# Neural Network

A neural network is a machine learning model that stacks simple neurons in layers and learns pattern-recognizing weights and biases from data to map inputs to outputs.

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

## DAWN: Robust Neural Blind Deconvolution via Alternating Wiener Guidance

DevFeed: [DAWN: Robust Neural Blind Deconvolution via Alternating Wiener Guidance](<https://devfeed.tech/articles/dawn-robust-neural-blind-deconvolution-via-alternating-wiener-guidance-145555.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25621>)

Author: Yida Pan, Yixuan Jia, Chaoyan Huang, Qing Qu, Ismail Alkhouri

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

Content type: article

Language: en

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

Topics: [Image](<https://devfeed.tech/topics/image.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Deep learning](<https://devfeed.tech/topics/deep-learning.md>)

Tags: [computer-vision](<https://devfeed.tech/tags/computer-vision.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>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [dip](<https://devfeed.tech/tags/dip.md>), [gaussian](<https://devfeed.tech/tags/gaussian.md>), [image](<https://devfeed.tech/tags/image.md>), [kernel](<https://devfeed.tech/tags/kernel.md>), [microscopy](<https://devfeed.tech/tags/microscopy.md>), [network](<https://devfeed.tech/tags/network.md>), [neural](<https://devfeed.tech/tags/neural.md>), [noise](<https://devfeed.tech/tags/noise.md>), [photography](<https://devfeed.tech/tags/photography.md>), [robust](<https://devfeed.tech/tags/robust.md>), [training-data](<https://devfeed.tech/tags/training-data.md>)

### AI overview

DAWN is a training-data-free deep image prior framework for blind image deconvolution, designed to reduce noise overfitting and improve blur-kernel estimates. It alternates between feeding model-based estimates into the network and refining them with the network's prior, using autoencoding regularization on both image and kernel networks. On natural and scientific images with Gaussian and Poisson noise, the authors report better image and kernel quality than state-of-the-art methods, including up to 4.5 dB over the next-best baseline on the Levin dataset at σ = 0.1. Without training images, DAWN is also reported to remain highly competitive with a diffusion model trained on many in-distribution FFHQ images.

### Source excerpt

arXiv:2611.25621v1 Announce Type: cross Abstract: Blind image deconvolution (BID) is an ill-posed inverse problem that aims to jointly recover a sharp image and blur kernel from a single degraded observation. Under nontrivial and high noise, which is common in photography and scientific imaging such as microscopy, BID remains particularly challenging, as existing methods suffer from noise overfitting and inaccurate kernel estimation. Existing training-data-free learning methods for BID, such as those based on the deep image prior (DIP), often utilize external guidance by including model-based estimates as additional loss terms, which underexploits the information these estimates carry especially under nontrivial noise. To address these challenges, we propose DAWN, a training-data-free DIP framework that alternates between injecting model-based estimates directly into the model input and refining this imperfect estimate with the network's own prior. To provide more reliable kernel estimates for these injections, we apply autoencoding regularization to both the image and kernel networks. On natural and scientific images under varying levels of Gaussian and Poisson noise, DAWN outperforms state-of-the-art methods in both image and kernel quality (achieving up to +4.5 dB over the next best baseline on the Levin dataset at $\sigma = 0.1$). Without any training images, DAWN remains highly competitive with a state-of-the-art diffusion model on FFHQ, despite the latter being trained on a massive number of in-distribution images.

## Attention Head Specialization: Phase Transitions in Sample Complexity

DevFeed: [Attention Head Specialization: Phase Transitions in Sample Complexity](<https://devfeed.tech/articles/attention-head-specialization-phase-transitions-in-sample-complexity-145153.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26228>)

Author: L\\'eo Catteau, O. Duranthon, Lenka Zdeborov\\'a

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

Content type: article

Language: en

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

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Transformer](<https://devfeed.tech/topics/transformer.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [attention](<https://devfeed.tech/tags/attention.md>), [bert](<https://devfeed.tech/tags/bert.md>), [bias](<https://devfeed.tech/tags/bias.md>), [cascade](<https://devfeed.tech/tags/cascade.md>), [changes](<https://devfeed.tech/tags/changes.md>), [complexity](<https://devfeed.tech/tags/complexity.md>), [cond-mat-dis-nn](<https://devfeed.tech/tags/cond-mat-dis-nn.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [data-structure](<https://devfeed.tech/tags/data-structure.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [emergence](<https://devfeed.tech/tags/emergence.md>), [error](<https://devfeed.tech/tags/error.md>), [features](<https://devfeed.tech/tags/features.md>), [function](<https://devfeed.tech/tags/function.md>), [learning](<https://devfeed.tech/tags/learning.md>), [model](<https://devfeed.tech/tags/model.md>), [multi-head](<https://devfeed.tech/tags/multi-head.md>), [online](<https://devfeed.tech/tags/online.md>), [online-learning](<https://devfeed.tech/tags/online-learning.md>), [sample](<https://devfeed.tech/tags/sample.md>), [specialization](<https://devfeed.tech/tags/specialization.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>)

### AI overview

The excerpt reports that BERT heads specialized to syntactic relations emerge as training dataset size increases. In a solvable high-dimensional multi-head attention model, specialization is a phase transition in sample complexity, marked by non-analytic changes in test error and other observables. Depending on data structure and regularization, heads may differentiate without conveying useful information, or specialize through successive transitions that recover progressively weaker latent features. The authors also relate these statistical transitions to stagewise specialization in online learning, separating statistical effects from training dynamics.

### Source excerpt

arXiv:2611.26228v1 Announce Type: new Abstract: Attention heads specialize into distinct functions during training, often emerging in stages together with sharp changes in model behavior. While this phenomenon has mostly been studied as a function of training time, we ask how head specialization depends on the amount of available training data. Empirically, we find that heads specialized to syntactic relations in BERT emerge as the dataset size increases. We then study this phenomenon in a solvable high-dimensional model of multi-head attention and establish that, in this model, head specialization is a genuine phase transition in sample complexity, corresponding to a non-analyticity in the test error and other observables. Depending on data structure and regularization, specialization can follow uninformative head differentiation or occur through a cascade of phase transitions that progressively recover weaker latent features, providing a statistical form of simplicity bias. Finally, we relate these statistical transitions to the stagewise emergence of specialized heads during online learning, disentangling statistical from dynamical origins of head specialization.

## Perfumer: A Large Language Model-based Agentic Framework Accelerating Structural Relaxation and Electronic Band Prediction for First-Principle Calculation

DevFeed: [Perfumer: A Large Language Model-based Agentic Framework Accelerating Structural Relaxation and Electronic Band Prediction for First-Principle Calculation](<https://devfeed.tech/articles/perfumer-a-large-language-model-based-agentic-framework-accelerating-structural-relaxation-and-electronic-band-prediction-for-first-principle-calculation-145166.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26349>)

Author: Zhiwei Li, Yuhao Zhou, Yikai Yang, Haonan Wang, Qing Ye, Jiancheng Lv, Kazunari Matsuda

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

Content type: article

Language: en

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

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [absolute](<https://devfeed.tech/tags/absolute.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [automatically](<https://devfeed.tech/tags/automatically.md>), [band](<https://devfeed.tech/tags/band.md>), [based](<https://devfeed.tech/tags/based.md>), [combines](<https://devfeed.tech/tags/combines.md>), [composition](<https://devfeed.tech/tags/composition.md>), [cond-mat-mtrl-sci](<https://devfeed.tech/tags/cond-mat-mtrl-sci.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [electronic](<https://devfeed.tech/tags/electronic.md>), [first](<https://devfeed.tech/tags/first.md>), [framework](<https://devfeed.tech/tags/framework.md>), [language](<https://devfeed.tech/tags/language.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [large](<https://devfeed.tech/tags/large.md>), [large-language-model](<https://devfeed.tech/tags/large-language-model.md>), [model](<https://devfeed.tech/tags/model.md>), [prediction](<https://devfeed.tech/tags/prediction.md>)

### AI overview

Perfumer links natural-language requests to workflows that combine Quantum ESPRESSO (QE) with predictive models for structural relaxation and electronic bands. In tested structures, pre-relaxation with Perfumer-relax followed by QE refinement produced a reported 2.75-fold overall speedup over QE alone. Perfumer-band predicts 12 band energies at requested wavevectors using a periodic graph neural network and latent Hermitian spectral head, preserving reciprocal-lattice periodicity and time-reversal symmetry without constructing a Hamiltonian. Trained on 125,153 band structures, it achieved a median per-material mean absolute error of 0.157 eV across 2,607 test materials, with inference taking about 97 ms per material on two consumer CPU cores. The described agent handles workflows from uploaded structures to automatically constructed candidates; the supplied excerpt does not establish performance beyond the reported tests.

