# Research Papers — 2025-10 · bundle page 201 of 217

> Full-text markdown of papers #10001–10050 of 10843 published in 2025-10, newest first. AI-readable mirrors of the full 10843-paper 2025-10 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2025-10/page/201
[← Previous bundle](https://aipod.dev/papers/archive/2025-10/page/200.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-10/page/202.md)

### Unlocking the power of partnership: How humans and machines can work together to improve face recognition

Authors: P. Jonathon Phillips, Geraldine Jeckeln, Amy N. Yates, Carina A. Hahn, Peter C. Fontana, Alice J. O'Toole
Published: 2025-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.02570v2

Human review of consequential decisions by face recognition algorithms creates a collaborative human-machine system. We establish the circumstances under which combining human and machine face identification decisions improves accuracy. Using data from expert and non-expert face identifiers, we show that the benefits of human-human and human-machine collaborations increase as the difference in baseline accuracy between collaborators decreases. This rule holds across a wide range of baseline abilities, from novices to professional forensic face examiners. An important consequence of the rule is that people who are substantially less accurate than the machine, can actually improve decision accuracy when they collaborate with the machine. In a group of individual people collaborating with a machine, "intelligent human-machine fusion" was implemented by selecting people with the potential to increase collaborative accuracy. Performance with intelligent human-machine fusion was more accurate than either the machine operating alone or fusing all humans with the machine. Eliminating the machine from consideration yielded less predictable results, with average performance at or below intelligent human-machine collaboration. However, intelligent human-machine fusion was consistently more effective at minimizing the impact of low-performing humans on accuracy. The results demonstrate a meaningful role for both humans and machines in assuring accurate face identification.

---

### Transcribe, Translate, or Transliterate: An Investigation of Intermediate Representations in Spoken Language Models

Authors: Tolúlopé Ògúnrèmí, Christopher D. Manning, Dan Jurafsky, Karen Livescu
Published: 2025-10-02
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.02569v2

Spoken language models (SLMs) that integrate speech with large language models (LMs) rely on modality adapters (MAs) to map the output of speech encoders to a representation that is understandable to the decoder LM. Yet we know very little about how these crucial MAs transform representations. Here we examine the MA output representation in three SLMs (SALMONN, Qwen2-Audio and Phi-4-Multimodal-Instruct). By finding the nearest decoder LM token to an MA representation, we uncover two strategies for MA representations. For models using a Whisper encoder, MAs appear to represent the meaning of the input using an English-based interlingua, allowing them to handle languages unseen in instruction tuning. For models that don't, like Phi-4-Multimodal-Instruct, MAs instead represent the phonetics of the input, but expressed with English words. We hypothesise that which arises depends on whether the speech encoder is trained only for speech recognition or also for translation.

---

### Identifying Asymptomatic Nodes in Network Epidemics using Graph Neural Networks

Authors: Conrado Catarcione Pinto, Amanda Camacho Novaes de Oliveira, Rodrigo Sapienza Luna, Daniel Ratton Figueiredo
Published: 2025-10-02
Categories: cs.SI, cs.NE, q-bio.PE
arXiv: https://arxiv.org/abs/2510.02568v1

Infected individuals in some epidemics can remain asymptomatic while still carrying and transmitting the infection. These individuals contribute to the spread of the epidemic and pose a significant challenge to public health policies. Identifying asymptomatic individuals is critical for measuring and controlling an epidemic, but periodic and widespread testing of healthy individuals is often too costly. This work tackles the problem of identifying asymptomatic individuals considering a classic SI (Susceptible-Infected) network epidemic model where a fraction of the infected nodes are not observed as infected (i.e., their observed state is identical to susceptible nodes). In order to classify healthy nodes as asymptomatic or susceptible, a Graph Neural Network (GNN) model with supervised learning is adopted where a set of node features are built from the network with observed infected nodes. The approach is evaluated across different network models, network sizes, and fraction of observed infections. Results indicate that the proposed methodology is robust across different scenarios, accurately identifying asymptomatic nodes while also generalizing to different network sizes and fraction of observed infections.

