# Research Papers — 2026-01 · bundle page 24 of 189

> Full-text markdown of papers #1151–1200 of 9403 published in 2026-01, newest first. AI-readable mirrors of the full 9403-paper 2026-01 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-01/page/24
[← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/23.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-01/page/25.md)

### Hardware-Triggered Backdoors

Authors: Jonas Möller, Erik Imgrund, Thorsten Eisenhofer, Konrad Rieck
Published: 2026-01-29
Categories: cs.LG, cs.CR
arXiv: https://arxiv.org/abs/2601.21902v2

Machine learning models are routinely deployed on a wide range of computing hardware. Although such hardware is typically expected to produce identical results, differences in its design can lead to small numerical variations during inference. In this work, we show that these variations can be exploited to create backdoors in machine learning models. The core idea is to shape the model's decision function such that it yields different predictions for the same input when executed on different hardware. This effect is achieved by locally moving the decision boundary close to a target input and then refining numerical deviations to flip the prediction on selected hardware. We empirically demonstrate that these hardware-triggered backdoors can be created reliably across common GPU accelerators. Our findings reveal a novel attack vector affecting the use of third-party models, and we investigate different defenses to counter this threat.

---

### TraceRouter: Robust Safety for Large Foundation Models via Path-Level Intervention

Authors: Chuancheng Shi, Shangze Li, Wenjun Lu, Wenhua Wu, Cong Wang, Zifeng Cheng, Fei Shen, Tat-Seng Chua
Published: 2026-01-29
Categories: cs.CV, cs.AI, cs.CY, cs.MM
arXiv: https://arxiv.org/abs/2601.21900v2

Despite their capabilities, large foundation models (LFMs) remain susceptible to adversarial manipulation. Current defenses predominantly rely on the "locality hypothesis", suppressing isolated neurons or features. However, harmful semantics act as distributed, cross-layer circuits, rendering such localized interventions brittle and detrimental to utility. To bridge this gap, we propose \textbf{TraceRouter}, a path-level framework that traces and disconnects the causal propagation circuits of illicit semantics. TraceRouter operates in three stages: (1) it pinpoints a sensitive onset layer by analyzing attention divergence; (2) it leverages sparse autoencoders (SAEs) and differential activation analysis to disentangle and isolate malicious features; and (3) it maps these features to downstream causal pathways via feature influence scores (FIS) derived from zero-out interventions. By selectively suppressing these causal chains, TraceRouter physically severs the flow of harmful information while leaving orthogonal computation routes intact. Extensive experiments demonstrate that TraceRouter significantly outperforms state-of-the-art baselines, achieving a superior trade-off between adversarial robustness and general utility. Our code will be publicly released. WARNING: This paper contains unsafe model responses.

---

### Breaking the Regional Barrier: Inductive Semantic Topology Learning for Worldwide Air Quality Forecasting

Authors: Zhiqing Cui, Siru Zhong, Ming Jin, Shirui Pan, Qingsong Wen, Yuxuan Liang
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21899v1

Global air quality forecasting grapples with extreme spatial heterogeneity and the poor generalization of existing transductive models to unseen regions. To tackle this, we propose OmniAir, a semantic topology learning framework tailored for global station-level prediction. By encoding invariant physical environmental attributes into generalizable station identities and dynamically constructing adaptive sparse topologies, our approach effectively captures long-range non-Euclidean correlations and physical diffusion patterns across unevenly distributed global networks. We further curate WorldAir, a massive dataset covering over 7,800 stations worldwide. Extensive experiments show that OmniAir achieves state-of-the-art performance against 18 baselines, maintaining high efficiency and scalability with speeds nearly 10 times faster than existing models, while effectively bridging the monitoring gap in data-sparse regions.

---

### Making Models Unmergeable via Scaling-Sensitive Loss Landscape

Authors: Minwoo Jang, Hoyoung Kim, Jabin Koo, Jungseul Ok
Published: 2026-01-29
Categories: cs.AI, cs.CR
arXiv: https://arxiv.org/abs/2601.21898v2

The rise of model hubs has made it easier to access reusable model components, making model merging a practical tool for combining capabilities. Yet, this modularity also creates a governance gap: downstream users can recompose released weights into unauthorized mixtures that bypass safety alignment or licensing terms. Because existing defenses are largely post-hoc and architecture-specific, they provide inconsistent protection across diverse architectures and release formats in practice. To close this gap, we propose Trap$^2$, an architecture-agnostic protection framework that encodes protection into updates during fine-tuning, regardless of whether they are released as adapters or full models. Instead of relying on architecture-dependent approaches, Trap$^2$ uses weight re-scaling as a simple proxy for the merging process. It keeps released weights effective in standalone use, but degrades them under re-scaling that often arises in merging, undermining unauthorized recomposition.

---

### A Low-Complexity Plug-and-Play Deep Learning Model for Generalizable Massive MIMO Precoding

Authors: Ali Hasanzadeh Karkan, Ahmed Ibrahim, Jean-François Frigon, François Leduc-Primeau
Published: 2026-01-29
Categories: cs.LG, eess.SP
arXiv: https://arxiv.org/abs/2601.21897v1

Massive multiple-input multiple-output (mMIMO) downlink precoding offers high spectral efficiency but remains challenging to deploy in practice because near-optimal algorithms such as the weighted minimum mean squared error (WMMSE) are computationally expensive, and sensitive to SNR and channel-estimation quality, while existing deep learning (DL)-based solutions often lack robustness and require retraining for each deployment site. This paper proposes a plug-and-play precoder (PaPP), a DL framework with a backbone that can be trained for either fully digital (FDP) or hybrid beamforming (HBF) precoding and reused across sites, transmit-power levels, and with varying amounts of channel estimation error, avoiding the need to train a new model from scratch at each deployment. PaPP combines a high-capacity teacher and a compact student with a self-supervised loss that balances teacher imitation and normalized sum-rate, trained using meta-learning domain-generalization and transmit-power-aware input normalization. Numerical results on ray-tracing data from three unseen sites show that the PaPP FDP and HBF models both outperform conventional and deep learning baselines, after fine-tuning with a small set of local unlabeled samples. Across both architectures, PaPP achieves more than 21$\times$ reduction in modeled computation energy and maintains good performance under channel-estimation errors, making it a practical solution for energy-efficient mMIMO precoding.

---

### Past- and Future-Informed KV Cache Policy with Salience Estimation in Autoregressive Video Diffusion

Authors: Hanmo Chen, Chenghao Xu, Xu Yang, Xuan Chen, Cheng Deng
Published: 2026-01-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.21896v3

Video generation is pivotal to digital media creation, and recent advances in autoregressive video generation have markedly enhanced the efficiency of real-time video synthesis. However, existing approaches generally rely on heuristic KV Cache policies, which ignore differences in token importance in long-term video generation. This leads to the loss of critical spatiotemporal information and the accumulation of redundant, invalid cache, thereby degrading video generation quality and efficiency. To address this limitation, we first observe that token contributions to video generation are highly time-heterogeneous and accordingly propose a novel Past- and Future-Informed KV Cache Policy (PaFu-KV). Specifically, PaFu-KV introduces a lightweight Salience Estimation Head distilled from a bidirectional teacher to estimate salience scores, allowing the KV cache to retain informative tokens while discarding less relevant ones. This policy yields a better quality-efficiency trade-off by shrinking KV cache capacity and reducing memory footprint at inference time. Extensive experiments on benchmarks demonstrate that our method preserves high-fidelity video generation quality while enables accelerated inference, thereby enabling more efficient long-horizon video generation. Our code will be released upon paper acceptance.