### Source excerpt

arXiv:2611.26349v1 Announce Type: cross Abstract: First-principles structural relaxation and band calculations require substantial computation and expertise. Here we present Perfumer, a large language model-based agentic framework that couples Quantum ESPRESSO with two predictive models. Across tested structures, pre-relaxation with Perfumer-relax model followed by QE refinement yielded a 2.75-fold overall speedup over QE alone. Perfumer-band model combines a periodic graph neural network with a latent Hermitian spectral head returns 12 band energies at any requested wavevector, preserving reciprocal lattice periodicity and time-reversal symmetry without constructing Hamiltonian. Trained on 125,153 self-curated band structures, Perfumer-band model reaches a median per-material mean absolute error of 0.157 eV on 2,607 test materials, including lanthanide, actinide and magnetic compounds, from held-out composition families spanning 85 elements, taking about 97 ms per material on two consumer CPU cores. The agent turns natural-language requests into relaxation-to-band workflows, from a single uploaded structure to automatically constructed new candidates and returns predicted properties.

## Statistical Inference with Ill-Conditioned Random Weight Neural Network Sieves

DevFeed: [Statistical Inference with Ill-Conditioned Random Weight Neural Network Sieves](<https://devfeed.tech/articles/statistical-inference-with-ill-conditioned-random-weight-neural-network-sieves-145176.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26410>)

Author: Giovanni Ballarin, Lyudmila Grigoryeva, Lukas Gonon

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

Content type: article

Language: en

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

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

Tags: [2](<https://devfeed.tech/tags/2.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [approximation](<https://devfeed.tech/tags/approximation.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [based](<https://devfeed.tech/tags/based.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [beta](<https://devfeed.tech/tags/beta.md>), [condition](<https://devfeed.tech/tags/condition.md>), [conditional](<https://devfeed.tech/tags/conditional.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [covariance](<https://devfeed.tech/tags/covariance.md>), [covering](<https://devfeed.tech/tags/covering.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [develop](<https://devfeed.tech/tags/develop.md>), [econ-em](<https://devfeed.tech/tags/econ-em.md>), [inference](<https://devfeed.tech/tags/inference.md>), [network](<https://devfeed.tech/tags/network.md>), [neural](<https://devfeed.tech/tags/neural.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [random](<https://devfeed.tech/tags/random.md>), [stat-me](<https://devfeed.tech/tags/stat-me.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>), [weight](<https://devfeed.tech/tags/weight.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

The paper shows that random weight neural network sieves become increasingly ill-conditioned as their dimension grows: under polynomial covering and Lipschitz conditions, the population covariance eigenvalues have geometry-dependent polynomial upper bounds, so the smallest eigenvalue vanishes. It develops theory that accounts for this behavior, including faster-than-Monte-Carlo L² approximation and stratified-sampling L∞ bounds under a β-Hölder source condition, plus ridge least-squares consistency guarantees based on effective dimension and stability conditions. For inference, it establishes asymptotic linearity and Gaussian limits for linear directions despite the vanishing eigenvalue; directional nondegeneracy is expressed through a ridge-normalized Rayleigh quotient and extended to smooth functionals using Riesz representation. Consistent ridge-adjusted variance estimation enables feasible studentized inference. The abstract reports numerical experiments on conditioning and finite-sample behavior.

### Source excerpt

arXiv:2611.26410v1 Announce Type: cross Abstract: Random weight neural networks (RWNNs) are shallow networks whose hidden-layer parameters are randomly generated and then kept fixed, while only the output coefficients are estimated. Conditional on the sampled parameters, RWNN regression is a linear sieve problem with random approximation space. We show that such random sieves cannot remain uniformly well-conditioned as their dimension grows. For parametrized feature families satisfying polynomial covering and Lipschitz conditions, the eigenvalues of the population covariance matrix are proven to have polynomial upper bounds determined by the covering geometry of the parameter space, leading to a vanishing smallest eigenvalue. We develop approximation, estimation, and inference theory that explicitly treats the ill-conditioning. Under a $\beta$-H\"older source condition, we obtain both faster-than-Monte-Carlo $L^2$ approximation rates and new $L^\infty$ approximation bounds based on stratified sampling. We then derive improved consistency guarantees for ridge-regularized least squares driven by ridge effective dimension, spectral, and leverage-based stability conditions. For inference, we establish asymptotic linearity and Gaussian limits for linear functional directions while allowing the smallest eigenvalue of the population covariance matrix to vanish. Nondegeneracy is directional: we characterize it with a ridge-normalized Rayleigh quotient and extend it to smooth functionals via Riesz representation. Consistency of ridge-adjusted variance estimators is proved, yielding feasible studentized inference. Numerical experiments illustrate the deterioration of random-feature conditioning and the finite-sample behavior of the resulting estimation and inference procedures.

## Neural Conditional Gaussian Model: Efficient Data Assimilation with Sequential Uncertainty Quantification for Nonlinear Dynamics

DevFeed: [Neural Conditional Gaussian Model: Efficient Data Assimilation with Sequential Uncertainty Quantification for Nonlinear Dynamics](<https://devfeed.tech/articles/neural-conditional-gaussian-model-efficient-data-assimilation-with-sequential-uncertainty-quantification-for-nonlinear-dynamics-145185.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26463>)

Author: Chuanqi Chen, Nan Chen

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

Content type: article

Language: en

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

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Data science and engineering](<https://devfeed.tech/topics/data-science-and-engineering.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [computer-science](<https://devfeed.tech/tags/computer-science.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data-driven](<https://devfeed.tech/tags/data-driven.md>), [decoder](<https://devfeed.tech/tags/decoder.md>), [encoder-decoder](<https://devfeed.tech/tags/encoder-decoder.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [gaussian](<https://devfeed.tech/tags/gaussian.md>), [math-ds](<https://devfeed.tech/tags/math-ds.md>), [model](<https://devfeed.tech/tags/model.md>), [neural](<https://devfeed.tech/tags/neural.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [uncertainty](<https://devfeed.tech/tags/uncertainty.md>), [uncertainty-quantification](<https://devfeed.tech/tags/uncertainty-quantification.md>)

### AI overview

Neural CGM learns a dynamical model by mapping physical states into a latent space with conditional Gaussian dynamics, enabling closed-form data-assimilation updates there. Nonlinear encoder-decoder mappings represent physical-space dynamics; a Gaussian decoder maps sequential latent posterior means and covariances back to physical variables using local moment propagation and learned corrections. These posterior statistics also enter a Gaussian negative log-likelihood during training, so the model learns from both estimated states and uncertainty. It can forecast and simulate without explicit governing equations. Benchmark experiments report accurate estimation and forecasting with substantially lower online assimilation cost than conventional nonlinear data-assimilation methods. The supplied abstract does not quantify these results.

### Source excerpt

arXiv:2611.26463v1 Announce Type: new Abstract: Data assimilation (DA) in partially observed nonlinear systems requires repeated estimation of both hidden states and their uncertainty. Conventional nonlinear filtering can be computationally expensive and often relies on an explicit dynamical model. In this paper, we develop the Neural Conditional Gaussian Model (Neural CGM), a data-driven dynamical model that maps unobserved physical states into a learned latent space, where the dynamics of the latent states follow a conditional Gaussian structure. This structure allows closed-form DA updates in the latent space, while nonlinear encoder-decoder mappings retain the flexibility needed to represent complex dynamics in physical space. A key feature of Neural CGM is its treatment of uncertainty in DA. The latent posterior mean and covariance are mapped to the physical variables through a Gaussian decoder that combines local moment propagation with learned nonlinear corrections. The resulting physical-space posterior statistics depend on the sequentially updated latent distribution rather than on a separate post-processing estimate. The same statistics enter a Gaussian negative log-likelihood during training, so both the posterior mean and uncertainty affect the learned model. Because Neural CGM learns the dynamical model itself, it can also be used for state forecasting and simulation without explicit knowledge of the governing equations. Numerical experiments on benchmark nonlinear dynamical systems show accurate state estimation and forecasting while substantially reducing online assimilation cost relative to conventional nonlinear DA methods. Neural CGM is therefore suited to repeated state estimation, uncertainty quantification, and prediction from partial observations.

