self-configuring-framework-based-on-nnu-net-140021.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22380>)

Author: Renjun Cai, Antonio Javier Gallego, F\\'elix Fuentes-Hurtado

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

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

Topics: [Medical imaging](<https://devfeed.tech/topics/medical-imaging.md>), [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [classification](<https://devfeed.tech/tags/classification.md>), [classifier](<https://devfeed.tech/tags/classifier.md>), [confidence](<https://devfeed.tech/tags/confidence.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [features](<https://devfeed.tech/tags/features.md>), [flare](<https://devfeed.tech/tags/flare.md>), [framework](<https://devfeed.tech/tags/framework.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [gpu-memory](<https://devfeed.tech/tags/gpu-memory.md>), [hidden](<https://devfeed.tech/tags/hidden.md>), [image](<https://devfeed.tech/tags/image.md>), [imaging](<https://devfeed.tech/tags/imaging.md>), [inference](<https://devfeed.tech/tags/inference.md>), [information](<https://devfeed.tech/tags/information.md>), [leaderboard](<https://devfeed.tech/tags/leaderboard.md>), [lightweight](<https://devfeed.tech/tags/lightweight.md>), [medical](<https://devfeed.tech/tags/medical.md>), [medical-imaging](<https://devfeed.tech/tags/medical-imaging.md>)

### AI overview

The framework extends nnU-Net to image-level classification by reusing a segmentation model's encoder features and predicted probabilities with a lightweight classifier. It tests two strategies: a frozen backbone with features augmented by region-of-interest information, predicted foreground volume, and segmentation confidence; and partial fine-tuning of the final encoder stage with an auxiliary segmentation loss. Five-fold experiments found that more complex configurations did not consistently improve classification. The selected framework scored 0.8056 on the public validation leaderboard and ranked first in the FLARE 2026 Task 2 final evaluation, with the highest DSC, NSD, and AUROC across four tasks, including two hidden tasks. Local inference stayed within the stated limits of 60 seconds per case and 4 GB of GPU memory.

### Source excerpt

arXiv:2611.22380v1 Announce Type: cross Abstract: We propose a self-configuring framework that extends the nnU-Net workflow to both segmentation and classification in 3D medical imaging. Building on nnU-Net's self-configuring method, the framework reuses representations learned by a segmentation model for image-level classification. Its encoder features and predicted segmentation probabilities are used by a lightweight classifier. Two information-sharing strategies are investigated. In the frozen-backbone branch, features are augmented with region-of-interest information, predicted foreground volume, and segmentation confidence. In the adaptation branch, the final encoder stage is partially fine-tuned for classification, with an auxiliary segmentation loss to preserve segmentation-related information. Additional five-fold experiments showed that more complex configurations did not consistently improve classification performance. On the public validation leaderboard, our selected framework achieved an Overall score of 0.8056. In the official FLARE 2026 Task 2 final evaluation, our framework ranked first overall, achieving the highest DSC, NSD, and AUROC across all four testing tasks, including two hidden testing tasks. Local inference measurements also remained below the challenge limits of $60 \mathrm{s}$ per case and $4 \mathrm{GB}$ GPU memory. These results demonstrate the practical value of reusing segmentation representations for classification while maintaining a simple and efficient framework.

## Simplicity Goes Far: Auditing Prompt Optimizers with Demonstration-Conditioned Prompt Search

DevFeed: [Simplicity Goes Far: Auditing Prompt Optimizers with Demonstration-Conditioned Prompt Search](<https://devfeed.tech/articles/simplicity-goes-far-auditing-prompt-optimizers-with-demonstration-conditioned-prompt-search-140056.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22605>)

Author: Bohe Chen

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

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

Topics: [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [prompt](<https://devfeed.tech/topics/prompt.md>), [DSPy](<https://devfeed.tech/topics/dspy.md>), [audit](<https://devfeed.tech/topics/audit.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [auditing](<https://devfeed.tech/tags/auditing.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [classification](<https://devfeed.tech/tags/classification.md>), [classification-accuracy](<https://devfeed.tech/tags/classification-accuracy.md>), [control](<https://devfeed.tech/tags/control.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [dspy](<https://devfeed.tech/tags/dspy.md>), [generate](<https://devfeed.tech/tags/generate.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [reflection](<https://devfeed.tech/tags/reflection.md>), [sampling](<https://devfeed.tech/tags/sampling.md>), [search](<https://devfeed.tech/tags/search.md>), [select](<https://devfeed.tech/tags/select.md>), [selection](<https://devfeed.tech/tags/selection.md>), [structured](<https://devfeed.tech/tags/structured.md>), [systems](<https://devfeed.tech/tags/systems.md>), [token-usage](<https://devfeed.tech/tags/token-usage.md>)