---

### Learn-to-Distance: Distance Learning for Detecting LLM-Generated Text

Authors: Hongyi Zhou, Jin Zhu, Kai Ye, Ying Yang, Erhan Xu, Chengchun Shi
Published: 2026-01-29
Categories: cs.CL, cs.AI, stat.ML
arXiv: https://arxiv.org/abs/2601.21895v2

Modern large language models (LLMs) such as GPT, Claude, and Gemini have transformed the way we learn, work, and communicate. Yet, their ability to produce highly human-like text raises serious concerns about misinformation and academic integrity, making it an urgent need for reliable algorithms to detect LLM-generated content. In this paper, we start by presenting a geometric approach to demystify rewrite-based detection algorithms, revealing their underlying rationale and demonstrating their generalization ability. Building on this insight, we introduce a novel rewrite-based detection algorithm that adaptively learns the distance between the original and rewritten text. Theoretically, we demonstrate that employing an adaptively learned distance function is more effective for detection than using a fixed distance. Empirically, we conduct extensive experiments with over 100 settings, and find that our approach demonstrates superior performance over baseline algorithms in the majority of scenarios. In particular, it achieves relative improvements from 54.3% to 75.4% over the strongest baseline across different target LLMs (e.g., GPT, Claude, and Gemini). A python implementation of our proposal is publicly available at https://github.com/Mamba413/L2D.

---

### Not All Code Is Equal: A Data-Centric Study of Code Complexity and LLM Reasoning

Authors: Lukas Twist, Shu Yang, Hanqi Yan, Jingzhi Gong, Di Wang, Helen Yannakoudakis, Jie M. Zhang
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21894v1

Large Language Models (LLMs) increasingly exhibit strong reasoning abilities, often attributed to their capacity to generate chain-of-thought-style intermediate reasoning. Recent work suggests that exposure to code can further enhance these skills, but existing studies largely treat code as a generic training signal, leaving open the question of which properties of code actually contribute to improved reasoning. To address this gap, we study the structural complexity of code, which captures control flow and compositional structure that may shape how models internalise multi-step reasoning during fine-tuning. We examine two complementary settings: solution-driven complexity, where complexity varies across multiple solutions to the same problem, and problem-driven complexity, where complexity reflects variation in the underlying tasks. Using cyclomatic complexity and logical lines of code to construct controlled fine-tuning datasets, we evaluate a range of open-weight LLMs on diverse reasoning benchmarks. Our findings show that although code can improve reasoning, structural properties strongly determine its usefulness. In 83% of experiments, restricting fine-tuning data to a specific structural complexity range outperforms training on structurally diverse code, pointing to a data-centric path for improving reasoning beyond scaling.

---

### Improving Classifier-Free Guidance of Flow Matching via Manifold Projection

Authors: Jian-Feng Cai, Haixia Liu, Zhengyi Su, Chao Wang
Published: 2026-01-29
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.21892v2

Classifier-free guidance (CFG) is a widely used technique for controllable generation in diffusion and flow-based models. Despite its empirical success, CFG relies on a heuristic linear extrapolation that is often sensitive to the guidance scale. In this work, we provide a principled interpretation of CFG through the lens of optimization. We demonstrate that the velocity field in flow matching corresponds to the gradient of a sequence of smoothed distance functions, which guides latent variables toward the scaled target image set. This perspective reveals that the standard CFG formulation is an approximation of this gradient, where the prediction gap, the discrepancy between conditional and unconditional outputs, governs guidance sensitivity. Leveraging this insight, we reformulate the CFG sampling as a homotopy optimization with a manifold constraint. This formulation necessitates a manifold projection step, which we implement via an incremental gradient descent scheme during sampling. To improve computational efficiency and stability, we further enhance this iterative process with Anderson Acceleration without requiring additional model evaluations. Our proposed methods are training-free and consistently refine generation fidelity, prompt alignment, and robustness to the guidance scale. We validate their effectiveness across diverse benchmarks, demonstrating significant improvements on large-scale models such as DiT-XL-2-256, Flux, and Stable Diffusion 3.5.

---

### VSE: Variational state estimation of complex model-free process

Authors: Gustav Norén, Anubhab Ghosh, Fredrik Cumlin, Saikat Chatterjee
Published: 2026-01-29
Categories: eess.SP, cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2601.21887v1

We design a variational state estimation (VSE) method that provides a closed-form Gaussian posterior of an underlying complex dynamical process from (noisy) nonlinear measurements. The complex process is model-free. That is, we do not have a suitable physics-based model characterizing the temporal evolution of the process state. The closed-form Gaussian posterior is provided by a recurrent neural network (RNN). The use of RNN is computationally simple in the inference phase. For learning the RNN, an additional RNN is used in the learning phase. Both RNNs help each other learn better based on variational inference principles. The VSE is demonstrated for a tracking application - state estimation of a stochastic Lorenz system (a benchmark process) using a 2-D camera measurement model. The VSE is shown to be competitive against a particle filter that knows the Lorenz system model and a recently proposed data-driven state estimation method that does not know the Lorenz system model.

---

### Adaptive Surrogate-Based Strategy for Accelerating Convergence Speed when Solving Expensive Unconstrained Multi-Objective Optimisation Problems

Authors: Tiwonge Msulira Banda, Alexandru-Ciprian Zăvoianu
Published: 2026-01-29
Categories: cs.NE
arXiv: https://arxiv.org/abs/2601.21885v1

Multi-Objective Evolutionary Algorithms (MOEAs) have proven effective at solving Multi-Objective Optimisation Problems (MOOPs). However, their performance can be significantly hindered when applied to computationally intensive industrial problems. To address this limitation, we propose an adaptive surrogate modelling approach designed to accelerate the early-stage convergence speed of state-of-the-art MOEAs. This is important because it ensures that a solver can identify optimal or near-optimal solutions with relatively few fitness function evaluations, thereby saving both time and computational resources. Our method employs a two-loop architecture. The outer loop runs a (baseline) host MOEA which carries out true fitness evaluations. The inner loop contains an Adaptive Accelerator that leverages data-driven machine learning (ML) surrogate models to approximate fitness functions. Integrated with NSGA-II and MOEA/D, our approach was tested on 31 widely known benchmark problems and a real-world North Sea fish abundance modelling case study. The results demonstrate that by incorporating Gaussian Process Regression, one-dimensional Convolutional Neural Networks, and Random Forest Regression, our proposed approach significantly accelerates the convergence speed of MOEAs in the early phases of optimisation.

---

### Managing Solution Stability in Decision-Focused Learning with Cost Regularization

Authors: Victor Spitzer, Francois Sanson
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21883v1

Decision-focused learning integrates predictive modeling and combinatorial optimization by training models to directly improve decision quality rather than prediction accuracy alone. Differentiating through combinatorial optimization problems represents a central challenge, and recent approaches tackle this difficulty by introducing perturbation-based approximations. In this work, we focus on estimating the objective function coefficients of a combinatorial optimization problem. Our study demonstrates that fluctuations in perturbation intensity occurring during the learning phase can lead to ineffective training, by establishing a theoretical link to the notion of solution stability in combinatorial optimization. We propose addressing this issue by introducing a regularization of the estimated cost vectors which improves the robustness and reliability of the learning process, as demonstrated by extensive numerical experiments.

---

### How Expressive Are Graph Neural Networks in the Presence of Node Identifiers?

Authors: Arie Soeteman, Michael Benedikt, Martin Grohe, Balder ten Cate
Published: 2026-01-29
Categories: cs.LO, cs.LG
arXiv: https://arxiv.org/abs/2601.21882v1

Graph neural networks (GNNs) are a widely used class of machine learning models for graph-structured data, based on local aggregation over neighbors. GNNs have close connections to logic. In particular, their expressive power is linked to that of modal logics and bounded-variable logics with counting. In many practical scenarios, graphs processed by GNNs have node features that act as unique identifiers. In this work, we study how such identifiers affect the expressive power of GNNs. We initiate a study of the key-invariant expressive power of GNNs, inspired by the notion of order-invariant definability in finite model theory: which node queries that depend only on the underlying graph structure can GNNs express on graphs with unique node identifiers? We provide answers for various classes of GNNs with local max- or sum-aggregation.

---

### astra-langchain4j: Experiences Combining LLMs and Agent Programming

Authors: Rem Collier, Katharine Beaumont, Andrei Ciortea
Published: 2026-01-29
Categories: cs.AI, cs.MA
arXiv: https://arxiv.org/abs/2601.21879v1

Given the emergence of Generative AI over the last two years and the increasing focus on Agentic AI as a form of Multi-Agent System it is important to explore both how such technologies can impact the use of traditional Agent Toolkits and how the wealth of experience encapsulated in those toolkits can influence the design of the new agentic platforms. This paper presents an overview of our experience developing a prototype large language model (LLM) integration for the ASTRA programming language. It presents a brief overview of the toolkit, followed by three example implementations, concluding with a discussion of the experiences garnered through the examples.