## DL101 The Representation Capability of Neural Network Models

DevFeed: [DL101 The Representation Capability of Neural Network Models](<https://devfeed.tech/articles/dl101-the-representation-capability-of-neural-network-models-144995.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24980>)

Author: Fernando Berzal

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

Content type: article

Language: en

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

Topics: [Deep learning](<https://devfeed.tech/topics/deep-learning.md>), [Deep neural networks](<https://devfeed.tech/topics/deep-neural-networks.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>)

Tags: [approximation](<https://devfeed.tech/tags/approximation.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ne](<https://devfeed.tech/tags/cs-ne.md>), [deep](<https://devfeed.tech/tags/deep.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [error](<https://devfeed.tech/tags/error.md>), [gradient-descent](<https://devfeed.tech/tags/gradient-descent.md>), [guarantees](<https://devfeed.tech/tags/guarantees.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [regions](<https://devfeed.tech/tags/regions.md>), [relu](<https://devfeed.tech/tags/relu.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>), [tangent](<https://devfeed.tech/tags/tangent.md>), [training](<https://devfeed.tech/tags/training.md>), [universal-approximation](<https://devfeed.tech/tags/universal-approximation.md>), [universality](<https://devfeed.tech/tags/universality.md>)

### AI overview

Universal approximation shows that shallow neural networks with non-polynomial activations can approximate any continuous function on compact sets, but it gives no guarantees about model size, successful training, or generalization. Barron's results provide quantitative approximation rates for functions with bounded Fourier norm. Montúfar's result shows that deep networks can form exponentially more linear regions as depth increases; the excerpt states that, for the same parameter count, they are more expressive than shallow networks. These results explain representational capacity, not whether training will achieve it.

### Source excerpt

arXiv:2611.24980v1 Announce Type: cross Abstract: Universal approximation theorems (Cybenko, Funahashi, Hornik & Leshno) provide the mathematical justification for why neural networks are powerful enough to model complex, nonlinear phenomena. Shallow networks can approximate any continuous function on compact sets. Leshno's theorem clarified that non-polynomial activations are necessary for universality, which explains why ReLU, logistic, hyperbolic tangent, and other common activations work (but pure polynomials do not). However, approximation theorems provide no efficiency, training, nor generalization guarantees. They do not indicate how many neurons are needed for a given accuracy, whether gradient descent will find the right parameters, or whether networks trained on finite data will generalize well. Barron provided quantitative error bounds for particular function classes (approximation rates for functions with bounded Fourier norm), whereas Mont\'ufar was the first to show that expressivity grows exponentially with depth (in terms of the number of linear regions). Deep networks are strictly more expressive than shallow ones with the same number of parameters. Depth is fundamentally more powerful than width, providing a rigorous mathematical explanation for the use of deep learning models.

## Revealing Pathway Interactions in Network Computations with Complete Path Contribution

DevFeed: [Revealing Pathway Interactions in Network Computations with Complete Path Contribution](<https://devfeed.tech/articles/revealing-pathway-interactions-in-network-computations-with-complete-path-contribution-145142.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26101>)

Author: Shenghua Liu, Youssef Faragalla, Surya Ganguli, Stephen A. Baccus

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

Content type: article

Language: en

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

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [model architecture](<https://devfeed.tech/topics/model-architecture.md>), [Tool](<https://devfeed.tech/topics/tool.md>)

Tags: [algorithm](<https://devfeed.tech/tags/algorithm.md>), [circuit](<https://devfeed.tech/tags/circuit.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [interaction](<https://devfeed.tech/tags/interaction.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [mouse](<https://devfeed.tech/tags/mouse.md>), [network](<https://devfeed.tech/tags/network.md>), [q-bio-nc](<https://devfeed.tech/tags/q-bio-nc.md>)

### AI overview

Complete Path Contribution decomposes a network's output into mathematically complete contributions from individual input-to-output paths, enabling direct study of higher-order pathway interactions. Summing path contributions also defines complete contributions for circuits at different levels of granularity. An algorithm finds the k strongest paths without enumerating them all; its runtime was empirically linear in k rather than exponential in network depth. The authors apply the method to benchmark image classifiers to quantify excitatory and inhibitory interactions, and to an interpretable mouse-retina model, where it reveals pathway interaction motifs associated with dynamic receptive fields under time-varying natural stimuli.

### Source excerpt

arXiv:2611.26101v1 Announce Type: new Abstract: Networks pass signals through many layers of interconnected units, so explaining what they compute means tracing how information travels from input to output and how pathways interact. Most existing approaches analyze only pairwise interactions between components such as units, layers or attention heads. Even when these measurements are combined post hoc into a circuit, they do not capture higher-order interactions along full causal chains (paths) from input to output, and so cannot address how exponentially many paths interact to construct network computation. This leaves a large set of mechanistic questions unanswerable, for example whether inhibitory pathways veto excitatory ones, and how the two combine to yield network decisions. To address this gap, we introduce Complete Path Contribution, a method to decompose the network output into mathematically complete contributions of individual paths to study path-level interactions directly and at scale. Partial summation then induces definitions of contributions for any circuit at any granularity level, with completeness guaranteed at every level. We further present an efficient algorithm that identifies the $k$ strongest paths without examining all of them, with runtime empirically linear in $k$ rather than exponential in depth. By applying Path Contribution to benchmark image classification networks, we quantify excitatory-inhibitory interactions and show how they collectively construct network decisions at a level of detail previously unmeasurable. Furthermore, in an interpretable model of the mouse retina, it reveals how dynamic receptive fields emerge from excitatory-inhibitory interaction motifs among pathways between different cell types in response to time-varying natural stimuli. Overall, Path Contribution reveals how a network's causal pathways interact at scale to transform input to output.

## PDE-State Conditioning for Learned Optimization Using Physics-Informed Neural Networks

DevFeed: [PDE-State Conditioning for Learned Optimization Using Physics-Informed Neural Networks](<https://devfeed.tech/articles/pde-state-conditioning-for-learned-optimization-using-physics-informed-neural-networks-145019.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25185>)

Author: Farzana Yasmin, Ricardo Vilalta

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

Content type: article

Language: en

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

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

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [neural](<https://devfeed.tech/tags/neural.md>), [neural-networks](<https://devfeed.tech/tags/neural-networks.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [physics](<https://devfeed.tech/tags/physics.md>), [physics-informed](<https://devfeed.tech/tags/physics-informed.md>)

### AI overview

The study reports that conditioning a meta-learned PINN optimizer on PDE-level state improved all four evaluated metrics over Adam and an unconditioned learned optimizer on Linear Advection, KdV, and Burgers. The proposed state is 11-dimensional and captures operator structure, residual geometry, and short-horizon update response; bounded feature-wise modulation incorporates it into the optimizer. Among the tested designs, the compact residual-geometry state G2 and bounded FiLM performed strongest overall. These findings are limited to the reported evaluation problems and metrics.

### Source excerpt

arXiv:2611.25185v1 Announce Type: new Abstract: Physics-informed neural network (PINN) optimization depends on global properties of the underlying PDE and how the physics residual evolves across the domain. Standard optimizers such as Adam use fixed gradient-based statistics, while learned PINN optimizers replace the fixed rule with a learned update built from local parameters and optimization-history features. However, none explicitly use PDE-level information. We study two coupled design questions: what PDE-level information should a learned optimizer observe, and how should that information influence its update? We introduce an 11-dimensional PDE state-capturing operator structure, residual geometry, and short-horizon update response, and use it to condition a structured meta-learned optimizer through bounded feature-wise modulation. Repeated design studies identify the compact residual-geometry state G2 and bounded FiLM as the strongest overall choices among those tested. The resulting PDE-State-conditioned optimizer improves all evaluation metrics (relative-L2, L1, solution MSE, and physics MSE) over Adam and the unconditioned learned optimizer on Linear Advection, KdV, and Burgers.

## Bayesian optimization on amortized search space

DevFeed: [Bayesian optimization on amortized search space](<https://devfeed.tech/articles/bayesian-optimization-on-amortized-search-space-145184.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26459>)

Author: Tristan Fauvel, Federico Pavone, Titouan Guillou, Pau Redon Mu\\~noz, Chlo\\'e Geoffroy

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

Content type: article

Language: en

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

Topics: [Operations research and optimization](<https://devfeed.tech/topics/operations-research-and-optimization.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [bayesian](<https://devfeed.tech/tags/bayesian.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [search](<https://devfeed.tech/tags/search.md>), [space](<https://devfeed.tech/tags/space.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>)

### AI overview

The method uses an amortized inverse map of an approximate mechanistic simulator to constrain Bayesian optimization to a learned, low-dimensional search space, while a Gaussian process kernel retains awareness of the original problem geometry. The authors characterize approximation bias and information gain and derive a regret rate dependent on intrinsic dimension, under additional geometric and Gaussian process assumptions. In Parkinson's disease dosing personalization, a neural network is trained offline on a PK/PD model calibrated to population data with quantified parameter uncertainty; online optimization then searches the learned representation instead of raw dosages. Numerical simulations report improved performance in this application and under simulator misspecification.