### AI overview

Demonstration-Conditioned Prompt Search (DCPS) isolates prompt generation and candidate selection from heavier optimizer mechanisms. It freezes the generator, conditions proposals on randomly sampled demonstrations, scores them on validation data, and retains the best candidate, without learned updates, reflection, or evolution. In StablePrompt, DCPS matches PPO classification accuracy within 0.5 percentage points at roughly one-third the measured cost. In DSPy-based systems on GPT-4.1-mini, it matches GEPA's macro accuracy under the reported budgets with 55% fewer rollouts and 32% lower average measured API cost. Results vary by task: proposal and selection recover much of heavier methods' performance on structured and instruction-following tasks, while added mechanisms retain advantages on some open-ended and reasoning-heavy tasks. The authors recommend comparing optimizers with generate-and-select controls and reporting rollouts, token use, and dollar costs alongside accuracy.

### Source excerpt

arXiv:2611.22605v1 Announce Type: new Abstract: Automatic prompt optimization is growing more elaborate. Recent methods stack policy-gradient updates, trace reflection, and evolutionary search. These methods share a generate-and-select step; without controlling for it, reported gains may conflate the contributions of optimizer mechanisms and larger sampling budgets. We isolate this step as Demonstration-Conditioned Prompt Search (DCPS), repurposing the Automatic Prompt Engineer (APE) recipe as a mechanism-attribution control: it freezes the generator, conditions it on randomly sampled demonstrations, scores candidates on validation data, and keeps the best---no learned update, reflection, or evolution. Across two optimizer families, DCPS replaces Proximal Policy Optimization (PPO) in StablePrompt, matching classification accuracy within 0.5 pp at roughly one-third the measured cost; in DSPy-based compound AI systems, under the reported budgets, it matches GEPA's macro accuracy on GPT-4.1-mini while using 55% fewer rollouts at 32% lower measured API cost on average. The audit reveals a task-dependent pattern: proposal and selection recover much of heavier optimizers' performance on structured and instruction-following tasks, while additional mechanisms retain advantages in some open-ended and reasoning-heavy settings. We recommend benchmarking prompt optimizers against generate-and-select controls and reporting rollout counts, token usage, and dollar costs alongside accuracy. Code and artifacts: https://github.com/bhc6/DCPS.

## Efficient Wide-Beam Generative Retrieval with Model-Native Trie Decoding and Shared-Context Attention

DevFeed: [Efficient Wide-Beam Generative Retrieval with Model-Native Trie Decoding and Shared-Context Attention](<https://devfeed.tech/articles/efficient-wide-beam-generative-retrieval-with-model-native-trie-decoding-and-shared-context-attention-140011.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22309>)

Author: Vedant Jhaveri, Arashpreet Singh Mor, Satyam Kumar, Alex Tsun, Qing Lan, Shaunak Mishra, Sriram Vasudevan

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

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

Topics: [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [vllm](<https://devfeed.tech/topics/vllm.md>), [Transformer](<https://devfeed.tech/topics/transformer.md>), [Cloud](<https://devfeed.tech/topics/cloud.md>)

Tags: [attention](<https://devfeed.tech/tags/attention.md>), [beam](<https://devfeed.tech/tags/beam.md>), [context](<https://devfeed.tech/tags/context.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [decoding](<https://devfeed.tech/tags/decoding.md>), [generative](<https://devfeed.tech/tags/generative.md>), [model](<https://devfeed.tech/tags/model.md>), [native](<https://devfeed.tech/tags/native.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [trie](<https://devfeed.tech/tags/trie.md>), [wide](<https://devfeed.tech/tags/wide.md>)

### AI overview

Trie-Head reduces two serving costs in wide-beam generative retrieval: it stores the identifier trie as validated GPU tensors, scores compact codebook candidates, and streams the full-vocabulary normalizer, avoiding Python constraint callbacks and vocabulary-wide candidate buffers. The authors say it preserves valid-token scores in real arithmetic without retraining. Shared-context attention reuses prompt key/value blocks across beams while keeping decode tails private. Both mechanisms are implemented in SGLang and evaluated for candidate sourcing in LinkedIn Hiring Assistant and qualification retrieval on the Qualification Retrieval Platform. At beam width 200 on the latter workload, the combined system reports up to 5.57x closed-loop throughput over a Python custom logit processor baseline and up to 2.63x over BF16 FlashRec.