---

### Evolution of Benchmark: Black-Box Optimization Benchmark Design through Large Language Model

Authors: Chen Wang, Sijie Ma, Zeyuan Ma, Yue-Jiao Gong
Published: 2026-01-29
Categories: cs.NE
arXiv: https://arxiv.org/abs/2601.21877v2

Benchmark Design in Black-Box Optimization (BBO) is a fundamental yet open-ended topic. Early BBO benchmarks are predominantly human-crafted, introducing expert bias and constraining diversity. Automating this design process can relieve the human-in-the-loop burden while enhancing diversity and objectivity. We propose Evolution of Benchmark (EoB), an automated BBO benchmark designer empowered by the large language model (LLM) and its program evolution capability. Specifically, we formulate benchmark design as a bi-objective optimization problem towards maximizing (i) landscape diversity and (ii) algorithm-differentiation ability across a portfolio of BBO solvers. Under this paradigm, EoB iteratively prompts LLM to evolve a population of benchmark programs and employs a reflection-based scheme to co-evolve the landscape and its corresponding program. Comprehensive experiments validate our EoB is a competitive candidate in multi-dimensional usages: 1) Benchmarking BBO algorithms; 2) Training and testing learning-assisted BBO algorithms; 3) Extending proxy for expensive real-world problems.

---

### Low-Rank Plus Sparse Matrix Transfer Learning under Growing Representations and Ambient Dimensions

Authors: Jinhang Chai, Xuyuan Liu, Elynn Chen, Yujun Yan
Published: 2026-01-29
Categories: cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2601.21873v2

Learning systems often expand their ambient features or latent representations over time, embedding earlier representations into larger spaces with limited new latent structure. We study transfer learning for structured matrix estimation under simultaneous growth of the ambient dimension and the intrinsic representation, where a well-estimated source task is embedded as a subspace of a higher-dimensional target task. We propose a general transfer framework in which the target parameter decomposes into an embedded source component, low-dimensional low-rank innovations, and sparse edits, and develop an anchored alternating projection estimator that preserves transferred subspaces while estimating only low-dimensional innovations and sparse modifications. We establish deterministic error bounds that separate target noise, representation growth, and source estimation error, yielding strictly improved rates when rank and sparsity increments are small. We demonstrate the generality of the framework by applying it to two canonical problems. For Markov transition matrix estimation from a single trajectory, we derive end-to-end theoretical guarantees under dependent noise. For structured covariance estimation under enlarged dimensions, we provide complementary theoretical analysis in the appendix and empirically validate consistent transfer gains.

---

### WebArbiter: A Principle-Guided Reasoning Process Reward Model for Web Agents

Authors: Yao Zhang, Shijie Tang, Zeyu Li, Zhen Han, Volker Tresp
Published: 2026-01-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.21872v2

Web agents hold great potential for automating complex computer tasks, yet their interactions involve long-horizon, sequential decision-making with irreversible actions. In such settings, outcome-based supervision is sparse and delayed, often rewarding incorrect trajectories and failing to support inference-time scaling. This motivates the use of Process Reward Models (WebPRMs) for web navigation, but existing approaches remain limited: scalar WebPRMs collapse progress into coarse, weakly grounded signals, while checklist-based WebPRMs rely on brittle template matching that fails under layout or semantic changes and often mislabels superficially correct actions as successful, providing little insight or interpretability. To address these challenges, we introduce WebArbiter, a reasoning-first, principle-inducing WebPRM that formulates reward modeling as text generation, producing structured justifications that conclude with a preference verdict and identify the action most conducive to task completion under the current context. Training follows a two-stage pipeline: reasoning distillation equips the model with coherent principle-guided reasoning, and reinforcement learning corrects teacher biases by directly aligning verdicts with correctness, enabling stronger generalization. To support systematic evaluation, we release WebPRMBench, a comprehensive benchmark spanning four diverse web environments with rich tasks and high-quality preference annotations. On WebPRMBench, WebArbiter-7B outperforms the strongest baseline, GPT-5, by 9.1 points. In reward-guided trajectory search on WebArena-Lite, it surpasses the best prior WebPRM by up to 6.4 points, underscoring its robustness and practical value in complex web tasks.

---

### On Forgetting and Stability of Score-based Generative models

Authors: Stanislas Strasman, Gabriel Cardoso, Sylvain Le Corff, Vincent Lemaire, Antonio Ocello
Published: 2026-01-29
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2601.21868v2

Understanding the stability and long-time behavior of generative models is a fundamental problem in modern machine learning. This paper provides quantitative bounds on the sampling error of score-based generative models by leveraging stability and forgetting properties of the Markov chain associated with the reverse-time dynamics. Under weak assumptions, we provide the two structural properties to ensure the propagation of initialization and discretization errors of the backward process: a Lyapunov drift condition and a Doeblin-type minorization condition. A practical consequence is quantitative stability of the sampling procedure, as the reverse diffusion dynamics induces a contraction mechanism along the sampling trajectory. Our results clarify the role of stochastic dynamics in score-based models and provide a principled framework for analyzing propagation of errors in such approaches.

---

### MoHETS: Long-term Time Series Forecasting with Mixture-of-Heterogeneous-Experts

Authors: Evandro S. Ortigossa, Guy Lutsker, Eran Segal
Published: 2026-01-29
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.21866v2

Real-world multivariate time series can exhibit intricate multi-scale structures, including global trends, local periodicities, and non-stationary regimes, which makes long-horizon forecasting challenging. Although sparse Mixture-of-Experts (MoE) approaches improve scalability and specialization, they typically rely on homogeneous MLP experts that poorly capture the diverse temporal dynamics of time series data. We address these limitations with MoHETS, an encoder-only Transformer that integrates sparse Mixture-of-Heterogeneous-Experts (MoHE) layers. MoHE routes temporal patches to a small subset of expert networks, combining a shared depthwise-convolution expert for sequence-level continuity with routed Fourier-based experts for patch-level periodic structures. MoHETS further improves robustness to non-stationary dynamics by incorporating exogenous information via cross-attention over covariate patch embeddings. Finally, we replace parameter-heavy linear projection heads with a lightweight convolutional patch decoder, improving parameter efficiency, reducing training instability, and allowing a single model to generalize across arbitrary forecast horizons. We validate across seven multivariate benchmarks and multiple horizons, with MoHETS consistently achieving state-of-the-art performance, reducing the average MSE by $12\%$ compared to strong recent baselines, demonstrating effective heterogeneous specialization for long-term forecasting.

---

### Knowing Bias, Doing Better: Mitigating Social Bias in LLMs via Know-Bias Neuron Enhancement

Authors: Jinhao Pan, Chahat Raj, Anjishnu Mukherjee, Sina Mansouri, Bowen Wei, Shloka Yada, Ziwei Zhu
Published: 2026-01-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.21864v2

Large language models (LLMs) exhibit social biases that reinforce harmful stereotypes, limiting their safe deployment. Most existing debiasing methods adopt a suppressive paradigm by modifying parameters, prompts, or neurons associated with biased behavior; however, such approaches are often brittle, weakly generalizable, data-inefficient, and prone to degrading general capability. We propose \textbf{KnowBias}, a lightweight and conceptually distinct framework that mitigates bias by strengthening, rather than suppressing, neurons encoding bias-knowledge. KnowBias identifies neurons encoding bias knowledge using a small set of bias-knowledge questions via attribution-based analysis, and selectively enhances them at inference time. This design enables strong debiasing while preserving general capabilities, generalizes across bias types and demographics, and is highly data efficient, requiring only a handful of simple yes/no questions and no retraining. Experiments across multiple benchmarks and LLMs demonstrate consistent state-of-the-art debiasing performance with minimal utility degradation. Data and code are available at https://github.com/JP-25/KnowBias.