### Source excerpt

arXiv:2611.26459v1 Announce Type: cross Abstract: In real-world black-box optimization problems, a mechanistic simulator is often available that approximately predicts the behavior of the system as a function of the parameters being optimized. Here, we show how an amortized inverse map of such a simulator can be used in Bayesian optimization to restrict the search space while using a Gaussian process kernel that is aware of the original problem geometry. This allows for efficiently leveraging the low intrinsic dimensionality of the optimization problem. We characterize its approximation bias and information gain, and derive an intrinsic-dimension-dependent regret rate under additional geometric and Gaussian Process assumptions. We illustrate the method with an application to drug dosing personalization in Parkinson's disease. We use, as a simulator, a pharmacokinetic/pharmacodynamic (PK/PD) model of a drug, a mechanistic model calibrated on population data with quantified parameter uncertainty. A dosing neural network is trained offline against the model, amortizing the map from patient PK/PD parameters to the regimen predicted to be optimal, and online Bayesian optimization then searches over this learned, low-dimensional representation rather than over raw dosages. Numerical simulations demonstrate the improved performance of our method in this application, as well as under simulator misspecification.

## SteerTrace: Tracing How Target Attacks Steer World Action Models

DevFeed: [SteerTrace: Tracing How Target Attacks Steer World Action Models](<https://devfeed.tech/articles/steertrace-tracing-how-target-attacks-steer-world-action-models-145575.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25719>)

Author: Yukun Dai, Tianshi Wang, Fengling Li, Jingjing Li, Xin Li, Lei Zhu

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

Content type: article

Language: en

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

Topics: [World models](<https://devfeed.tech/topics/world-models.md>), [Computer vision](<https://devfeed.tech/topics/computer-vision.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Physical AI](<https://devfeed.tech/topics/physical-ai.md>)

Tags: [a-safety](<https://devfeed.tech/tags/a-safety.md>), [actions](<https://devfeed.tech/tags/actions.md>), [attack](<https://devfeed.tech/tags/attack.md>), [attacks](<https://devfeed.tech/tags/attacks.md>), [bounded](<https://devfeed.tech/tags/bounded.md>), [budget](<https://devfeed.tech/tags/budget.md>), [camera](<https://devfeed.tech/tags/camera.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [effect](<https://devfeed.tech/tags/effect.md>), [experts](<https://devfeed.tech/tags/experts.md>), [ffn](<https://devfeed.tech/tags/ffn.md>), [frame](<https://devfeed.tech/tags/frame.md>), [future](<https://devfeed.tech/tags/future.md>), [inside](<https://devfeed.tech/tags/inside.md>), [layers](<https://devfeed.tech/tags/layers.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [noise](<https://devfeed.tech/tags/noise.md>), [target](<https://devfeed.tech/tags/target.md>), [tracing](<https://devfeed.tech/tags/tracing.md>), [world](<https://devfeed.tech/tags/world.md>)

### AI overview

SteerTrace reports that a single per-model perturbation bounded at 4/255, shared across frames, tasks, and camera views, steers actions toward an attacker-chosen target across eight world action models on LIBERO and RoboTwin, bringing task success close to zero. Same-budget noise and BadWAM cause little shift toward the target. Tracing the effect shows different dependencies: under the standard attack, restoring future tokens removes 61-94% of the targeted effect in five models whose action experts attend to them. When predicted futures are kept close to clean predictions, current-frame tokens become more important in three models: restoring them removes 87-99% of the effect, versus 9-15% for future tokens. In six models, restoring 0.36% of video-expert FFN neurons in layers 10-29 removes 72-96% of the effect. The abstract also reports that residual directions 2-16 contain only 11-18% of shift energy, yet removing them cancels most of the effect.

### Source excerpt

arXiv:2611.25719v1 Announce Type: cross Abstract: World action models (WAMs) predict future videos and actions with one backbone, and their imagined future is often seen as a safety signal. We introduce SteerTrace, a targeted universal attack on WAMs that also traces its effect inside the model. One $\ell_\infty$-bounded (4/255) perturbation per model, shared across frames, tasks, and camera views, pushes the executed action toward an attacker-chosen target. On eight WAMs from four families (FastWAM, Faster-WAM, OpenWAM and LingBot-VA) on LIBERO and RoboTwin, it drives task success close to zero and moves the attacked action toward the target. Same-budget noise and BadWAM produce little action shift in the target direction. We use the signed targeted effect to measure how much of the effect each intervention removes or reproduces. Under the standard attack, restoring future tokens removes 61-94% of the targeted effect in the five models whose action experts attend to them. When the attack is constrained to keep the predicted future close to the clean prediction, the targeted effect can become more dependent on current-frame tokens. In three of these five models, restoring current-frame tokens removes 87-99% of the effect, compared with 9-15% for future tokens. In six models, restoring just 0.36% of the FFN neurons in layers 10-29 of the video expert removes 72-96% of the targeted effect. Residual directions 2-16 hold only 11-18% of the shift energy, yet removing them cancels most of the targeted effect.

## Learning to Verify: Signature-Guided Training for Adaptive Adversarial Detection

DevFeed: [Learning to Verify: Signature-Guided Training for Adaptive Adversarial Detection](<https://devfeed.tech/articles/learning-to-verify-signature-guided-training-for-adaptive-adversarial-detection-145031.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25271>)

Author: Arash Vashagh, Roozbeh Razavi-Far

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

Content type: article

Language: en

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

Topics: [Adversarial attacks](<https://devfeed.tech/topics/adversarial-attacks.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [ablation](<https://devfeed.tech/tags/ablation.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [adversarial](<https://devfeed.tech/tags/adversarial.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [detection](<https://devfeed.tech/tags/detection.md>), [learning](<https://devfeed.tech/tags/learning.md>), [training](<https://devfeed.tech/tags/training.md>), [verify](<https://devfeed.tech/tags/verify.md>)

### AI overview

The approach keeps the classifier fixed and trains a separate verifier to assess its predictions. During training, class-specific activation signatures guide the creation of difficult adversarial examples; knowledge distillation helps the verifier retain performance on clean inputs. At inference, it checks predictions without using the signatures. On the evaluated datasets, it substantially outperformed confidence-based and feature-based detectors against the tested adaptive L-infinity attacks, including attacks unseen in training. Ablations attribute most of the robustness gain to adaptive verifier training, while signature guidance mainly helps clean accuracy. Additional L2 evaluations expose limits to robustness across perturbation types.

### Source excerpt

arXiv:2611.25271v1 Announce Type: new Abstract: Adversarial example detectors often fail when attackers explicitly attempt to bypass them. We propose a dual-model approach that keeps the original classifier fixed and trains a separate verifier to assess its predictions. During training, we identify patterns of internal neural network activations that distinguish genuine examples of each class from adversarial examples targeting that class. These patterns, called activation signatures, guide the generation of challenging adversarial examples that mislead the classifier, challenge the verifier, and internally resemble the target class. The verifier learns to recognize these examples while preserving its performance on clean inputs through knowledge distillation. At inference, the verifier evaluates the classifier's prediction without requiring activation signatures. We evaluate the approach on publicly available datasets against several adaptive attacks, including attacks unseen during training. The proposed verifier substantially outperforms directly attacked confidence-based and feature-based detectors under the evaluated L-infinity attacks. Ablation studies show that adaptive verifier training is the main source of this improvement, while signature guidance primarily benefits clean accuracy. Additional L2 evaluations reveal limitations in robustness across perturbation types.

## Safeguarded Learning-Assisted Nonlinear Conjugate Gradient Methods

DevFeed: [Safeguarded Learning-Assisted Nonlinear Conjugate Gradient Methods](<https://devfeed.tech/articles/safeguarded-learning-assisted-nonlinear-conjugate-gradient-methods-145093.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25707>)

Author: Morteza Kimiaei, Saman Babaie--Kafaki

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

Content type: article

Language: en

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

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [AI Development](<https://devfeed.tech/topics/ai-development.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [example](<https://devfeed.tech/tags/example.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [learning](<https://devfeed.tech/tags/learning.md>), [lightweight](<https://devfeed.tech/tags/lightweight.md>), [machine](<https://devfeed.tech/tags/machine.md>), [math-oc](<https://devfeed.tech/tags/math-oc.md>), [online-learning](<https://devfeed.tech/tags/online-learning.md>), [optimization](<https://devfeed.tech/tags/optimization.md>), [performance](<https://devfeed.tech/tags/performance.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [problems](<https://devfeed.tech/tags/problems.md>), [reduce](<https://devfeed.tech/tags/reduce.md>), [robustness](<https://devfeed.tech/tags/robustness.md>), [storage](<https://devfeed.tech/tags/storage.md>), [time](<https://devfeed.tech/tags/time.md>)

### AI overview

The framework adds two online extreme learning machine models to classical nonlinear conjugate gradient methods while retaining their standard two-term structure. One predicts a coefficient during restart recovery or persistent stagnation; the other proposes a bounded correction to the fallback step size after a finite-precision line-search failure. Predictions are accepted only if they meet descent, scale, and out-of-distribution safeguards; otherwise deterministic recovery remains in place. The models train on trajectory data without extra function or gradient evaluations. Across six variants tested on 480 CUTEst problems, the abstract reports improved robustness at broadly comparable evaluation costs. A conductivity-identification example also suggests potential savings in evaluations and computation time when evaluations are expensive.