### Source excerpt

arXiv:2611.22309v1 Announce Type: new Abstract: Generative retrieval (GR) predicts catalog entries as short semantic identifiers (SIDs). Trie-constrained beam search keeps these identifiers valid, but wide beams expose two serving costs: vocabulary-wide score tensors that are mostly discarded, and repeated attention over the same prompt. We present Trie-Head, a model-native output interface that stores the identifier trie as validated GPU tensors. Compared with a Python custom logit processor (CLP), it reduces CPU--GPU coordination and memory traffic by avoiding Python constraint callbacks and vocabulary-wide candidate buffers. It scores compact codebook candidates and streams the full-vocabulary normalizer, preserving the original model's valid-token scores in real arithmetic without retraining. Shared-context attention complements this output path by reusing prompt key/value blocks across beams while preserving their private decode tails. We implement both mechanisms in SGLang and evaluate candidate sourcing for LinkedIn Hiring Assistant and qualification retrieval for the Qualification Retrieval Platform. The combined system improves closed-loop throughput by up to 5.57x over the CLP baseline and up to 2.63x over BF16 FlashRec at a common beam width of 200 on the Qualification Retrieval Platform workload.

## Harnesses That Improve Themselves: A Unifying Framework for Self-Improving Agent Harnesses

DevFeed: [Harnesses That Improve Themselves: A Unifying Framework for Self-Improving Agent Harnesses](<https://devfeed.tech/articles/harnesses-that-improve-themselves-a-unifying-framework-for-self-improving-agent-harnesses-140008.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22299>)

Author: Deep Shah, Nehal Kathrotia

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

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

Topics: [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [Code Modularity](<https://devfeed.tech/topics/code-modularity.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>), [Error Handling](<https://devfeed.tech/topics/error-handling.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [actions](<https://devfeed.tech/tags/actions.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-control](<https://devfeed.tech/tags/agent-control.md>), [agent-harness](<https://devfeed.tech/tags/agent-harness.md>), [agent-harnesses](<https://devfeed.tech/tags/agent-harnesses.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [bench](<https://devfeed.tech/tags/bench.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [bloat](<https://devfeed.tech/tags/bloat.md>), [bottleneck](<https://devfeed.tech/tags/bottleneck.md>), [bugs](<https://devfeed.tech/tags/bugs.md>), [code](<https://devfeed.tech/tags/code.md>), [components](<https://devfeed.tech/tags/components.md>), [context](<https://devfeed.tech/tags/context.md>), [control](<https://devfeed.tech/tags/control.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [framework](<https://devfeed.tech/tags/framework.md>), [self-improving-agent](<https://devfeed.tech/tags/self-improving-agent.md>)

### AI overview

The framework treats an agent harness as four editable runtime parts--context and state, agent control, action interfaces, and verification and recovery--updated through a loop that diagnoses failures, proposes patches, validates them, and deploys changes. The authors report that tracing long execution histories can uncover reusable mechanisms such as pre-stop verifiers, loop breakers, and persistent memory. They argue that gated modular updates generalize better across models and limit token cost compared with full-harness rewrites, which can overfit or break across modules. On SWE-bench Lite, they report that gated modular evolution outperformed prompt reflection and full-harness rewrites across models and repositories at a fraction of the token cost. The abstract also describes defenses against prompt bloat, benchmark gaming, and evaluator tampering.

### Source excerpt

arXiv:2611.22299v1 Announce Type: new Abstract: Whether a foundation model succeeds as an autonomous agent depends as much on its runtime software as on its weights. This runtime, the agent harness, manages context and memory, steers multi-turn control, executes tool actions in sandboxes, and verifies work before stopping. Because models differ in how they use tools and how they fail, a static harness tuned for one model or task often stalls, wastes tokens, or regresses on another. This bottleneck has driven a shift from prompt tuning to self-improving agent harnesses: systems that inspect their own execution traces, separate runtime bugs from model limits, and rewrite their own code. We present a unifying architectural and design framework for self-improving agent harnesses. The framework pairs four editable runtime components (context and state, agent control, action interfaces, and verification and recovery) with a four-stage loop that diagnoses failures, proposes code patches, validates candidates, and deploys updates. Using this framework, we show how optimizers trace errors across long trajectories and discover reusable runtime mechanisms such as pre-stop verifiers, loop breakers, and persistent memory. We further show why full-harness rewrites overfit or break across modules, and how modular updates with validation gates generalize across models while bounding token cost. To keep self-modification safe, we establish practical defenses against recurring failure modes including prompt bloat, benchmark gaming, and evaluator tampering. Finally, experiments on SWE-bench Lite validate this framework, showing that gated modular evolution outperforms prompt reflection and full-harness rewrites across models and repositories at a fraction of the token cost.