---

### Pathwise Learning of Stochastic Dynamical Systems with Partial Observations

Authors: Nicole Tianjiao Yang
Published: 2026-01-29
Categories: math.OC, stat.ML
arXiv: https://arxiv.org/abs/2601.21860v3

The reconstruction and inference of stochastic dynamical systems from data is a fundamental task in inverse problems and statistical learning. While surrogate modeling advances computational methods to approximate these dynamics, standard approaches typically require high-fidelity training data. In many practical settings, however, the data are indirectly observed through noisy and nonlinear measurements. The challenge lies not only in approximating the coefficients of the SDEs, but in simultaneously inferring the posterior updates given the observations. In this work, we present an amortized path generation method to address these challenges and solve nonlinear stochastic filtering from noisy observations. We first derive a variational inference formulation that solves filtering distribution for a given noisy observation path. This leads to a controlled SDE representation in which the feedback control is identified through the score structure of a pathwise Zakai equation. Motivated by this representation, we construct a conditional generative model that learns, in an amortized manner over observation paths, to transport a prior latent path measure toward the corresponding posterior path measure. We demonstrate the method on nonlinear stochastic systems with multimodal posterior structure, chaotic dynamics, and sparse observations, showing that the learned conditional path generator enables uncertainty quantification for both filtering marginals and trajectory-dependent functionals.

---

### Trajectory-Guided Diffusion for Foreground-Preserving Background Generation in Multi-Layer Documents

Authors: Taewon Kang
Published: 2026-01-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.21857v1

We present a diffusion-based framework for document-centric background generation that achieves foreground preservation and multi-page stylistic consistency through latent-space design rather than explicit constraints. Instead of suppressing diffusion updates or applying masking heuristics, our approach reinterprets diffusion as the evolution of stochastic trajectories through a structured latent space. By shaping the initial noise and its geometric alignment, background generation naturally avoids designated foreground regions, allowing readable content to remain intact without auxiliary mechanisms. To address the long-standing issue of stylistic drift across pages, we decouple style control from text conditioning and introduce cached style directions as persistent vectors in latent space. Once selected, these directions constrain diffusion trajectories to a shared stylistic subspace, ensuring consistent appearance across pages and editing iterations. This formulation eliminates the need for repeated prompt-based style specification and provides a more stable foundation for multi-page generation. Our framework admits a geometric and physical interpretation, where diffusion paths evolve on a latent manifold shaped by preferred directions, and foreground regions are rarely traversed as a consequence of trajectory initialization rather than explicit exclusion. The proposed method is training-free, compatible with existing diffusion backbones, and produces visually coherent, foreground-preserving results across complex documents. By reframing diffusion as trajectory design in latent space, we offer a principled approach to consistent and structured generative modeling.

---

### Blind Ultrasound Image Enhancement via Self-Supervised Physics-Guided Degradation Modeling

Authors: Shujaat Khan, Syed Muhammad Atif, Jaeyoung Huh, Syed Saad Azhar
Published: 2026-01-29
Categories: eess.IV, cs.CV, stat.ML
arXiv: https://arxiv.org/abs/2601.21856v1

Ultrasound (US) interpretation is hampered by multiplicative speckle, acquisition blur from the point-spread function (PSF), and scanner- and operator-dependent artifacts. Supervised enhancement methods assume access to clean targets or known degradations; conditions rarely met in practice. We present a blind, self-supervised enhancement framework that jointly deconvolves and denoises B-mode images using a Swin Convolutional U-Net trained with a \emph{physics-guided} degradation model. From each training frame, we extract rotated/cropped patches and synthesize inputs by (i) convolving with a Gaussian PSF surrogate and (ii) injecting noise via either spatial additive Gaussian noise or complex Fourier-domain perturbations that emulate phase/magnitude distortions. For US scans, clean-like targets are obtained via non-local low-rank (NLLR) denoising, removing the need for ground truth; for natural images, the originals serve as targets. Trained and validated on UDIAT~B, JNU-IFM, and XPIE Set-P, and evaluated additionally on a 700-image PSFHS test set, the method achieves the highest PSNR/SSIM across Gaussian and speckle noise levels, with margins that widen under stronger corruption. Relative to MSANN, Restormer, and DnCNN, it typically preserves an extra $\sim$1--4\,dB PSNR and 0.05--0.15 SSIM in heavy Gaussian noise, and $\sim$2--5\,dB PSNR and 0.05--0.20 SSIM under severe speckle. Controlled PSF studies show reduced FWHM and higher peak gradients, evidence of resolution recovery without edge erosion. Used as a plug-and-play preprocessor, it consistently boosts Dice for fetal head and pubic symphysis segmentation. Overall, the approach offers a practical, assumption-light path to robust US enhancement that generalizes across datasets, scanners, and degradation types.

---

### LEMUR: Learned Multi-Vector Retrieval

Authors: Elias Jääsaari, Ville Hyvönen, Teemu Roos
Published: 2026-01-29
Categories: cs.IR, cs.LG
arXiv: https://arxiv.org/abs/2601.21853v2

Multi-vector representations generated by late interaction models, such as ColBERT, enable superior retrieval quality compared to single-vector representations in information retrieval applications. In multi-vector retrieval systems, both queries and documents are encoded using one embedding per token, and similarity between queries and documents is measured by the MaxSim similarity measure. However, the improved quality of multi-vector retrieval comes at the expense of significantly increased search latency. In this work, we introduce LEMUR, a simple yet efficient framework for multi-vector similarity search. LEMUR consists of two consecutive problem reductions: First, we formulate multi-vector similarity search as a supervised learning problem that can be solved using a one-hidden-layer neural network. Second, we reduce inference under this model to single-vector similarity search in its latent space, enabling the use of existing single-vector search indexes to accelerate retrieval. LEMUR is an order of magnitude faster than prior multi-vector similarity search methods. Our code is available at https://github.com/ejaasaari/lemur

---

### Visual Disentangled Diffusion Autoencoders: Scalable Counterfactual Generation for Foundation Models

Authors: Sidney Bender, Marco Morik
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21851v2

Foundation models, despite their robust zero-shot capabilities, remain vulnerable to spurious correlations and 'Clever Hans' strategies. Existing mitigation methods often rely on unavailable group labels or computationally expensive gradient-based adversarial optimization. To address these limitations, we propose Visual Disentangled Diffusion Autoencoders (DiDAE), a novel framework integrating frozen foundation models with disentangled dictionary learning for efficient, gradient-free counterfactual generation directly for the foundation model. DiDAE first edits foundation model embeddings in interpretable disentangled directions of the disentangled dictionary and then decodes them via a diffusion autoencoder. This allows the generation of multiple diverse, disentangled counterfactuals for each factual, much faster than existing baselines, which generate single entangled counterfactuals. When paired with Counterfactual Knowledge Distillation, DiDAE-CFKD achieves state-of-the-art performance in mitigating shortcut learning, improving downstream performance on unbalanced datasets.

---

### READY: Reward Discovery for Meta-Black-Box Optimization

Authors: Zechuan Huang, Zhiguang Cao, Hongshu Guo, Yue-Jiao Gong, Zeyuan Ma
Published: 2026-01-29
Categories: cs.LG, cs.NE
arXiv: https://arxiv.org/abs/2601.21847v1

Meta-Black-Box Optimization (MetaBBO) is an emerging avenue within Optimization community, where algorithm design policy could be meta-learned by reinforcement learning to enhance optimization performance. So far, the reward functions in existing MetaBBO works are designed by human experts, introducing certain design bias and risks of reward hacking. In this paper, we use Large Language Model~(LLM) as an automated reward discovery tool for MetaBBO. Specifically, we consider both effectiveness and efficiency sides. On effectiveness side, we borrow the idea of evolution of heuristics, introducing tailored evolution paradigm in the iterative LLM-based program search process, which ensures continuous improvement. On efficiency side, we additionally introduce multi-task evolution architecture to support parallel reward discovery for diverse MetaBBO approaches. Such parallel process also benefits from knowledge sharing across tasks to accelerate convergence. Empirical results demonstrate that the reward functions discovered by our approach could be helpful for boosting existing MetaBBO works, underscoring the importance of reward design in MetaBBO. We provide READY's project at https://anonymous.4open.science/r/ICML_READY-747F.