---

### Agentic Additive Manufacturing Alloy Evaluation

Authors: Peter Pak, Achuth Chandrasekhar, Amir Barati Farimani
Published: 2025-10-02
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.02567v3

Agentic systems enable the intelligent use of research tooling, augmenting a researcher's ability to investigate and propose novel solutions to existing problems. Within Additive Manufacturing (AM), alloy selection and evaluation remains a complex challenge, often requiring expertise in the various domains of materials science, thermodynamic simulations, and experimental analysis. Large Language Model (LLM) enabled agents can facilitate this endeavor by utilizing their extensive knowledge base to dispatch tool calls via Model Context Protocol (MCP) to perform actions such as thermophysical property diagram calculations and lack of fusion process map generation. In addition, the multi-agent system can effectively reason through complex user prompts and provide analysis on the lack of fusion process window of common alloys such as SS316L and IN718 along with proposed composition variants of known alloys. These agents can dynamically adjust their task trajectory to the outcomes of tool call results, effectively enabling autonomous decision-making in practical environments. This work aims to showcase the benefits of adopting a LLM enabled multi-agent system to automate and accelerate the task of evaluating proposed additive manufacturing alloys, both novel and known.

---

### PhysHMR: Learning Humanoid Control Policies from Vision for Physically Plausible Human Motion Reconstruction

Authors: Qiao Feng, Yiming Huang, Yufu Wang, Jiatao Gu, Lingjie Liu
Published: 2025-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.02566v1

Reconstructing physically plausible human motion from monocular videos remains a challenging problem in computer vision and graphics. Existing methods primarily focus on kinematics-based pose estimation, often leading to unrealistic results due to the lack of physical constraints. To address such artifacts, prior methods have typically relied on physics-based post-processing following the initial kinematics-based motion estimation. However, this two-stage design introduces error accumulation, ultimately limiting the overall reconstruction quality. In this paper, we present PhysHMR, a unified framework that directly learns a visual-to-action policy for humanoid control in a physics-based simulator, enabling motion reconstruction that is both physically grounded and visually aligned with the input video. A key component of our approach is the pixel-as-ray strategy, which lifts 2D keypoints into 3D spatial rays and transforms them into global space. These rays are incorporated as policy inputs, providing robust global pose guidance without depending on noisy 3D root predictions. This soft global grounding, combined with local visual features from a pretrained encoder, allows the policy to reason over both detailed pose and global positioning. To overcome the sample inefficiency of reinforcement learning, we further introduce a distillation scheme that transfers motion knowledge from a mocap-trained expert to the vision-conditioned policy, which is then refined using physically motivated reinforcement learning rewards. Extensive experiments demonstrate that PhysHMR produces high-fidelity, physically plausible motion across diverse scenarios, outperforming prior approaches in both visual accuracy and physical realism.

---

### On The Expressive Power of GNN Derivatives

Authors: Yam Eitan, Moshe Eliasof, Yoav Gelberg, Fabrizio Frasca, Guy Bar-Shalom, Haggai Maron
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.02565v1

Despite significant advances in Graph Neural Networks (GNNs), their limited expressivity remains a fundamental challenge. Research on GNN expressivity has produced many expressive architectures, leading to architecture hierarchies with models of increasing expressive power. Separately, derivatives of GNNs with respect to node features have been widely studied in the context of the oversquashing and over-smoothing phenomena, GNN explainability, and more. To date, these derivatives remain unexplored as a means to enhance GNN expressivity. In this paper, we show that these derivatives provide a natural way to enhance the expressivity of GNNs. We introduce High-Order Derivative GNN (HOD-GNN), a novel method that enhances the expressivity of Message Passing Neural Networks (MPNNs) by leveraging high-order node derivatives of the base model. These derivatives generate expressive structure-aware node embeddings processed by a second GNN in an end-to-end trainable architecture. Theoretically, we show that the resulting architecture family's expressive power aligns with the WL hierarchy. We also draw deep connections between HOD-GNN, Subgraph GNNs, and popular structural encoding schemes. For computational efficiency, we develop a message-passing algorithm for computing high-order derivatives of MPNNs that exploits graph sparsity and parallelism. Evaluations on popular graph learning benchmarks demonstrate HOD-GNN's strong performance on popular graph learning tasks.