### Source excerpt

arXiv:2611.25707v1 Announce Type: cross Abstract: A learning-assisted framework for Nonlinear Conjugate Gradient (NCG) methods is developed in which lightweight online prediction is used to support, rather than replace, the classical optimization process. The proposed approach retains the standard two-term NCG structure and introduces two independent online extreme learning machine models. The first model predicts an NCG coefficient during restart recovery or when persistent stagnation is detected, whereas the second model is used only after a line-search failure in finite precision arithmetic to propose a bounded correction to the deterministic fallback step size. Neural predictions are accepted only when they satisfy deterministic numerical safeguards, including descent, scale, and out-of-distribution requirements; otherwise, the corresponding deterministic recovery or fallback mechanism is retained. Both learning models are updated online using quantities already generated along the optimization trajectory and therefore require no additional function or gradient evaluations for training. These learning frameworks are applied to six well-known classical NCG methods, producing corresponding low-storage learning-assisted variants with deterministic safeguards. Numerical experiments on 480 CUTEst test problems show improved robustness for all six variants at broadly comparable evaluation costs. A PDE-constrained conductivity-identification example further illustrates the potential of the proposed approach to reduce evaluation costs and computational time for problems with expensive function and gradient evaluations. These findings suggest that restricted online learning can improve the robustness and computational performance of classical NCG methods while preserving their low-storage structure and deterministic optimization safeguards, and is particularly promising when function or gradient evaluations are expensive.

## Infer3D: Inverse Generative Modeling Enables Robust 3D Reconstruction

DevFeed: [Infer3D: Inverse Generative Modeling Enables Robust 3D Reconstruction](<https://devfeed.tech/articles/infer3d-inverse-generative-modeling-enables-robust-3d-reconstruction-145004.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25061>)

Author: Amani Kiruga, Mark Hamilton, Joshua B. Tenenbaum, William T. Freeman, Vincent Sitzmann, Yilun Du, Ayush Tewari

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

Content type: article

Language: en

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

Topics: [3D](<https://devfeed.tech/topics/3d.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Inference](<https://devfeed.tech/topics/inference.md>), [Image](<https://devfeed.tech/topics/image.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [3d-reconstruction](<https://devfeed.tech/tags/3d-reconstruction.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [generative](<https://devfeed.tech/tags/generative.md>), [modeling](<https://devfeed.tech/tags/modeling.md>), [ood](<https://devfeed.tech/tags/ood.md>), [robust](<https://devfeed.tech/tags/robust.md>)

### AI overview

Infer3D reconstructs 3D scenes from a single image by optimizing the latent variables of a generative model to match the input, rather than relying only on fast feedforward inference. A frozen 3D lifting network constrains optimization to a valid, multi-view-consistent 3D manifold to prevent flat outputs. On controlled synthetic and real-world tasks, the authors report better out-of-distribution reconstruction accuracy than feedforward networks trained on equal or larger datasets. Its adaptive strategy spends more test-time compute on ambiguous out-of-distribution examples while routing in-distribution inputs efficiently; the abstract does not quantify this compute or accuracy tradeoff.

### Source excerpt

arXiv:2611.25061v1 Announce Type: cross Abstract: Humans can understand the 3D structure of scenes even from highly unfamiliar or extreme viewpoints. Cognitive science suggests this generalization stems from a top-down generative inference strategy. In contrast, state-of-the-art 3D reconstruction neural networks typically rely on amortized, feedforward inference. While fast, this pattern-matching approach degrades when faced with out-of-distribution (OOD) scene conditions unseen during training. Inspired by the generative nature of human vision, we propose Infer3D, an inverse generative modeling framework that recovers the latent variables of a generative model to best match the monocular input. To prevent degenerate flat outputs, we couple the generative image prior with a frozen 3D lifting network that constrains the optimization space to a valid, multi-view-consistent 3D manifold. Across a variety of controlled synthetic and real-world tasks, Infer3D demonstrates superior OOD reconstruction accuracy compared to feedforward networks trained on equal or larger datasets. Furthermore, mirroring how human perception requires more time to process confusing stimuli, we introduce an adaptive inference strategy that expends more test-time compute on ambiguous OOD examples while efficiently routing in-distribution inputs. Project website: https://amanikiruga.github.io/infer3d-site/

## AdsNet: A Graph Neural Network for Direct Adsorption Energy Prediction

DevFeed: [AdsNet: A Graph Neural Network for Direct Adsorption Energy Prediction](<https://devfeed.tech/articles/adsnet-a-graph-neural-network-for-direct-adsorption-energy-prediction-145085.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25646>)

Author: Changzhi Ai, Frank Abild-Pedersen, Johannes Voss

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

Content type: article

Language: en

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

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [graph theory](<https://devfeed.tech/topics/graph-theory.md>), [3D](<https://devfeed.tech/topics/3d.md>), [Hyperparameter optimization](<https://devfeed.tech/topics/hyperparameter-optimization.md>), [concept-learning](<https://devfeed.tech/topics/concept-learning.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [algorithm](<https://devfeed.tech/tags/algorithm.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [automatic](<https://devfeed.tech/tags/automatic.md>), [catalyst](<https://devfeed.tech/tags/catalyst.md>), [complexity](<https://devfeed.tech/tags/complexity.md>), [computational-complexity](<https://devfeed.tech/tags/computational-complexity.md>), [cond-mat-mtrl-sci](<https://devfeed.tech/tags/cond-mat-mtrl-sci.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [contributions](<https://devfeed.tech/tags/contributions.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [dft](<https://devfeed.tech/tags/dft.md>), [energy](<https://devfeed.tech/tags/energy.md>), [fast](<https://devfeed.tech/tags/fast.md>), [framework](<https://devfeed.tech/tags/framework.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [gradient-descent](<https://devfeed.tech/tags/gradient-descent.md>), [graph](<https://devfeed.tech/tags/graph.md>), [neural](<https://devfeed.tech/tags/neural.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [physics-comp-ph](<https://devfeed.tech/tags/physics-comp-ph.md>), [prediction](<https://devfeed.tech/tags/prediction.md>)

### AI overview

AdsNet predicts adsorption energies in one forward pass using an equivariant message-passing network over adsorbate atoms and nearby surface atoms within a cutoff. Its local graph uses distances and directional vectors, so computation is decoupled from total system size while retaining sensitivity to local geometry. On 33,920 DFT structures covering 26 transition metals and six adsorbates, it reports an energy RMSE of 0.032 eV and force RMSE of 0.047 eV/Å. Automatically differentiated forces support projected-gradient relaxation with adaptive line search, and predicted energies can be decomposed into individual local-graph atom contributions. These results are reported for the described dataset; the excerpt does not provide further validation details.

### Source excerpt

arXiv:2611.25646v1 Announce Type: cross Abstract: Fast and accurate prediction of adsorption energies is critical for high-throughput computational catalyst screening, existing methods compromise either on evaluation speed (full-graph machine learning potentials) or geometric sensitivity (coordination-based models like the alpha-scheme). Here, we present AdsNet, an equivariant graph neural network that directly predicts adsorption energies in a single forward pass, which operates on a local graph restricted only to adsorbate atoms and its immediate surface neighbors within a cutoff radius. This local graph formulation decouples computational complexity from total system size while encoding continuous geometric features (interatomic distances and directional vectors) enabling sensitivity to subtle variations in local bonding environments. AdsNet adopts a message-passing architecture with scalar and vector node features over multiple interaction layers. Trained on AdsTraj, a dataset of 33,920 density functional theory (DFT) calculated structures spanning 26 transition metals and 6 adsorbates (H*, C*, O*, CO*, OH*, HOCO*), the model achieves an adsorption energy root-mean-square error (RMSE) of 0.032 eV and a force RMSE of 0.047 eV/A. Analytic forces obtained via automatic differentiation enable structure relaxation through a projected gradient descent algorithm with adaptive line search, eliminating entire geometry optimizations. Additionally, AdsNet uniquely decomposes predicted adsorption energies into contributions from individual atoms within the local adsorbate graph - an interpretability feature unattainable with conventional machine-learning potentials. AdsNet thus establishes an accurate, highly efficient, and interpretable framework for large-scale adsorption energy prediction and local interface relaxation across various material surfaces.