---

### Constrained Meta Reinforcement Learning with Provable Test-Time Safety

Authors: Tingting Ni, Maryam Kamgarpour
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21845v2

Meta reinforcement learning (RL) allows agents to leverage experience across a distribution of tasks on which the agent can train at will, enabling faster learning of optimal policies on new test tasks. Despite its success in improving sample complexity on test tasks, many real-world applications, such as robotics and healthcare, impose safety constraints during testing. Constrained meta RL provides a promising framework for integrating safety into meta RL. An open question in constrained meta RL is how to ensure safety of the policy on the real-world test task, while reducing the sample complexity and thus, enabling faster learning of optimal policies. To address this gap, we propose an algorithm that refines policies learned during training, with provable safety and sample complexity guarantees for learning a near optimal policy on the test tasks. We further derive a matching lower bound, showing that this sample complexity is tight.

---

### Bridging Forecast Accuracy and Inventory KPIs: A Simulation-Based Software Framework

Authors: So Fukuhara, Abdallah Alabdallah, Nuwan Gunasekara, Slawomir Nowaczyk
Published: 2026-01-29
Categories: cs.AI, cs.SE
arXiv: https://arxiv.org/abs/2601.21844v2

Efficient management of spare parts inventory is crucial in the automotive aftermarket, where demand is highly intermittent and uncertainty drives substantial cost and service risks. Forecasting is therefore central, but the quality of forecasting models should be judged not by statistical accuracy (e.g., MAE, RMSE) but rather by its impact on key operational performance indicators (KPIs), such as total cost and service level. Yet most existing work evaluates models exclusively using accuracy metrics, and the relationship between these metrics and KPIs remains poorly understood. To address this gap, we propose a decision-centric simulation software framework that enables systematic evaluation of forecasting models in realistic inventory management setting. The framework comprises: (i) a synthetic demand generator tailored to spare-parts demand characteristics, (ii) a flexible forecasting module that can host arbitrary predictive models, and (iii) an inventory control simulator that consumes the forecasts and computes operational KPIs. This closed-loop setup enables researchers to evaluate models not only in terms of statistical error but also in terms of downstream inventory implications. Using a wide range of simulation scenarios, we show that improvements in accuracy metrics do not necessarily lead to better KPIs, and that models with similar error profiles can induce different cost-service trade-offs. We analyze these discrepancies to characterize how forecast performance affects inventory outcomes and derive guidance for model selection. Overall, the framework links demand forecasting and inventory management, shifting evaluation from predictive accuracy toward operational relevance in the automotive aftermarket and related domains. An open-source implementation of the software is available at https://github.com/caisr-hh/TruckParts-Demand-Inventory-Simulator/releases/tag/IDA_2026.

---

### Embodied Task Planning via Graph-Informed Action Generation with Large Language Models

Authors: Xiang Li, Ning Yan, Masood Mortazavi
Published: 2026-01-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.21841v3

While Large Language Models (LLMs) have demonstrated strong zero-shot reasoning capabilities, their deployment as embodied agents still faces fundamental challenges in long-horizon planning. Unlike open-ended text generation, embodied agents must decompose high-level intents into actionable sub-goals while adhering to the constraints of a dynamic environment. Standard LLM planners frequently fail to maintain strategy coherence over extended horizons due to context window limitations or hallucinate state transitions that violate environment constraints. We propose GiG, a planning framework that structures embodied agents' memory using a Graph-in-Graph architecture. Our approach employs a Graph Neural Network (GNN) to encode environmental states into embeddings, organizing these embeddings into action-connected execution trace graphs within an experience memory bank. GiG enables retrieval of structurally-similar priors, allowing agents to ground current decisions in relevant past structural patterns. Furthermore, we introduce a bounded lookahead module that leverages symbolic transition logic to enhance the agent's planning capabilities through grounded action projections. We evaluate our framework on three embodied planning benchmarks-Robotouille Synchronous, Robotouille Asynchronous, and ALFWorld. Our method outperforms state-of-the-art baselines, achieving Pass@1 performance gains of up to 22% on Robotouille Synchronous, 37% on Asynchronous, and 15% on ALFWorld while maintaining comparable or lower computational cost.

---

### Test-Time Compute Games

Authors: Ander Artola Velasco, Dimitrios Rontogiannis, Stratis Tsirtsis, Manuel Gomez-Rodriguez
Published: 2026-01-29
Categories: cs.CY, cs.AI, cs.GT, cs.LG
arXiv: https://arxiv.org/abs/2601.21839v2

Test-time compute has emerged as a promising strategy to enhance the reasoning abilities of large language models (LLMs). However, this strategy has in turn increased how much users pay cloud-based providers offering LLM-as-a-service, since providers charge users for the amount of test-time compute they use to generate an output. In our work, we show that the market of LLM-as-a-service is socially inefficient: providers have a financial incentive to increase the amount of test-time compute, even if this increase contributes little to the quality of the outputs. To address this inefficiency, we introduce a reverse second-price auction mechanism where providers bid their offered price and (expected) quality for the opportunity to serve a user, and users pay proportionally to the marginal value generated by the winning provider relative to the second-highest bidder. To illustrate and complement our theoretical results, we conduct experiments with multiple instruct models from the $\texttt{Llama}$ and $\texttt{Qwen}$ families, as well as reasoning models distilled from $\texttt{DeepSeek-R1}$, on math and science benchmark datasets.

---

### Trustworthy Intelligent Education: A Systematic Perspective on Progress, Challenges, and Future Directions

Authors: Xiaoshan Yu, Shangshang Yang, Ziwen Wang, Haiping Ma, Xingyi Zhang
Published: 2026-01-29
Categories: cs.CY, cs.AI, cs.IR
arXiv: https://arxiv.org/abs/2601.21837v1

In recent years, trustworthiness has garnered increasing attention and exploration in the field of intelligent education, due to the inherent sensitivity of educational scenarios, such as involving minors and vulnerable groups, highly personalized learning data, and high-stakes educational outcomes. However, existing research either focuses on task-specific trustworthy methods without a holistic view of trustworthy intelligent education, or provides survey-level discussions that remain high-level and fragmented, lacking a clear and systematic categorization. To address these limitations, in this paper, we present a systematic and structured review of trustworthy intelligent education. Specifically, We first organize intelligent education into five representative task categories: learner ability assessment, learning resource recommendation, learning analytics, educational content understanding, and instructional assistance. Building on this task landscape, we review existing studies from five trustworthiness perspectives, including safety and privacy, robustness, fairness, explainability, and sustainability, and summarize and categorize the research methodologies and solution strategies therein. Finally, we summarize key challenges and discuss future research directions. This survey aims to provide a coherent reference framework and facilitate a clearer understanding of trustworthiness in intelligent education.

---

### Scalable Linearized Laplace Approximation via Surrogate Neural Kernel

Authors: Luis A. Ortega, Simón Rodríguez-Santana, Daniel Hernández-Lobato
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21835v2

We introduce a scalable method to approximate the kernel of the Linearized Laplace Approximation (LLA). For this, we use a surrogate deep neural network (DNN) that learns a compact feature representation whose inner product replicates the Neural Tangent Kernel (NTK). This avoids the need to compute large Jacobians. Training relies solely on efficient Jacobian-vector products, allowing to compute predictive uncertainty on large-scale pre-trained DNNs. Experimental results show similar or improved uncertainty estimation and calibration compared to existing LLA approximations. Notwithstanding, biasing the learned kernel significantly enhances out-of-distribution detection. This remarks the benefits of the proposed method for finding better kernels than the NTK in the context of LLA to compute prediction uncertainty given a pre-trained DNN.