---

### Oracle-RLAIF: An Improved Fine-Tuning Framework for Multi-modal Video Models using Reinforcement Learning from Ranking Feedback

Authors: Derek Shi, Ruben Glatt, Christine Klymko, Shubham Mohole, Hongjun Choi, Shashank Kushwaha, Sam Sakla, Felipe Leno da Silva
Published: 2025-10-02
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2510.02561v2

Recent advances in large video-language models (VLMs) rely on extensive fine-tuning techniques that strengthen alignment between textual and visual comprehension. Leading pipelines typically pair supervised fine-tuning (SFT) with reinforcement learning from preference data to enhance video comprehension. However, as VLMs scale in parameter size, so does the cost of gathering enough human feedback. To make fine-tuning more cost-effective, recent frameworks explore reinforcement learning with AI feedback (RLAIF), which replace human preference with AI as a judge. Current RLAIF frameworks rely on a specialized reward model trained with video narratives to create calibrated scalar rewards -- an expensive and restrictive pipeline. We propose Oracle-RLAIF, a novel framework that replaces the trained reward model with a more general Oracle ranker which acts as a drop-in model ranking candidate model responses rather than scoring them. Alongside Oracle-RLAIF, we introduce $GRPO_{rank}$, a novel rank-based loss function based on Group Relative Policy Optimization (GRPO) that directly optimizes ordinal feedback with rank-aware advantages. Empirically, we demonstrate that Oracle-RLAIF consistently outperforms leading VLMs using existing fine-tuning methods when evaluated across various video comprehension benchmarks. Oracle-RLAIF paves the path to creating flexible and data-efficient frameworks for aligning large multi-modal video models with reinforcement learning from rank rather than score.

---

### AttentiveGRUAE: An Attention-Based GRU Autoencoder for Temporal Clustering and Behavioral Characterization of Depression from Wearable Data

Authors: Nidhi Soley, Vishal M Patel, Casey O Taylor
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.02558v2

In this study, we present AttentiveGRUAE, a novel attention-based gated recurrent unit (GRU) autoencoder designed for temporal clustering and prediction of outcome from longitudinal wearable data. Our model jointly optimizes three objectives: (1) learning a compact latent representation of daily behavioral features via sequence reconstruction, (2) predicting end-of-period depression rate through a binary classification head, and (3) identifying behavioral subtypes through Gaussian Mixture Model (GMM) based soft clustering of learned embeddings. We evaluate AttentiveGRUAE on longitudinal sleep data from 372 participants (GLOBEM 2018-2019), and it demonstrates superior performance over baseline clustering, domain-aligned self-supervised, and ablated models in both clustering quality (silhouette score = 0.70 vs 0.32-0.70) and depression classification (AUC = 0.74 vs 0.50-0.67). Additionally, external validation on cross-year cohorts from 332 participants (GLOBEM 2020-2021) confirms cluster reproducibility (silhouette score = 0.63, AUC = 0.61) and stability. We further perform subtype analysis and visualize temporal attention, which highlights sleep-related differences between clusters and identifies salient time windows that align with changes in sleep regularity, yielding clinically interpretable explanations of risk.

---

### Orchestrating Human-AI Teams: The Manager Agent as a Unifying Research Challenge

Authors: Charlie Masters, Advaith Vellanki, Jiangbo Shangguan, Bart Kultys, Jonathan Gilmore, Alastair Moore, Stefano V. Albrecht
Published: 2025-10-02
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.02557v1