## A General Harness for Protein Foundation Model Fitness Prediction

DevFeed: [A General Harness for Protein Foundation Model Fitness Prediction](<https://devfeed.tech/articles/a-general-harness-for-protein-foundation-model-fitness-prediction-145382.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.34654>)

Author: Yang Tan, Qijia Tian, Gangyu Sun, Bozitao Zhong, Mingchen Li, Yuanxi Yu, Nanqing Dong, Liang Hong

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

Content type: article

Language: en

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

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [alignment](<https://devfeed.tech/tags/alignment.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [correlation](<https://devfeed.tech/tags/correlation.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [fitness](<https://devfeed.tech/tags/fitness.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [foundation-model](<https://devfeed.tech/tags/foundation-model.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [function](<https://devfeed.tech/tags/function.md>), [general](<https://devfeed.tech/tags/general.md>), [harness](<https://devfeed.tech/tags/harness.md>), [model](<https://devfeed.tech/tags/model.md>), [model-agnostic](<https://devfeed.tech/tags/model-agnostic.md>), [models](<https://devfeed.tech/tags/models.md>), [multiple](<https://devfeed.tech/tags/multiple.md>), [mutation](<https://devfeed.tech/tags/mutation.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [protein](<https://devfeed.tech/tags/protein.md>), [q-bio-qm](<https://devfeed.tech/tags/q-bio-qm.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [score](<https://devfeed.tech/tags/score.md>)

### AI overview

VenusREM-Harness (VRH) improves protein fitness predictions by combining frozen protein foundation model scores with family-specific multiple sequence alignment evidence, weighted by model uncertainty, then correcting and shrinking scores using structural confidence and solvent exposure. It is model-agnostic and training-free. Across 1,211 assays and 3.1 million variants, all 71 tested configurations improved Spearman correlation on ProteinGym, VenusMutHub, and VenusViroHub, with an average gain of 0.073 and broad gains across five metrics. The VRH-based VenusREM2 ranked highest across the reported function, taxon, MSA-depth, and mutation-depth categories, scoring 0.556 average Spearman on ProteinGym, 0.038 above the prior best.

### Source excerpt

arXiv:2609.34654v2 Announce Type: replace Abstract: Accurate fitness prediction is central to protein engineering and understanding sequence-function relationships. With advances in deep learning, protein foundation models (PFMs) have become widely used for this task. Recent analyses, however, show that these models share preferences reflecting their training corpora, while unreliable inputs can further distort fitness predictions. Family-specific evolutionary evidence and structural context can help address these limitations by providing complementary constraints on model scores, motivating VenusREM-Harness (VRH), a general, model-agnostic, training-free Retrieval-Enhanced Mutation harness. It fuses frozen model scores with multiple sequence alignment (MSA) evidence according to model uncertainty, then applies gated background correction and score shrinkage based on structural confidence and solvent exposure. Across 1,211 assays and 3.1 million measured variants from ProteinGym, VenusMutHub, and the newly curated viral benchmark VenusViroHub, all 71 configurations improve Spearman correlation on all 3 benchmarks by 0.073 on average, with broad gains across 5 metrics. Extended analyses relate retrieval gains to model-MSA preference differences, assess domain-level gains and immune-escape cases, and quantify computational speedups. Built with VRH, VenusREM2 is the first to rank highest in all function, taxon, MSA-depth, and mutation-depth categories, with a ProteinGym Average Spearman of 0.556, 0.038 above the prior best.

## Continuous-time learning of probability distributions: a case study in a digital trial of young children with type 1 diabetes

DevFeed: [Continuous-time learning of probability distributions: a case study in a digital trial of young children with type 1 diabetes](<https://devfeed.tech/articles/continuous-time-learning-of-probability-distributions-a-case-study-in-a-digital-trial-of-young-children-with-type-1-diabetes-144882.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2603.24427>)

Author: Antonio \\'Alvarez-L\\'opez, Marcos Matabuena

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

Content type: article

Language: en

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

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [case](<https://devfeed.tech/tags/case.md>), [case-study](<https://devfeed.tech/tags/case-study.md>), [children](<https://devfeed.tech/tags/children.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [continuous-glucose-monitoring](<https://devfeed.tech/tags/continuous-glucose-monitoring.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [diabetes](<https://devfeed.tech/tags/diabetes.md>), [digital](<https://devfeed.tech/tags/digital.md>), [digital-health](<https://devfeed.tech/tags/digital-health.md>), [distributions](<https://devfeed.tech/tags/distributions.md>), [gaussian](<https://devfeed.tech/tags/gaussian.md>), [insulin-delivery](<https://devfeed.tech/tags/insulin-delivery.md>), [learning](<https://devfeed.tech/tags/learning.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [neural](<https://devfeed.tech/tags/neural.md>), [ode](<https://devfeed.tech/tags/ode.md>), [probability](<https://devfeed.tech/tags/probability.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>), [time](<https://devfeed.tech/tags/time.md>), [type](<https://devfeed.tech/tags/type.md>)

### AI overview

The paper models how glucose measurement distributions change continuously over time using CGM data recorded every five minutes. It represents each distribution as a Gaussian mixture and uses a neural ODE to evolve the mixture weights, fitting the model by matching distributions with maximum mean discrepancy. Applied to data from a 13-week trial of children with type 1 diabetes comparing t:slim X2 closed-loop insulin delivery with standard therapy, the method reportedly detects treatment-related improvements in glucose dynamics that conventional analytical approaches may miss. The supplied text does not provide quantitative results or details of the comparison.

### Source excerpt

arXiv:2603.24427v2 Announce Type: replace-cross Abstract: Understanding how biomarker distributions evolve over time is a central challenge in digital health and chronic disease monitoring. For example, in diabetes, changes in the distribution of glucose measurements can reveal patterns of disease progression and treatment response that conventional summary measures miss. Motivated by a 13-week clinical trial comparing the closed-loop insulin delivery system t:slim X2 with standard therapy in children with type 1 diabetes, we propose a probabilistic framework to model the continuous-time evolution of time-indexed distributions using continuous glucose monitoring (CGM) data collected every five minutes. We represent the glucose distribution as a Gaussian mixture, with time-varying mixture weights governed by a neural ODE. We estimate the model parameters using a distribution-matching criterion based on the maximum mean discrepancy. The resulting framework is interpretable, computationally efficient, and sensitive to subtle temporal distributional changes. In the CGM trial data analyzed here, the method detects treatment-related improvements in glucose dynamics that are difficult to capture with traditional analytical approaches.

## Deviations from the NTK Regime in CNNs under Joint Depth-Width Scaling

DevFeed: [Deviations from the NTK Regime in CNNs under Joint Depth-Width Scaling](<https://devfeed.tech/articles/deviations-from-the-ntk-regime-in-cnns-under-joint-depth-width-scaling-145079.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25624>)

Author: Maria Matveev, Pascal Esser, Gitta Kutyniok, Mariia Seleznova

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

Content type: article

Language: en

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

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [cnns](<https://devfeed.tech/tags/cnns.md>), [convolutional-neural-networks](<https://devfeed.tech/tags/convolutional-neural-networks.md>), [critical](<https://devfeed.tech/tags/critical.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dispersion](<https://devfeed.tech/tags/dispersion.md>), [infinite](<https://devfeed.tech/tags/infinite.md>), [kernel](<https://devfeed.tech/tags/kernel.md>), [rotation](<https://devfeed.tech/tags/rotation.md>), [scaling](<https://devfeed.tech/tags/scaling.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

Numerical experiments in CNNs find that NTK dispersion at initialization rises with the depth-to-width ratio L/C in critical and chaotic regimes, while remaining small in the ordered regime. At critical initialization, dispersion curves approximately collapse when plotted against L/C. During training, kernel rotation generally increases with L/C, including in the ordered regime. The results suggest L/C may govern deviations from the infinite-width limit in some regimes, while showing that initial NTK concentration and training-time stability scale differently.

### Source excerpt

arXiv:2611.25624v1 Announce Type: new Abstract: The concentration and training-time constancy of the neural tangent kernel (NTK) depend on the interplay between network width, depth, and initialization. While joint depth-width scaling has been studied in fully-connected, residual, and recurrent networks, its effects in convolutional neural networks (CNNs) remain less understood. We investigate these effects numerically in CNNs, examining NTK dispersion at initialization and kernel rotation during training. Let depth $L$ be the number of convolutional layers, and width $C$ the number of channels per layer. We find that NTK dispersion increases with $L/C$ in the critical and chaotic initialization regimes, but remains small in the ordered regime. At critical initialization, dispersion curves approximately collapse as a function of $L/C$, suggesting a nontrivial joint scaling regime. During training, kernel rotation generally increases with $L/C$, even in the ordered regime where initial dispersion is small. Together, these results suggest that $L/C$ may govern NTK deviations from the infinite-width limit in certain regimes, and that NTK concentration at initialization and stability during training exhibit different scaling behaviors.