---

### Goal-Driven Adaptive Sampling Strategies for Machine Learning Models Predicting Fields

Authors: Jigar Parekh, Philipp Bekemeyer
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21832v1

Machine learning models are widely regarded as a way forward to tackle multi-query challenges that arise once expensive black-box simulations such as computational fluid dynamics are investigated. However, ensuring the desired level of accuracy for a certain task at minimal computational cost, e.g. as few black-box samples as possible, remains a challenges. Active learning strategies are used for scalar quantities to overcome this challenges and different so-called infill criteria exists and are commonly employed in several scenarios. Even though needed in various field an extension of active learning strategies towards field predictions is still lacking or limited to very specific scenarios and/or model types. In this paper we propose an active learning strategy for machine learning models that are capable if predicting field which is agnostic to the model architecture itself. For doing so, we combine a well-established Gaussian process model for a scalar reference value and simultaneously aim at reducing the epistemic model error and the difference between scalar and field predictions. Different specific forms of the above-mentioned approach are introduced and compared to each other as well as only scalar-valued based infill. Results are presented for the NASA common research model for an uncertainty propagation task showcasing high level of accuracy at significantly smaller cost compared to an approach without active learning.

---

### Generative Modeling of Discrete Data Using Geometric Latent Subspaces

Authors: Daniel Gonzalez-Alvarado, Jonas Cassel, Stefania Petra, Christoph Schnörr
Published: 2026-01-29
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2601.21831v3

We propose a geometric latent-subspace framework for generative modeling of discrete data. Specifically, we introduce latent subspaces in the exponential parameter space of product manifolds of categorical distributions as a novel approach to learning low-dimensional representations of high-dimensional discrete data. The resulting low-dimensional latent space captures statistical dependencies and removes redundant degrees of freedom among the categorical variables. We equip the parameter domain with a Riemannian geometry such that the latent subspace and induced data manifold are related isometrically, enabling consistent flow matching. Exploiting this structure, we propose a geometry-aware dimensionality reduction objective, called geometric PCA (GPCA), which we formulate as a regularized cross-entropy minimization that encourages small Riemannian distances between the data and their reconstructions. In particular, under the induced geometry, geodesics correspond to straight lines in the latent parameter space, allowing flow matching to be performed directly in reduced coordinates. Empirical results show that low-dimensional latent representations suffice to accurately model high-dimensional discrete data and enable substantially more computationally efficient flow matching.

---

### Looking Beyond Accuracy: A Holistic Benchmark of ECG Foundation Models

Authors: Francesca Filice, Edoardo De Rose, Simone Bartucci, Francesco Calimeri, Simona Perri
Published: 2026-01-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.21830v1

The electrocardiogram (ECG) is a cost-effective, highly accessible and widely employed diagnostic tool. With the advent of Foundation Models (FMs), the field of AI-assisted ECG interpretation has begun to evolve, as they enable model reuse across different tasks by relying on embeddings. However, to responsibly employ FMs, it is crucial to rigorously assess to which extent the embeddings they produce are generalizable, particularly in error-sensitive domains such as healthcare. Although prior works have already addressed the problem of benchmarking ECG-expert FMs, they focus predominantly on the evaluation of downstream performance. To fill this gap, this study aims to find an in-depth, comprehensive benchmarking framework for FMs, with a specific focus on ECG-expert ones. To this aim, we introduce a benchmark methodology that complements performance-based evaluation with representation-level analysis, leveraging SHAP and UMAP techniques. Furthermore, we rely on the methodology for carrying out an extensive evaluation of several ECG-expert FMs pretrained via state-of-the-art techniques over different cross-continental datasets and data availability settings; this includes ones featuring data scarcity, a fairly common situation in real-world medical scenarios. Experimental results show that our benchmarking protocol provides a rich insight of ECG-expert FMs' embedded patterns, enabling a deeper understanding of their representational structure and generalizability.

---

### Mil-SCORE: Benchmarking Long-Context Geospatial Reasoning and Planning in Large Language Models

Authors: Aadi Palnitkar, Mingyang Mao, Nicholas Waytowich, Vinicius G. Goecks, Xiaomin Lin
Published: 2026-01-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.21826v3

As large language models (LLMs) are applied to increasingly longer and more complex tasks, there is a growing need for realistic long-context benchmarks that require selective reading and integration of heterogeneous, multi-modal information sources. This need is especially acute for geospatial planning problems, such as those found in planning for large-scale military operations, which demand fast and accurate reasoning over maps, orders, intelligence reports, and other distributed data. To address this gap, we present MilSCORE (Military Scenario Contextual Reasoning), to our knowledge the first scenario-level dataset of expert-authored, multi-hop questions grounded in a complex, simulated military planning scenario used for training. MilSCORE is designed to evaluate high-stakes decision-making and planning, probing LLMs' ability to combine tactical and spatial reasoning across multiple sources and to reason over long-horizon, geospatially rich context. The benchmark includes a diverse set of question types across seven categories targeting both factual recall and multi-step reasoning about constraints, strategy, and spatial analysis. We provide an evaluation protocol and report baseline results for a range of contemporary vision-language models. Our findings highlight substantial headroom on MilSCORE, indicating that current systems struggle with realistic, scenario-level long-context planning, and positioning MilSCORE as a challenging testbed for future work.

---

### DASH: Deterministic Attention Scheduling for High-throughput Reproducible LLM Training

Authors: Xinwei Qiang, Hongmin Chen, Shixuan Sun, Jingwen Leng, Xin Liu, Minyi Guo
Published: 2026-01-29
Categories: cs.LG, cs.DC
arXiv: https://arxiv.org/abs/2601.21824v1

Determinism is indispensable for reproducibility in large language model (LLM) training, yet it often exacts a steep performance cost. In widely used attention implementations such as FlashAttention-3, the deterministic backward pass can incur up to a 37.9% throughput reduction relative to its non-deterministic counterpart, primarily because gradient accumulation operations must be serialized to guarantee numerical consistency. This performance loss stems from suboptimal scheduling of compute and gradient-reduction phases, leading to significant hardware underutilization. To address this challenge, we formulate the backward pass of deterministic attention as a scheduling problem on a Directed Acyclic Graph (DAG) and derive schedules that minimize the critical path length. Building on this formulation, we present DASH (Deterministic Attention Scheduling for High-Throughput), which encapsulates two complementary scheduling strategies: (i) Descending Q-Tile Iteration, a reversed query-block traversal that shrinks pipeline stalls in causal attention, and (ii) Shift Scheduling, a theoretically optimal schedule within our DAG model that reduces pipeline stalls for both full and causal masks. Our empirical evaluations on NVIDIA H800 GPUs demonstrate that DASH narrows the performance gap of deterministic attention. The proposed strategies improve the throughput of the attention backward pass by up to 1.28$\times$ compared to the baseline, significantly advancing the efficiency of reproducible LLM training. Our code is open-sourced at https://github.com/SJTU-Liquid/deterministic-FA3.

---

### General Self-Prediction Enhancement for Spiking Neurons

Authors: Zihan Huang, Zijie Xu, Yihan Huang, Shanshan Jia, Tong Bu, Yiting Dong, Wenxuan Liu, Jianhao Ding, Zhaofei Yu, Tiejun Huang
Published: 2026-01-29
Categories: cs.NE
arXiv: https://arxiv.org/abs/2601.21823v1

Spiking Neural Networks (SNNs) are highly energy-efficient due to event-driven, sparse computation, but their training is challenged by spike non-differentiability and trade-offs among performance, efficiency, and biological plausibility. Crucially, mainstream SNNs ignore predictive coding, a core cortical mechanism where the brain predicts inputs and encodes errors for efficient perception. Inspired by this, we propose a self-prediction enhanced spiking neuron method that generates an internal prediction current from its input-output history to modulate membrane potential. This design offers dual advantages, it creates a continuous gradient path that alleviates vanishing gradients and boosts training stability and accuracy, while also aligning with biological principles, which resembles distal dendritic modulation and error-driven synaptic plasticity. Experiments show consistent performance gains across diverse architectures, neuron types, time steps, and tasks demonstrating broad applicability for enhancing SNNs.