## Retrieval-Augmented Correction for Generalist Medical Image Segmentation

DevFeed: [Retrieval-Augmented Correction for Generalist Medical Image Segmentation](<https://devfeed.tech/articles/retrieval-augmented-correction-for-generalist-medical-image-segmentation-140128.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23189>)

Author: Vincenzo Marcian\\\`o, Xiaoming Zhang, Michela Antonelli, Sebastien Ourselin, Maria A Zuluaga

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

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

Topics: [Medical imaging](<https://devfeed.tech/topics/medical-imaging.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [Vectors & LLMs](<https://devfeed.tech/topics/vectors-llms.md>), [Image](<https://devfeed.tech/topics/image.md>)

Tags: [context](<https://devfeed.tech/tags/context.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [generalist](<https://devfeed.tech/tags/generalist.md>), [input](<https://devfeed.tech/tags/input.md>), [knowledge-base](<https://devfeed.tech/tags/knowledge-base.md>), [mask](<https://devfeed.tech/tags/mask.md>), [medical](<https://devfeed.tech/tags/medical.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [public](<https://devfeed.tech/tags/public.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [ultrasound](<https://devfeed.tech/tags/ultrasound.md>)

### AI overview

GRACE refines medical image segmentation masks by retrieving verified prior cases from a shared knowledge base, which can serve as a shape prior. Its refiner is trained once across multiple anatomies; in a leave-one-anatomy-out experiment, verified cases for an unseen solid organ were added to the index while the refiner remained frozen. Across 18 public CT, MRI, and ultrasound datasets and seven upstream segmentors, GRACE added up to 0.35 DSC for weak inputs (baseline 0.386), while changes were small on a strong segmentor (baseline 0.932). Transfer to modalities absent from both training and the index was partial and anatomy-dependent, and remains an open research direction.

### Source excerpt

arXiv:2611.23189v1 Announce Type: cross Abstract: Automated segmentation is routinely used as the first step in quantitative image-based clinical workflows, where predicted masks are subsequently reviewed and corrected by human operators. Automated correction of the residual structural errors is currently addressed one anatomy at a time, so the number of trained correctors scales with the number of segmentation targets. We present GRACE, a retrieval-augmented refiner that edits an input mask under context drawn from a shared knowledge base of verified prior cases. The refiner is trained once on a multi-organ corpus spanning compact organs, cavities, branching vessels, and tumour lesions. In the held-out dataset-transfer evaluation, three held-out datasets query a fixed knowledge base built from training data only. A separate leave-one-anatomy-out experiment tests parameter-free index extension: verified cases of an unseen solid organ enter the knowledge base, and the refiner stays frozen. A light shape-aware projection head on a frozen multimodal encoder maps embedding similarity to the shape agreement of the masks, so the knowledge base returns cases that the refiner can use as a shape prior. We evaluate GRACE on 18 public datasets spanning CT, MRI and ultrasound, using nine refinement configurations and seven upstream segmentors that span weak, mid, and strong operating regimes. We report each change with the DSC of its input mask. On the weak tier, the input DSC is 0.386 and GRACE adds up to +0.35 DSC. On a strong segmentor, the input DSC is 0.932 and the mean change is small. Transfer to an imaging modality that is absent from training and from the index is partial and depends on the anatomy, and remains an open research direction. Our results suggest retrieval-conditioned refinement is a practical alternative to anatomy-specific correction models. Code: https://github.com/robustml-eurecom/grace

## Efficient Self-Supervised Adaptation of 3D Abdominal Vision-Language Model for Institution-Specific HCC Classification via Full Fine-Tuning and PEFT

DevFeed: [Efficient Self-Supervised Adaptation of 3D Abdominal Vision-Language Model for Institution-Specific HCC Classification via Full Fine-Tuning and PEFT](<https://devfeed.tech/articles/efficient-self-supervised-adaptation-of-3d-abdominal-vision-language-model-for-institution-specific-hcc-classification-via-full-fine-tuning-and-peft-141782.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22504>)

Author: Febryan Putra Kartika, Cheng-Yu Ma, Ying-Jia Lin, Chi-Tung Cheng, Kuan-Fu Chen

Published: 2026-11-18T05:00:00Z

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>), [lora](<https://devfeed.tech/topics/lora.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [acquisition](<https://devfeed.tech/tags/acquisition.md>), [annotations](<https://devfeed.tech/tags/annotations.md>), [classification](<https://devfeed.tech/tags/classification.md>), [contrastive-learning](<https://devfeed.tech/tags/contrastive-learning.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [lora](<https://devfeed.tech/tags/lora.md>), [medical](<https://devfeed.tech/tags/medical.md>), [peft](<https://devfeed.tech/tags/peft.md>), [pretraining](<https://devfeed.tech/tags/pretraining.md>), [temperature](<https://devfeed.tech/tags/temperature.md>), [training](<https://devfeed.tech/tags/training.md>), [vision-language-model](<https://devfeed.tech/tags/vision-language-model.md>), [vision-language-models-vlms](<https://devfeed.tech/tags/vision-language-models-vlms.md>)