While agentic AI has advanced in automating individual tasks, managing complex multi-agent workflows remains a challenging problem. This paper presents a research vision for autonomous agentic systems that orchestrate collaboration within dynamic human-AI teams. We propose the Autonomous Manager Agent as a core challenge: an agent that decomposes complex goals into task graphs, allocates tasks to human and AI workers, monitors progress, adapts to changing conditions, and maintains transparent stakeholder communication. We formalize workflow management as a Partially Observable Stochastic Game and identify four foundational challenges: (1) compositional reasoning for hierarchical decomposition, (2) multi-objective optimization under shifting preferences, (3) coordination and planning in ad hoc teams, and (4) governance and compliance by design. To advance this agenda, we release MA-Gym, an open-source simulation and evaluation framework for multi-agent workflow orchestration. Evaluating GPT-5-based Manager Agents across 20 workflows, we find they struggle to jointly optimize for goal completion, constraint adherence, and workflow runtime - underscoring workflow management as a difficult open problem. We conclude with organizational and ethical implications of autonomous management systems.

---

### ToolTweak: An Attack on Tool Selection in LLM-based Agents

Authors: Jonathan Sneh, Ruomei Yan, Jialin Yu, Philip Torr, Yarin Gal, Sunando Sengupta, Eric Sommerlade, Alasdair Paren, Adel Bibi
Published: 2025-10-02
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2510.02554v1

As LLMs increasingly power agents that interact with external tools, tool use has become an essential mechanism for extending their capabilities. These agents typically select tools from growing databases or marketplaces to solve user tasks, creating implicit competition among tool providers and developers for visibility and usage. In this paper, we show that this selection process harbors a critical vulnerability: by iteratively manipulating tool names and descriptions, adversaries can systematically bias agents toward selecting specific tools, gaining unfair advantage over equally capable alternatives. We present ToolTweak, a lightweight automatic attack that increases selection rates from a baseline of around 20% to as high as 81%, with strong transferability between open-source and closed-source models. Beyond individual tools, we show that such attacks cause distributional shifts in tool usage, revealing risks to fairness, competition, and security in emerging tool ecosystems. To mitigate these risks, we evaluate two defenses: paraphrasing and perplexity filtering, which reduce bias and lead agents to select functionally similar tools more equally. All code will be open-sourced upon acceptance.

---

### Knowledge-Graph Based RAG System Evaluation Framework

Authors: Sicheng Dong, Vahid Zolfaghari, Nenad Petrovic, Alois Knoll
Published: 2025-10-02
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.02549v1

Large language models (LLMs) has become a significant research focus and is utilized in various fields, such as text generation and dialog systems. One of the most essential applications of LLM is Retrieval Augmented Generation (RAG), which greatly enhances generated content's reliability and relevance. However, evaluating RAG systems remains a challenging task. Traditional evaluation metrics struggle to effectively capture the key features of modern LLM-generated content that often exhibits high fluency and naturalness. Inspired by the RAGAS tool, a well-known RAG evaluation framework, we extended this framework into a KG-based evaluation paradigm, enabling multi-hop reasoning and semantic community clustering to derive more comprehensive scoring metrics. By incorporating these comprehensive evaluation criteria, we gain a deeper understanding of RAG systems and a more nuanced perspective on their performance. To validate the effectiveness of our approach, we compare its performance with RAGAS scores and construct a human-annotated subset to assess the correlation between human judgments and automated metrics. In addition, we conduct targeted experiments to demonstrate that our KG-based evaluation method is more sensitive to subtle semantic differences in generated outputs. Finally, we discuss the key challenges in evaluating RAG systems and highlight potential directions for future research.

---

### Exploring OCR-augmented Generation for Bilingual VQA

Authors: JoonHo Lee, Sunho Park
Published: 2025-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.02543v1

We investigate OCR-augmented generation with Vision Language Models (VLMs), exploring tasks in Korean and English toward multilingualism. To support research in this domain, we train and release KLOCR, a strong bilingual OCR baseline trained on 100M instances to augment VLMs with OCR ability. To complement existing VQA benchmarks, we curate KOCRBench for Korean VQA, and analyze different prompting methods. Extensive experiments show that OCR-extracted text significantly boosts performance across open source and commercial models. Our work offers new insights into OCR-augmented generation for bilingual VQA. Model, code, and data are available at https://github.com/JHLee0513/KLOCR.