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

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

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

Author: Bassam El Said, Adam Sobey

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

Content type: article

Language: en

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

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

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

### AI overview

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

### Source excerpt

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

## SpikingGamma: Temporally Precise Online SNN Training Through Smoothed Temporal Delays

DevFeed: [SpikingGamma: Temporally Precise Online SNN Training Through Smoothed Temporal Delays](<https://devfeed.tech/articles/spikinggamma-temporally-precise-online-snn-training-through-smoothed-temporal-delays-144872.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2602.01978>)

Author: Roel Koopman, Sebastian Otte, Sander Boht\\'e

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

Content type: article

Language: en

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

Topics: [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Green Software](<https://devfeed.tech/topics/green-software.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [backpropagation](<https://devfeed.tech/tags/backpropagation.md>), [bank](<https://devfeed.tech/tags/bank.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [bptt](<https://devfeed.tech/tags/bptt.md>), [complex](<https://devfeed.tech/tags/complex.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ne](<https://devfeed.tech/tags/cs-ne.md>), [delay](<https://devfeed.tech/tags/delay.md>), [delta](<https://devfeed.tech/tags/delta.md>), [energy](<https://devfeed.tech/tags/energy.md>), [event](<https://devfeed.tech/tags/event.md>), [event-driven](<https://devfeed.tech/tags/event-driven.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [latency](<https://devfeed.tech/tags/latency.md>), [learning](<https://devfeed.tech/tags/learning.md>), [online](<https://devfeed.tech/tags/online.md>), [snn](<https://devfeed.tech/tags/snn.md>), [temporal](<https://devfeed.tech/tags/temporal.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

SpikingGamma aims to make online training of feedforward spiking neural networks more stable at fine temporal resolution. Neurons keep a compact bank of smoothed-delay states and communicate using sigma-delta spike coding. The method backpropagates online through a continuous reconstruction signal, avoiding surrogate gradients through the discontinuous spike. The abstract reports stable temporal-pattern learning with minimal spiking, competitive accuracy on complex tasks and benchmarks, and robustness to model temporal resolution. It presents SpikingGamma as an alternative to recurrent SNNs trained with surrogate gradients and as a possible route to neuromorphic hardware; the supplied text gives no specific benchmark results or hardware measurements.

### Source excerpt

arXiv:2602.01978v2 Announce Type: replace-cross Abstract: Neuromorphic hardware implementations of Spiking Neural Networks (SNNs) promise energy-efficient, low-latency AI through sparse, event-driven computation. Yet, training SNNs under fine temporal discretization remains challenging, hindering low-latency responsiveness and the mapping of software-trained SNNs to efficient hardware. In current approaches, spiking neurons are modeled as self-recurrent units, embedded into recurrent networks to maintain state over time, and trained with BPTT or RTRL variants based on surrogate gradients. These methods scale poorly with temporal resolution, while online approximations exhibit instability for long sequences and tend to fail at capturing temporal patterns. To address these limitations, we develop SpikingGamma, an SNN architecture in which neurons maintain a compact bank of smoothed-delay states and communicate through sigma-delta spike-coding. We show that in feedforward networks, SpikingGamma supports direct online error backpropagation through the continuous reconstruction signal, avoiding surrogate gradients through the spike discontinuity. This enables stable learning of temporal patterns with minimal spiking and scales feedforward SNNs to complex tasks and benchmarks with competitive accuracy, all while remaining robust to the temporal resolution of the model. Our approach offers both an alternative to recurrent SNNs trained with surrogate gradients, and a novel route for mapping SNNs to neuromorphic hardware.

## Hierarchical Pooling for Sheaf Neural Networks

DevFeed: [Hierarchical Pooling for Sheaf Neural Networks](<https://devfeed.tech/articles/hierarchical-pooling-for-sheaf-neural-networks-144910.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.20932>)

Author: Dionisia Naddeo, Carlo Abate, Pietro Li\\\`o, Nicola Toschi, Filippo Maria Bianchi

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

Content type: article

Language: en

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

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [aggregation](<https://devfeed.tech/tags/aggregation.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [features](<https://devfeed.tech/tags/features.md>), [graph](<https://devfeed.tech/tags/graph.md>), [graph-neural-networks](<https://devfeed.tech/tags/graph-neural-networks.md>), [learning](<https://devfeed.tech/tags/learning.md>), [nodes](<https://devfeed.tech/tags/nodes.md>), [representation-learning](<https://devfeed.tech/tags/representation-learning.md>), [scale](<https://devfeed.tech/tags/scale.md>), [snn](<https://devfeed.tech/tags/snn.md>), [structure](<https://devfeed.tech/tags/structure.md>)

### AI overview

Existing sheaf neural networks can learn layer-specific diffusion, but keep operating on the same graph. Hierarchical Sheaf Pooling (HiSP) adds pooling: it groups fine-scale nodes, reduces their features using the fine-scale sheaf, builds a coarse graph, and learns a new sheaf at that resolution. Interleaving pooling with sheaf diffusion lets the network learn distinct diffusion dynamics across graph resolutions, supporting multiscale problems where aggregation changes both the entities and their relations.

### Source excerpt

arXiv:2606.20932v2 Announce Type: replace Abstract: Sheaf neural networks (SNNs) extend graph neural networks by operating on cellular sheaves, which assign to each node and edge its own vector space, along with linear restriction maps from node spaces to the spaces associated with their incident edges. These maps determine the sheaf Laplacian and can be learned at each layer, enabling layer-specific diffusion dynamics. Existing SNN architectures, however, always diffuse on the same graph, with restriction maps that may change across layers but continue to act on the original nodes and edges. We introduce Hierarchical Sheaf Pooling (HiSP), a pooling layer that enables hierarchical representation learning in SNNs. Following the Select-Reduce-Connect-Lift framework, HiSP groups fine-scale nodes, reduces their features using the fine-scale sheaf structure, and constructs the corresponding coarse graph. A new sheaf is then learned at the coarse resolution. By interleaving sheaf diffusion and pooling, HiSP enables hierarchical SNNs to learn distinct diffusion dynamics across layers and graph resolutions. This construction supports multiscale problems in which aggregation changes both the interacting entities and their relations.

## Learning with Local Search MCMC Layers

DevFeed: [Learning with Local Search MCMC Layers](<https://devfeed.tech/articles/learning-with-local-search-mcmc-layers-144853.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2505.14240>)

Author: Germain Vivier-Ardisson, Mathieu Blondel, Axel Parmentier

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

Content type: article

Language: en

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

Topics: [Combinatorial optimization](<https://devfeed.tech/topics/combinatorial-optimization.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Operations research and optimization](<https://devfeed.tech/topics/operations-research-and-optimization.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [layers](<https://devfeed.tech/tags/layers.md>), [learning](<https://devfeed.tech/tags/learning.md>), [local](<https://devfeed.tech/tags/local.md>), [mcmc](<https://devfeed.tech/tags/mcmc.md>), [search](<https://devfeed.tech/tags/search.md>)

### AI overview

The paper turns local search neighborhoods into proposal distributions for MCMC over feasible solutions, yielding differentiable, stochastic combinatorial layers and associated loss functions for learning with inexact solvers. The approach is motivated by NP-hard problems where exact solvers are costly; the abstract reports that replacing them with local search strongly reduces learning computation. Demonstrations include dynamic vehicle routing with time windows, binary vector and k-subset prediction, and decision-focused learning for multidimensional knapsack. The excerpt does not provide details on theoretical guarantees or comparative results.

### Source excerpt

arXiv:2505.14240v4 Announce Type: replace Abstract: Integrating combinatorial optimization layers into neural networks has recently attracted significant research interest. However, many existing approaches lack theoretical guarantees or fail to perform adequately when relying on inexact solvers. This is a critical limitation, as many operations research problems are NP-hard, often necessitating the use of neighborhood-based local search heuristics. In this paper, we introduce a principled approach for learning with such inexact solvers. Inspired by the connection between simulated annealing and Metropolis-Hastings, we transform the problem specific neighborhood systems used in local search heuristics into proposal distributions, implementing MCMC on the set of feasible solutions. This allows us to construct differentiable, stochastic combinatorial layers and associated loss functions. Replacing an exact solver with a local search strongly reduces the computational burden of learning on many applications. We demonstrate our approach on a dynamic vehicle routing problem with time windows, binary vector and k-subset prediction tasks, as well as a multi-dimensional knapsack decision-focused learning problem.