### AI overview

The study reports that a pretrained 3D medical vision-language model can be adapted to institution-specific HCC classification using self-supervised contrastive learning on paired CT scans and radiology reports, without lesion annotations or HCC labels as training targets. Under cross-institutional domain shift, full fine-tuning achieved AUC 0.82 and F1 0.71; PEFT achieved AUC 0.79 and F1 0.70. Full fine-tuning converged quickly but overfit within a few epochs. PEFT used ConvLoRA for the image encoder and LoRA for the text encoder, updated about 1% of model parameters, and required more training steps. Both improved on the pretrained model, though results remained below those reported for supervised HCC classification. The abstract also reports that contrastive temperature initialization significantly affected performance.

### Source excerpt

arXiv:2611.22504v1 Announce Type: cross Abstract: Medical vision-language models (VLMs) capture cross-modal relationships between image and text, yet their adaptation to institution-specific clinical tasks remains underexplored. In this study, we fine-tuned a pretrained 3D medical VLM for hepatocellular carcinoma (HCC) classification using paired abdominal CT scans and radiology reports from a different institution with acquisition characteristics that differ from the model's original pretraining corpus. We compared two adaptation strategies: full fine-tuning and parameter-efficient fine-tuning (PEFT), motivated by the common use of PEFT to reduce computational cost and enable adaptation under limited-data constraints. Both approaches achieve strong downstream HCC classification performance despite the cross-institutional domain shift, with PEFT reaching an AUC of 0.79 and F1 of 0.70, and full fine-tuning achieving an AUC of 0.82 and F1 of 0.71. Although these results remain below the performance reported by supervised HCC classification approaches, our method achieves improvements over the pretrained model through CT-report contrastive learning without requiring lesion-level annotations or HCC classification labels as training targets. Full fine-tuning converges rapidly but overfits within a few epochs, whereas PEFT (ConvLoRA for the image encoder and LoRA for the text encoder) attains comparable performance while updating approximately 1% of the model parameters, although requiring more training steps. We also examine the role of contrastive temperature, observing that temperature initialization significantly affects classification performance. This study demonstrates that a 3D medical VLM can be efficiently adapted to institution-specific HCC classification using self-supervised CT-report contrastive learning, while highlighting the practical trade-offs between full fine-tuning and parameter-efficient fine-tuning.

## A Chain-of-Thought Prompt Engineering Framework for Automated Technical Specification Generation in Enterprise Integration: A Real-World Case Study

DevFeed: [A Chain-of-Thought Prompt Engineering Framework for Automated Technical Specification Generation in Enterprise Integration: A Real-World Case Study](<https://devfeed.tech/articles/a-chain-of-thought-prompt-engineering-framework-for-automated-technical-specification-generation-in-enterprise-integration-a-real-world-case-study-139577.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21617>)

Author: Aryan Dhingra, Gokul Menon, Anusha Antony

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

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

Topics: [Prompt Engineering](<https://devfeed.tech/topics/prompt-engineering.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>)

Tags: [automated](<https://devfeed.tech/tags/automated.md>), [case-study](<https://devfeed.tech/tags/case-study.md>), [chain-of-thought](<https://devfeed.tech/tags/chain-of-thought.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [enterprise-integration](<https://devfeed.tech/tags/enterprise-integration.md>), [framework](<https://devfeed.tech/tags/framework.md>), [generation](<https://devfeed.tech/tags/generation.md>), [integration](<https://devfeed.tech/tags/integration.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [prompt-engineering](<https://devfeed.tech/tags/prompt-engineering.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [specification](<https://devfeed.tech/tags/specification.md>), [study](<https://devfeed.tech/tags/study.md>), [technical](<https://devfeed.tech/tags/technical.md>)

### AI overview

In a case study of 15 production interfaces at a large Australian distributor migrating more than 250 integrations to Azure Integration Services, a GitHub Copilot framework cut technical specification authoring effort from 793 to 300 person-hours (62.2% pooled reduction). It grounds generation in functional specifications, organizational standards, and an approved template, using a six-stage pipeline with human validation and version-controlled prompts. Reviewer-assessed accuracy was 74.3%, and developer satisfaction averaged 3.7/5. Results worsened as interfaces became more complex: effort reduction declined from 70.6% to 58.9%, and accuracy from 91% to 62.5%; human work remained concentrated in diagrams, sequence flows, and cross-system content. The abstract also reports comparisons involving GPT-4.1/5, Claude Sonnet 4, and Copilot Ask and Agent modes, but gives no comparative results.