---

### Even Faster Kernel Matrix Linear Algebra via Density Estimation

Authors: Rikhav Shah, Sandeep Silwal, Haike Xu
Published: 2025-10-02
Categories: cs.DS, cs.LG, math.NA
arXiv: https://arxiv.org/abs/2510.02540v2

This paper studies the use of kernel density estimation (KDE) for linear algebraic tasks involving the kernel matrix of a collection of $n$ data points in $\mathbb R^d$. In particular, we improve upon existing algorithms for computing the following up to $(1+\varepsilon)$ relative error: matrix-vector products, matrix-matrix products, the spectral norm, and sum of all entries. The runtimes of our algorithms depend on the dimension $d$, the number of points $n$, and the target error $\varepsilon$. Importantly, the dependence on $n$ in each case is far lower when accessing the kernel matrix through KDE queries as opposed to reading individual entries. Our improvements over existing best algorithms (particularly those of Backurs, Indyk, Musco, and Wagner '21) for these tasks reduce the polynomial dependence on $\varepsilon$, and additionally decreases the dependence on $n$ in the case of computing the sum of all entries of the kernel matrix. We complement our upper bounds with several lower bounds for related problems, which provide (conditional) quadratic time hardness results and additionally hint at the limits of KDE based approaches for the problems we study.

---

### Hierarchical Semantic Retrieval with Cobweb

Authors: Anant Gupta, Karthik Singaravadivelan, Zekun Wang
Published: 2025-10-02
Categories: cs.CL, cs.IR
arXiv: https://arxiv.org/abs/2510.02539v2

Neural document retrieval often treats a corpus as a flat cloud of vectors scored at a single granularity, leaving corpus structure underused and explanations opaque. We use Cobweb--a hierarchy-aware framework--to organize sentence embeddings into a prototype tree and rank documents via coarse-to-fine traversal. Internal nodes act as concept prototypes, providing multi-granular relevance signals and a transparent rationale through retrieval paths. We instantiate two inference approaches: a generalized best-first search and a lightweight path-sum ranker. We evaluate our approaches on MS MARCO and QQP with encoder (e.g., BERT/T5) and decoder (GPT-2) representations. Our results show that our retrieval approaches match the dot product search on strong encoder embeddings while remaining robust when kNN degrades: with GPT-2 vectors, dot product performance collapses whereas our approaches still retrieve relevant results. Overall, our experiments suggest that Cobweb provides competitive effectiveness, improved robustness to embedding quality, scalability, and interpretable retrieval via hierarchical prototypes.

---

### PHORECAST: Enabling AI Understanding of Public Health Outreach Across Populations

Authors: Rifaa Qadri, Anh Nhat Nhu, Swati Ramnath, Laura Yu Zheng, Raj Bhansali, Sylvette La Touche-Howard, Tracy Marie Zeeger, Tom Goldstein, Ming Lin
Published: 2025-10-02
Categories: cs.CY, cs.AI
arXiv: https://arxiv.org/abs/2510.02535v2

Understanding how diverse individuals and communities respond to persuasive messaging holds significant potential for advancing personalized and socially aware machine learning. While Large Vision and Language Models (VLMs) offer promise, their ability to emulate nuanced, heterogeneous human responses, particularly in high stakes domains like public health, remains underexplored due in part to the lack of comprehensive, multimodal dataset. We introduce PHORECAST (Public Health Outreach REceptivity and CAmpaign Signal Tracking), a multimodal dataset curated to enable fine-grained prediction of both individuallevel behavioral responses and community-wide engagement patterns to health messaging. This dataset supports tasks in multimodal understanding, response prediction, personalization, and social forecasting, allowing rigorous evaluation of how well modern AI systems can emulate, interpret, and anticipate heterogeneous public sentiment and behavior. By providing a new dataset to enable AI advances for public health, PHORECAST aims to catalyze the development of models that are not only more socially aware but also aligned with the goals of adaptive and inclusive health communication

---

### Learning Multi-Index Models with Hyper-Kernel Ridge Regression

Authors: Shuo Huang, Hippolyte Labarrière, Ernesto De Vito, Tomaso Poggio, Lorenzo Rosasco
Published: 2025-10-02
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2510.02532v2