---

### CORE:Toward Ubiquitous 6G Intelligence Through Collaborative Orchestration of Large Language Model Agents Over Hierarchical Edge

Authors: Zitong Yu, Boquan Sun, Yang Li, Zheyan Qu, Xing Zhang
Published: 2026-01-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.21822v1

Rapid advancements in sixth-generation (6G) networks and large language models (LLMs) have paved the way for ubiquitous intelligence, wherein seamless connectivity and distributed artificial intelligence (AI) have revolutionized various aspects of our lives.However, realizing this vision faces significant challenges owing to the fragmented and heterogeneous computing resources across hierarchical networks, which are insufficient for individual LLM agents to perform complex reasoning tasks.To address this issue, we propose Collaborative Orchestration Role at Edge (CORE), an innovative framework that employs a collaborative learning system in which multiple LLMs, each assigned a distinct functional role, are distributed across mobile devices and tiered edge servers. The system integrates three optimization modules, encompassing real-time perception,dynamic role orchestration, and pipeline-parallel execution, to facilitate efficient and rapid collaboration among distributed agents. Furthermore, we introduce a novel role affinity scheduling algorithm for dynamically orchestrating LLM role assignments across the hierarchical edge infrastructure, intelligently matching computational demands with available dispersed resources.Finally, comprehensive case studies and performance evaluations across various 6G application scenarios demonstrated the efficacy of CORE, revealing significant enhancements in the system efficiency and task completion rates. Building on these promising outcomes, we further validated the practical applicability of CORE by deploying it on a real-world edge-computing platform,that exhibits robust performance in operational environments.

---

### MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methods

Authors: Honglin Lin, Zheng Liu, Yun Zhu, Chonghan Qin, Juekai Lin, Xiaoran Shang, Conghui He, Wentao Zhang, Lijun Wu
Published: 2026-01-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.21821v1

Recent advances in Vision Language Models (VLMs) have driven significant progress in visual reasoning. However, open-source VLMs still lag behind proprietary systems, largely due to the lack of high-quality reasoning data. Existing datasets offer limited coverage of challenging domains such as STEM diagrams and visual puzzles, and lack consistent, long-form Chain-of-Thought (CoT) annotations essential for eliciting strong reasoning capabilities. To bridge this gap, we introduce MMFineReason, a large-scale multimodal reasoning dataset comprising 1.8M samples and 5.1B solution tokens, featuring high-quality reasoning annotations distilled from Qwen3-VL-235B-A22B-Thinking. The dataset is established via a systematic three-stage pipeline: (1) large-scale data collection and standardization, (2) CoT rationale generation, and (3) comprehensive selection based on reasoning quality and difficulty awareness. The resulting dataset spans STEM problems, visual puzzles, games, and complex diagrams, with each sample annotated with visually grounded reasoning traces. We fine-tune Qwen3-VL-Instruct on MMFineReason to develop MMFineReason-2B/4B/8B versions. Our models establish new state-of-the-art results for their size class. Notably, MMFineReason-4B succesfully surpasses Qwen3-VL-8B-Thinking, and MMFineReason-8B even outperforms Qwen3-VL-30B-A3B-Thinking while approaching Qwen3-VL-32B-Thinking, demonstrating remarkable parameter efficiency. Crucially, we uncover a "less is more" phenomenon via our difficulty-aware filtering strategy: a subset of just 7\% (123K samples) achieves performance comparable to the full dataset. Notably, we reveal a synergistic effect where reasoning-oriented data composition simultaneously boosts general capabilities.

---

### A Judge-Aware Ranking Framework for Evaluating Large Language Models without Ground Truth

Authors: Mingyuan Xu, Xinzi Tan, Jiawei Wu, Doudou Zhou
Published: 2026-01-29
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2601.21817v3

Evaluating large language models (LLMs) on open-ended tasks without ground-truth labels is increasingly done via the LLM-as-a-judge paradigm. A critical but under-modeled issue is that judge LLMs differ substantially in reliability; treating all judges equally can yield biased leaderboards and misleading uncertainty estimates. More data can make evaluation more confidently wrong under misspecified aggregation. We propose a judge-aware ranking framework that extends the Bradley-Terry-Luce model by introducing judge-specific discrimination parameters, jointly estimating latent model quality and judge reliability from pairwise comparisons without reference labels. We establish identifiability up to natural normalizations and prove consistency and asymptotic normality of the maximum likelihood estimator, enabling confidence intervals for score differences and rank comparisons. Across multiple public benchmarks and a newly collected dataset, our method improves agreement with human preferences, achieves higher data efficiency than unweighted baselines, and produces calibrated uncertainty quantification for LLM rankings.

---

### Nonparametric LLM Evaluation from Preference Data

Authors: Dennis Frauen, Athiya Deviyani, Mihaela van der Schaar, Stefan Feuerriegel
Published: 2026-01-29
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.21816v2

Evaluating the performance of large language models (LLMs) from human preference data is crucial for obtaining LLM leaderboards. However, many existing approaches either rely on restrictive parametric assumptions or lack valid uncertainty quantification when flexible machine learning methods are used. In this paper, we propose a nonparametric statistical framework, called DMLRank, for comparing and ranking LLMs from preference data using debiased machine learning (DML). For this, we introduce generalized average ranking scores (GARS), which generalize commonly used ranking models, including the Bradley-Terry model or PageRank/ Rank centrality, with complex human responses such as ties. DMLRank comes with the following advantages: (i)~It produces statistically efficient estimates of GARS ranking scores. (ii) It naturally allows the incorporation of black-box machine learning methods for estimation. (iii) It can be combined with pre-trained LLM evaluators (e.g., using LLM-as-a-judge). (iv) It suggests optimal policies for collecting preference data under budget constraints. We demonstrate these advantages both theoretically and empirically using both synthetic and real-world preference datasets. In summary, our framework provides practitioners with powerful, state-of-the-art methods for comparing or ranking LLMs for leaderboards.

---

### Moral Outrage Shapes Commitments Beyond Attention: Multimodal Moral Emotions on YouTube in Korea and the US

Authors: Seongchan Park, Jaehong Kim, Hyeonseung Kim, Heejin Bin, Sue Moon, Wonjae Lee
Published: 2026-01-29
Categories: cs.CY, cs.AI, cs.CL, cs.SI
arXiv: https://arxiv.org/abs/2601.21815v2

Understanding how media rhetoric shapes audience engagement is crucial in the attention economy. This study examines how moral emotional framing by mainstream news channels on YouTube influences user behavior across Korea and the United States. To capture the platform's multimodal nature, combining thumbnail images and video titles, we develop a multimodal moral emotion classifier by fine tuning a vision language model. The model is trained on human annotated multimodal datasets in both languages and applied to approximately 400,000 videos from major news outlets. We analyze engagement levels including views, likes, and comments, representing increasing degrees of commitment. The results show that other condemning rhetoric expressions of moral outrage that criticize others morally consistently increase all forms of engagement across cultures, with effects ranging from passive viewing to active commenting. These findings suggest that moral outrage is a particularly effective emotional strategy, attracting not only attention but also active participation. We discuss concerns about the potential misuse of other condemning rhetoric, as such practices may deepen polarization by reinforcing in group and out group divisions. To facilitate future research and ensure reproducibility, we publicly release our Korean and English multimodal moral emotion classifiers.

---

### A Decomposable Forward Process in Diffusion Models for Time-Series Forecasting

Authors: Francisco Caldas, Sahil Kumar, Cláudia Soares
Published: 2026-01-29
Categories: stat.ML, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.21812v2

We introduce a model-agnostic forward diffusion process for time-series forecasting that decomposes signals into spectral components, preserving structured temporal patterns such as seasonality more effectively than standard diffusion. Unlike prior work that modifies the network architecture or diffuses directly in the frequency domain, our proposed method alters only the diffusion process itself, making it compatible with existing diffusion backbones (e.g., DiffWave, TimeGrad, CSDI). By staging noise injection according to component energy, it maintains high signal-to-noise ratios for dominant frequencies throughout the diffusion trajectory, thereby improving the recoverability of long-term patterns. This strategy enables the model to maintain the signal structure for a longer period in the forward process, leading to improved forecast quality. Across standard forecasting benchmarks, we show that applying spectral decomposition strategies, such as the Fourier or Wavelet transform, consistently improves upon diffusion models using the baseline forward process, with negligible computational overhead. The code for this paper is available at https://anonymous.4open.science/r/D-FDP-4A29.