### Source excerpt

arXiv:2611.21617v1 Announce Type: new Abstract: Technical Specification Documents (TSDs) translate functional requirements into implementable designs and are mandatory artifacts in governed Enterprise Application Integration (EAI) programs, yet their production remains a manual, expertise-intensive bottleneck. Large Language Models (LLMs) have been studied extensively for code generation, but long-form, template-governed engineering documents remain comparatively unexplored. This paper reports an industrial case study, conducted during a migration of more than 250 enterprise integrations to Microsoft Azure Integration Services at a large Australian wholesale distributor, that introduces a chain-of-thought prompt-engineering framework for automated TSD generation using GitHub Copilot. The framework combines a two-phase generation architecture that grounds the model in the functional specification, organizational standards, and the approved template; a six-stage prompting pipeline with human validation checkpoints; and Prompt-as-a-Code, under which prompts are version-controlled and reviewed as engineering artifacts. Applied to a stratified sample of fifteen production interfaces, the framework reduced TSD authoring effort from 793 to 300 person-hours, a 62.2% pooled reduction and a 64.0% mean per-document reduction (Wilcoxon signed-rank, n = 15, p < 0.001), at 74.3% reviewer-assessed accuracy and 3.7/5 developer satisfaction. Benefits degrade monotonically with interface complexity: between the simplest and most complex bands, effort reduction falls from 70.6% to 58.9% and accuracy from 91% to 62.5%, and the residual human effort concentrates in diagrams, sequence flows, and cross-system integration content. We additionally report comparative observations of GPT-4.1/5 and Claude Sonnet 4, as well as Copilot's Ask and Agent interaction modes, offering reusable, practice-grounded guidance for enterprise settings.

## Fairness-Aware Test-Time Prompt Tuning

DevFeed: [Fairness-Aware Test-Time Prompt Tuning](<https://devfeed.tech/articles/fairness-aware-test-time-prompt-tuning-141598.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2608.25707>)

Author: Yoann Launay, Parameswaran Kamalaruban, Tom Kempton, Stuart Burrell, David Sutton

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

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

Topics: [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [responsible-ai](<https://devfeed.tech/topics/responsible-ai.md>), [peft](<https://devfeed.tech/topics/peft.md>)

Tags: [clip](<https://devfeed.tech/tags/clip.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [fairness](<https://devfeed.tech/tags/fairness.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [test-time-adaptation](<https://devfeed.tech/tags/test-time-adaptation.md>), [vision-language-models](<https://devfeed.tech/tags/vision-language-models.md>)

### AI overview

FairTPT adapts CLIP at test time to address fairness under subpopulation shifts without retraining or demographic attributes. Through soft-prompt tuning, it jointly minimizes target marginal entropy and maximizes spurious marginal entropy. The authors report that standard episodic test-time adaptation generally worsens disparities, while excessive suppression of spurious attributes can cause catastrophic forgetting. Monitoring test-time changes in target loss within the linear regime can prevent this collapse while retaining fairness gains on reactive data and overall performance. The abstract reports that FairTPT outperforms the state-of-the-art episodic test-time debiasing methods.

### Source excerpt

arXiv:2608.25707v2 Announce Type: replace Abstract: Vision-language models have displayed remarkable capabilities in multi-modal understanding and are increasingly used in critical applications where economic and practical deployment constraints prohibit re-training or fine-tuning. However, these models can also exhibit systematic biases that disproportionately affect protected demographic groups and existing approaches to addressing these biases require extensive model retraining and access to demographic attributes. There is a clear need to develop test-time adaptation (TTA) approaches that improve the fairness characteristics of pretrained models under distributional shift. In this paper, we evaluate how episodic TTA affects fairness in CLIP classification under subpopulation shifts and develop FairTPT, a novel fairness-aware episodic TTA method that jointly minimizes target marginal entropy while maximizing spurious marginal entropy through soft-prompt tuning. We find that standard episodic TTA generally exacerbates disparities between majority and minority groups, that blinding a model to spurious attributes without degrading target performance is inherently challenging, and that excessive blinding can lead to catastrophic forgetting. This model collapse can be prevented by monitoring test-time changes in target loss within the linear regime, while still achieving fairness improvements on reactive data and preserving overall performance. FairTPT outperforms all state-of-the-art episodic test-time debiasing methods and establishes a foundation for robust TTA, which is essential for achieving fairness in practice.