## Virasoro Symmetry in Neural Network Field Theories

DevFeed: [Virasoro Symmetry in Neural Network Field Theories](<https://devfeed.tech/articles/virasoro-symmetry-in-neural-network-field-theories-144869.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2512.24420>)

Author: Brandon Robinson

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

Content type: article

Language: en

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

Topics: [quantum mechanics](<https://devfeed.tech/topics/quantum-mechanics.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [boundary](<https://devfeed.tech/tags/boundary.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [fields](<https://devfeed.tech/tags/fields.md>), [hep-th](<https://devfeed.tech/tags/hep-th.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [stress](<https://devfeed.tech/tags/stress.md>), [symmetry](<https://devfeed.tech/tags/symmetry.md>)

### AI overview

The work constructs a central-charge-one Virasoro algebra from a neural free-field architecture. A logarithmic momentum law and smooth spectral window produce a weight-one current kernel; its radial-ordering discontinuity determines the oscillator pairing, and limiting current correlations match Fock vacuum correlations. The resulting Sugawara stress tensor and charged vertex insertions give the Virasoro algebra and its highest-weight sectors. Error bounds control finite-mode approximations, while a finite-width auxiliary Fock readout is defined when a Hermitian mixed block is positive. The abstract also describes extensions to fermionic and supersymmetric sectors, ghost systems, and conformal or superconformal gluing via shared reflected features. Numerical comparisons support the current pairing and field correlations; detailed analysis is in the appendices.

### Source excerpt

arXiv:2512.24420v4 Announce Type: replace-cross Abstract: We construct the Virasoro algebra from a neural free-field architecture. A logarithmic momentum law and a smooth spectral window give the weight-one current kernel, whose radial-ordering discontinuity fixes the oscillator pairing. The limiting neural current correlations agree with the Fock vacuum correlations. Their Sugawara stress tensor gives the Virasoro algebra with central charge one, and charged vertex insertions identify its highest-weight sectors. Bounds on regulator and feature errors control finite-mode approximations to this representation. At finite width, an explicit adjoint-compatible projection defines an auxiliary Fock readout when its Hermitian mixed block is positive. Cauchy features extend the construction to fermionic and supersymmetric sectors, including a Clifford zero mode in the Ramond representation, while first-order stress tensors give fermionic and bosonic ghost systems. Shared reflected features impose scalar boundary conditions in every realization and yield conformal and superconformal gluing in the limiting fields. Numerical comparisons support the current pairing and field correlations; their detailed analysis is collected in the appendices.

## From Data Alignment to Parameter Generation: Few-Shot Cross-City Traffic Prediction with Conditional Diffusion

DevFeed: [From Data Alignment to Parameter Generation: Few-Shot Cross-City Traffic Prediction with Conditional Diffusion](<https://devfeed.tech/articles/from-data-alignment-to-parameter-generation-few-shot-cross-city-traffic-prediction-with-conditional-diffusion-145073.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25592>)

Author: Shengchun Wang, Yuhang Lin, Tong Li, Pengyuan Zhou, Andy Chi Lok Tai, Wenjie Chen, Zhu Xiao, Zhu Han

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

Content type: article

Language: en

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

Topics: [few-shot](<https://devfeed.tech/topics/few-shot.md>), [Internet Traffic](<https://devfeed.tech/topics/internet-traffic.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>)

Tags: [cities](<https://devfeed.tech/tags/cities.md>), [clustering](<https://devfeed.tech/tags/clustering.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [few-shot](<https://devfeed.tech/tags/few-shot.md>), [imagery](<https://devfeed.tech/tags/imagery.md>), [large-language-model](<https://devfeed.tech/tags/large-language-model.md>), [node](<https://devfeed.tech/tags/node.md>), [parameter](<https://devfeed.tech/tags/parameter.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [region](<https://devfeed.tech/tags/region.md>), [robustness](<https://devfeed.tech/tags/robustness.md>), [traffic](<https://devfeed.tech/tags/traffic.md>)

### AI overview

CDPG addresses few-shot mobile traffic forecasting in data-scarce cities by generating base-station-specific predictor parameters instead of aligning raw traffic data across cities. It learns a conditional distribution of strong source-city predictor parameters and conditions generation on spatial context from satellite imagery and LLM-derived descriptions, plus temporal patterns identified through unsupervised clustering. The authors report that across three large-scale datasets CDPG outperforms cross-city few-shot baselines by up to 9.72% on standard forecasting metrics and remains robust for few-shot multi-step prediction. These findings are reported in the abstract; details of the metrics and experimental setup are not provided here.

### Source excerpt

arXiv:2611.25592v1 Announce Type: new Abstract: Accurate mobile traffic forecasting is crucial for network service management, supporting proactive capacity provisioning, congestion mitigation, and energy-efficient operation. However, many cities and regions are data-scarce (e.g., newly deployed or poorly instrumented areas), thus making traffic prediction a few-shot learning problem and limiting the practicality of standard data-driven models. Existing cross-city solutions primarily rely on data-space alignment to match traffic distributions between data-rich and data-scarce cities, but such alignment is often brittle under real-world distribution shifts, measurement noise, and city-specific bias. To address these limitations, we propose a Conditional Diffusion-based Parameter Generation (CDPG) framework. Instead of aligning raw data, CDPG learns the conditional distribution of high-performing predictor parameters for source cities and directly generates node-specific model parameters for the target city. To support reliable adaptation for network traffic prediction, CDPG conditions parameter generation on multi-modal spatio-temporal context. Specifically, CDPG captures spatial environment cues by combining satellite imagery with large language model (LLM)-derived semantic descriptions, and summarizes complex temporal behaviors via unsupervised pattern clustering. These contextual components guide CDPG in producing parameters tailored to each base station's operational context, thereby improving robustness under cross-city and cross-region shifts. Extensive experiments on three large-scale datasets show that CDPG consistently outperforms state-of-the-art cross-city few-shot baselines by up to 9.72% on standard forecasting metrics and remains robust for few-shot multi-step prediction. The code is available at https://github.com/LingoodBoy/CDPG.

## A Complete Geometric Message-Passing Architecture for Point Clouds

DevFeed: [A Complete Geometric Message-Passing Architecture for Point Clouds](<https://devfeed.tech/articles/a-complete-geometric-message-passing-architecture-for-point-clouds-145173.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26384>)

Author: Nadav Dym, Tal Amir, Gia Chi Dang, Christopher Morris

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

Content type: article

Language: en

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

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [3D](<https://devfeed.tech/topics/3d.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [complete](<https://devfeed.tech/tags/complete.md>), [completeness](<https://devfeed.tech/tags/completeness.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [learning](<https://devfeed.tech/tags/learning.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [message](<https://devfeed.tech/tags/message.md>), [point](<https://devfeed.tech/tags/point.md>), [theoretical](<https://devfeed.tech/tags/theoretical.md>)

### AI overview

The paper presents a simple equivariant message-passing architecture for 3D point clouds and proves it is complete: it can distinguish inputs unless they differ only by the rigid-motion and permutation symmetries considered. Experiments show it separates challenging point clouds that distance-based message passing provably cannot distinguish. The authors argue this expressivity may be achievable without computationally intensive high-order features, and report a completeness proof for a slightly modified SEGNN as initial evidence that the approach may extend to other equivariant architectures. The supplied text is the abstract.

### Source excerpt

arXiv:2611.26384v1 Announce Type: new Abstract: Three-dimensional point clouds are a fundamental data modality, and a large family of machine learning architectures has been developed to respect their natural symmetries to rigid motions and permutations. A principled way to compare these architectures, which vary widely in complexity and cost, is through their expressive power. That is, an architecture is complete if it can distinguish any two inputs not related by the symmetries it respects. Simple distance-based message-passing networks are known to be incomplete, and while there exist complete alternatives, they typically rely on computationally intensive mechanisms like high-order representations of the rotation group or higher-order Weisfeiler-Leman-type tests. In this work, we devise a novel, simple, equivariant message-passing architecture for point clouds, and prove that it is complete. We complement our theoretical result with experiments showing that our architecture separates challenging point clouds that distance-based message passing provably cannot distinguish. Our results suggest that high-order equivariant features may not be necessary to enhance the expressivity of point cloud architectures, and can help settle completeness for other equivariant architectures for which such results are currently unknown. We give initial evidence for this by proving completeness of a slightly modified version of SEGNN.

[Next page](<https://devfeed.tech/topics/neural-network.md?cursor=WyIyMDI2LTExLTIwVDA1OjAwOjAwKzAwOjAwIiwgIjgyNTA0MjU5LWZmY2UtNDQ4OC1hOTU3LTdhMmMwOTRlN2E2YSJd>)