---

### Distribution-Aware Reward Estimation for Test-Time Reinforcement Learning

Authors: Bodong Du, Xuanqi Huang, Xiaomeng Li
Published: 2026-01-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.21804v1

Test-time reinforcement learning (TTRL) enables large language models (LLMs) to self-improve on unlabeled inputs, but its effectiveness critically depends on how reward signals are estimated without ground-truth supervision. Most existing TTRL methods rely on majority voting (MV) over rollouts to produce deterministic rewards, implicitly assuming that the majority rollout provides a reliable learning signal. We show that this assumption is fragile: MV reduces the rollout distribution into a single outcome, discarding information about non-majority but correct actions candidates, and yields systematically biased reward estimates. To address this, we propose Distribution-AwareReward Estimation (DARE), which shifts reward estimation from a single majority outcome to the full empirical rollout distribution. DARE further augments this distribution-based reward with an exploration bonus and a distribution pruning mechanism for non-majority rollout exploration and reward denoise, yielding a more informative and robust reward estimation. Extensive experiments on challenging reasoning benchmarks show that DARE improves optimization stability and final performance over recent baselines, achieving relative improvements of 25.3% on challenging AIME 2024 and 5.3% on AMC.

---

### Quantifying Retriever-Generator Alignment in RAG with Local Explanations

Authors: Korbinian Randl, Guido Rocchietti, Aron Henriksson, Ziawasch Abedjan, Tony Lindgren, John Pavlopoulos
Published: 2026-01-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.21803v2

Retrieval-Augmented Generation (RAG) systems combine dense retrievers and language models to ground their outputs in external documents. However, the interaction between these components remains opaque, creating challenges for deployment in high-stakes domains. We present RAG-E, an end-to-end explainability framework that quantifies retriever-generator alignment through mathematically grounded attribution methods. Our approach adapts Integrated Gradients for retriever analysis, proposes a Monte Carlo-stabilized Shapley Value approximation for generator attribution, and introduces the Weighted Attribution-Relevance Gap (WARG) metric to measure how closely the generator's document usage aligns with retriever rankings. Experiments on PopQA, QAMPARI, and TREC CAST datasets reveal substantial misalignment: depending on the model and setting, generators often ignore top-ranked documents and rely on documents ranked as less relevant. We show that WARG captures retriever-generator alignment better than Pearson and Spearman correlations and can serve as an indicator of RAG performance. RAG-E and WARG provide a practical framework for auditing this interaction, enabling more reliable and transparent RAG systems.

---

### A Unified XAI-LLM Approach for EndotrachealSuctioning Activity Recognition

Authors: Hoang Khang Phan, Quang Vinh Dang, Noriyo Colley, Christina Garcia, Nhat Tan Le
Published: 2026-01-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.21802v1

Endotracheal suctioning (ES) is an invasive yet essential clinical procedure that requires a high degree of skill to minimize patient risk - particularly in home care and educational settings, where consistent supervision may be limited. Despite its critical importance, automated recognition and feedback systems for ES training remain underexplored. To address this gap, this study proposes a unified, LLM-centered framework for video-based activity recognition benchmarked against conventional machine learning and deep learning approaches, and a pilot study on feedback generation. Within this framework, the Large Language Model (LLM) serves as the central reasoning module, performing both spatiotemporal activity recognition and explainable decision analysis from video data. Furthermore, the LLM is capable of verbalizing feedback in natural language, thereby translating complex technical insights into accessible, human-understandable guidance for trainees. Experimental results demonstrate that the proposed LLM-based approach outperforms baseline models, achieving an improvement of approximately 15-20\% in both accuracy and F1 score. Beyond recognition, the framework incorporates a pilot student-support module built upon anomaly detection and explainable AI (XAI) principles, which provides automated, interpretable feedback highlighting correct actions and suggesting targeted improvements. Collectively, these contributions establish a scalable, interpretable, and data-driven foundation for advancing nursing education, enhancing training efficiency, and ultimately improving patient safety.

---

### BioAgent Bench: An AI Agent Evaluation Suite for Bioinformatics

Authors: Dionizije Fa, Marko Culjak, Bruno Pandza, Mateo Cupic
Published: 2026-01-29
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.21800v4

We introduce BioAgent Bench, an evaluation suite designed for measuring the performance and robustness of AI agents in common bioinformatics tasks. The suite consists of manually curated end-to-end tasks (e.g., RNA-seq, variant calling, metagenomics) accompanied by task-specific prompts and concrete output artifacts to support automated assessment. We evaluate frontier closed- and open-weight models across multiple agent harnesses, and use an LLM-based grader to score pipeline progress and outcome validity. We find that agents based on frontier LLMs can complete multi-step bioinformatics pipelines without elaborate custom scaffolding, often producing the requested final artifacts reliably. However, robustness tests reveal failure modes under controlled perturbations (corrupted inputs, decoy files, and prompt bloat), indicating that correct high-level pipeline construction does not guarantee reliable step-level reasoning. By releasing the code and the complementary resources constituting our suite, our primary goal is to accelerate the development of cost-effective yet reliable local agents, capable of handling complex bioinformatics workflows often involving sensitive patient data or unpublished intellectual property.

---

### CG-MLLM: Captioning and Generating 3D content via Multi-modal Large Language Models

Authors: Junming Huang, Chi Wang, Letian Li, Guangkai Xu, Donglin Huang, Hao Chen, Qiang Dai, Weiwei Xu
Published: 2026-01-29
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.21798v2

Large Language Models(LLMs) have revolutionized text generation and multimodal perception,but their capabilities in 3D content generation remain underexplored. Existing methods compromise by producing either low-resolution meshes or coarse structural proxies, failing to capture finegrained geometry natively. In this paper, we propose CG-MLLM, a novel Multi-modal Large Language Model (MLLM) capable of 3D captioning and high-resolution 3D generation in a single framework. Leveraging the Mixture-ofTransformer architecture, CG-MLLM decouples disparate modeling needs, where the Token-level Autoregressive (TokenAR) Transformer handles token-level content, and the Block-level Autoregressive (BlockAR) Transformer handles blocklevel content. By integrating a pre-trained visionlanguage backbone with a specialized 3D VAE latent space, CG-MLLM facilitates long-context interactions between standard tokens and spatial blocks within a single integrated architecture. Experimental results show that CG-MLLM significantly outperforms existing MLLMs in generating high-fidelity 3D objects, effectively bringing high-resolution 3D content creation into the mainstream LLM paradigm. Beyond generation, we further observe that learning to produce 3D content transfers back to perception, strengthening the model's image-based 3D understanding.

---

### Enhancing Conversational Agents via Task-Oriented Adversarial Memory Adaptation

Authors: Yimin Deng, Yuqing Fu, Derong Xu, Yejing Wang, Wei Ni, Jingtong Gao, Xiaopeng Li, Chengxu Liu, Xiao Han, Guoshuai Zhao, Xiangyu Zhao, Li Zhu et al.
Published: 2026-01-29
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.21797v1

Conversational agents struggle to handle long conversations due to context window limitations. Therefore, memory systems are developed to leverage essential historical information. Existing memory systems typically follow a pipeline of offline memory construction and update, and online retrieval. Despite the flexible online phase, the offline phase remains fixed and task-independent. In this phase, memory construction operates under a predefined workflow and fails to emphasize task relevant information. Meanwhile, memory updates are guided by generic metrics rather than task specific supervision. This leads to a misalignment between offline memory preparation and task requirements, which undermines downstream task performance. To this end, we propose an Adversarial Memory Adaptation mechanism (AMA) that aligns memory construction and update with task objectives by simulating task execution. Specifically, first, a challenger agent generates question answer pairs based on the original dialogues. The constructed memory is then used to answer these questions, simulating downstream inference. Subsequently, an evaluator agent assesses the responses and performs error analysis. Finally, an adapter agent analyzes the error cases and performs dual level updates on both the construction strategy and the content. Through this process, the memory system receives task aware supervision signals in advance during the offline phase, enhancing its adaptability to downstream tasks. AMA can be integrated into various existing memory systems, and extensive experiments on long dialogue benchmark LoCoMo demonstrate its effectiveness.

---