## Synergizing Evidence Retrieval and Utilization for Self-Improving Multimodal Scientific Document Reasoning

DevFeed: [Synergizing Evidence Retrieval and Utilization for Self-Improving Multimodal Scientific Document Reasoning](<https://devfeed.tech/articles/synergizing-evidence-retrieval-and-utilization-for-self-improving-multimodal-scientific-document-reasoning-140084.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22837>)

Author: Frank Nie, Ethan B. Liu, Yuan Zhu, Jindong Han

Published: 2026-11-18T05:00:00Z

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>), [information retrieval](<https://devfeed.tech/topics/information-retrieval.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>)

Tags: [acquisition](<https://devfeed.tech/tags/acquisition.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [budget](<https://devfeed.tech/tags/budget.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [framework](<https://devfeed.tech/tags/framework.md>), [grounded](<https://devfeed.tech/tags/grounded.md>), [guidance](<https://devfeed.tech/tags/guidance.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [ontology](<https://devfeed.tech/tags/ontology.md>), [optimize](<https://devfeed.tech/tags/optimize.md>), [organization](<https://devfeed.tech/tags/organization.md>), [question](<https://devfeed.tech/tags/question.md>), [questions](<https://devfeed.tech/tags/questions.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [reward](<https://devfeed.tech/tags/reward.md>), [settings](<https://devfeed.tech/tags/settings.md>), [sources](<https://devfeed.tech/tags/sources.md>), [space](<https://devfeed.tech/tags/space.md>), [tables](<https://devfeed.tech/tags/tables.md>)

### AI overview

The framework separates evidence retrieval from evidence use for multimodal scientific document reasoning. A Retriever selects source-aligned text, figure, and table evidence under a context budget; a Perceiver organizes selected evidence and the question into a question-conditioned latent ontology; and a Reasoner uses that guidance alongside the original sources to generate grounded answers. During training, the three components are alternately optimized with a shared answering reward. Retrieval Potential Feedback and Latent Efficacy Feedback diagnose whether questions need better retrieval or latent guidance and direct targeted resampling. Across four benchmarks and seven evaluation settings, it reports average scores of 69.9 with a 7B Reasoner and 75.0 with an 8B Reasoner, exceeding the strongest baseline by 4.0 and 9.1 points, respectively.

### Source excerpt

arXiv:2611.22837v1 Announce Type: new Abstract: Scientific document reasoning requires an agent to acquire and integrate evidence distributed across text, figures, and tables. Retrieval can identify relevant sources while missing complementary support, and answer-level feedback alone does not reveal whether an error stems from evidence acquisition or evidence use. We present an agentic Retriever--Perceiver--Reasoner framework for multimodal scientific document reasoning. The Retriever sequentially selects source-aligned evidence under a context budget. Given the selected evidence and question, the Perceiver constructs a question-conditioned latent ontology, an implicit organization of multimodal evidence in a shared continuous space. The Reasoner combines this latent guidance with the original sources to produce grounded answers. We train the three components through diagnostic-guided co-evolution: alternating updates optimize them with a shared answering reward, while Retrieval Potential Feedback (RPF) and Latent Efficacy Feedback (LEF) identify questions that need better retrieval or latent guidance and direct targeted resampling. Across four benchmarks and seven evaluation settings, our framework obtains average scores of 69.9 and 75.0 with 7B and 8B Reasoner backbones, exceeding the strongest baseline by 4.0 and 9.1 points, respectively.

## Hidden-State Similarity Predicts Re-Elicitation After Inoculation Prompting

DevFeed: [Hidden-State Similarity Predicts Re-Elicitation After Inoculation Prompting](<https://devfeed.tech/articles/hidden-state-similarity-predicts-re-elicitation-after-inoculation-prompting-140120.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23096>)

Author: Rose N. Kwon, Samy Mammeri, Sabyasachi Sahoo, Christian Gagn\\'e, Thomas Jiralerspong

Published: 2026-11-18T05:00:00Z

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>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [peft](<https://devfeed.tech/topics/peft.md>)

Tags: [activation](<https://devfeed.tech/tags/activation.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [eval](<https://devfeed.tech/tags/eval.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [misalignment](<https://devfeed.tech/tags/misalignment.md>), [models](<https://devfeed.tech/tags/models.md>), [prompting](<https://devfeed.tech/tags/prompting.md>), [prompts](<https://devfeed.tech/tags/prompts.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [study](<https://devfeed.tech/tags/study.md>)

### AI overview

Inoculation prompting can suppress harmful behavior learned during narrow-task fine-tuning, but evaluation prompts may bring it back. The study finds that textual similarity to the inoculation prompt is an incomplete predictor of re-elicitation; prompts are more likely to trigger suppressed behavior when they produce activation states similar to those induced by the inoculation context. The authors suggest activation-space analysis may help identify when suppressed behaviors remain accessible under evaluation prompts.

### Source excerpt

arXiv:2611.23096v1 Announce Type: new Abstract: Fine-tuning on narrow harmful tasks can cause emergent misalignment, where models generalize harmful behavior beyond the training distribution. Inoculation prompting can reduce this effect by explicitly eliciting the undesired behavior during training, but recent work shows that the behavior can reappear when evaluation prompts contain cues from the training context. We study what makes such prompts effective triggers. We find that, surprisingly, textual similarity to the inoculation prompt is an incomplete predictor. Instead, prompts are more likely to re-elicit suppressed behavior when they induce activation states similar to those produced by the inoculation context. These findings advance our understanding of how inoculation prompting modulates conditional misalignment, and suggest that activation space analysis can help identify when suppressed behaviors remain accessible under eval-time prompts.

## How we fine-tuned a faster model for Patterns on Baseten

DevFeed: [How we fine-tuned a faster model for Patterns on Baseten](<https://devfeed.tech/articles/how-we-fine-tuned-a-faster-model-for-patterns-on-baseten-142009.md>)

Original publisher: [Read original article](<https://www.datadoghq.com/blog/fine-tuning-patterns-baseten/>)

Author: Fouad Wahabi; Vincent Cornet; Alex Barksdale

Published: 2026-11-18T00:00:00Z

Content type: article

Language: en

Sources: [Datadog](<https://devfeed.tech/sources/datadog-the-monitor-blog.md>)

Topics: [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [ai observability](<https://devfeed.tech/topics/ai-observability.md>), [LLM observability](<https://devfeed.tech/topics/llm-observability.md>), [LLM optimization](<https://devfeed.tech/topics/llm-optimization.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [baseten](<https://devfeed.tech/topics/baseten.md>), [lora](<https://devfeed.tech/topics/lora.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [post-training](<https://devfeed.tech/topics/post-training.md>), [model-deployment](<https://devfeed.tech/topics/model-deployment.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [78](<https://devfeed.tech/tags/78.md>), [about](<https://devfeed.tech/tags/about.md>), [actions](<https://devfeed.tech/tags/actions.md>), [adapter](<https://devfeed.tech/tags/adapter.md>), [after](<https://devfeed.tech/tags/after.md>), [agent-observability](<https://devfeed.tech/tags/agent-observability.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [baseten](<https://devfeed.tech/tags/baseten.md>), [datadog](<https://devfeed.tech/tags/datadog.md>), [faster](<https://devfeed.tech/tags/faster.md>), [fine](<https://devfeed.tech/tags/fine.md>), [generation](<https://devfeed.tech/tags/generation.md>), [model](<https://devfeed.tech/tags/model.md>), [one](<https://devfeed.tech/tags/one.md>), [patterns](<https://devfeed.tech/tags/patterns.md>), [production](<https://devfeed.tech/tags/production.md>), [time](<https://devfeed.tech/tags/time.md>), [traces](<https://devfeed.tech/tags/traces.md>), [under](<https://devfeed.tech/tags/under.md>), [using](<https://devfeed.tech/tags/using.md>)

### AI overview

Datadog fine-tuned Qwen 3.5 9B for Patterns' repetitive interaction summarization, reducing average generation time on 150 held-out cases from 44.7 seconds to 0.89 seconds. Patterns' production LLM spans supplied examples pairing interactions and clustering instructions with existing summaries; rows were split by trace to limit test leakage. On Baseten, the base model often exhausted its token budget before producing a parseable answer: only 48% of its outputs finished, and 44.7% passed the LLM judge. After LoRA fine-tuning with early stopping based on validation loss, all 150 outputs finished and 93.3% passed. Datadog's pup finetune workflow supported dataset review, benchmarking, training, and evaluation, with approval before paid steps. The reported evaluation used five held-out traces and suggests the model generalized across the tested clustering instructions.

### Source excerpt

Datadog fine-tuned a summarization model using production traces, cutting average generation time from about 45 seconds to under one second.

[Next page](<https://devfeed.tech/topics/enhancing-ai-model-performance.md?cursor=WyIyMDI2LTExLTE4VDAwOjAwOjAwKzAwOjAwIiwgIjk5ZjhkYWQxLWI1NTEtNDdlMi1hNDQzLTIxZjFjMTU1YjQ1YyJd>)