Deep neural networks excel in high-dimensional problems, outperforming models such as kernel methods, which suffer from the curse of dimensionality. However, the theoretical foundations of this success remain poorly understood. We follow the idea that the compositional structure of the learning task is the key factor determining when deep networks outperform other approaches. Taking a step towards formalizing this idea, we consider a simple compositional model, namely the multi-index model (MIM). In this context, we introduce and study hyper-kernel ridge regression (HKRR), an approach blending neural networks and kernel methods. Our main contribution is a sample complexity result demonstrating that HKRR can adaptively learn MIM, overcoming the curse of dimensionality. Further, we exploit the kernel nature of the estimator to develop ad hoc optimization approaches. Indeed, we contrast alternating minimization and alternating gradient methods both theoretically and numerically. These numerical results complement and reinforce our theoretical findings.

---

### Multimodal Function Vectors for Visual Relations

Authors: Shuhao Fu, Esther Goldberg, Ying Nian Wu, Hongjing Lu
Published: 2025-10-02
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.02528v2

Large Multimodal Models (LMMs) demonstrate impressive in-context learning abilities from few multimodal demonstrations, yet the internal mechanisms supporting such task learning remain opaque. Building on prior work of Large Language Models, we show that a small subset of attention heads in Large Multimodal Models is responsible for transmitting representations of visual relations. The activations of these attention heads, termed function vectors, can be extracted and manipulated to alter an LMM's performance on relational tasks. First, using synthetic and real image datasets, we apply causal mediation analysis to identify attention heads that strongly influence relational predictions, and extract multimodal function vectors that improve zero-shot accuracy at inference time. We further demonstrate that these multimodal function vectors can be fine-tuned with a modest amount of training data, while keeping LMM parameters frozen, to significantly outperform in-context learning baselines. Finally, we show that relation-specific function vectors can be linearly combined to solve analogy problems involving novel and untrained visual relations, highlighting the strong generalization ability of this approach. Through experiments on two LMMs, including OpenFlamingo and Qwen3-VL, our results show that these models encode visual relational knowledge within localized internal structures, which can be systematically extracted and optimized, thereby advancing our understanding of model modularity and enhancing control over relational reasoning in LMMs.

---

### Self-supervised diffusion model fine-tuning for costate initialization using Markov chain Monte Carlo

Authors: Jannik Graebner, Ryne Beeson
Published: 2025-10-02
Categories: astro-ph.EP, astro-ph.IM, cs.LG, eess.SY, math.OC
arXiv: https://arxiv.org/abs/2510.02527v1

Global search and optimization of long-duration, low-thrust spacecraft trajectories with the indirect method is challenging due to a complex solution space and the difficulty of generating good initial guesses for the costate variables. This is particularly true in multibody environments. Given data that reveals a partial Pareto optimal front, it is desirable to find a flexible manner in which the Pareto front can be completed and fronts for related trajectory problems can be found. In this work we use conditional diffusion models to represent the distribution of candidate optimal trajectory solutions. We then introduce into this framework the novel approach of using Markov Chain Monte Carlo algorithms with self-supervised fine-tuning to achieve the aforementioned goals. Specifically, a random walk Metropolis algorithm is employed to propose new data that can be used to fine-tune the diffusion model using a reward-weighted training based on efficient evaluations of constraint violations and missions objective functions. The framework removes the need for separate focused and often tedious data generation phases. Numerical experiments are presented for two problems demonstrating the ability to improve sample quality and explicitly target Pareto optimality based on the theory of Markov chains. The first problem does so for a transfer in the Jupiter-Europa circular restricted three-body problem, where the MCMC approach completes a partial Pareto front. The second problem demonstrates how a dense and superior Pareto front can be generated by the MCMC self-supervised fine-tuning method for a Saturn-Titan transfer starting from the Jupiter-Europa case versus a separate dedicated global search.

---

### Unraveling Syntax: Language Modeling and the Substructure of Grammars

Authors: Laura Ying Schulz, Daniel Mitropolsky, Tomaso Poggio
Published: 2025-10-02
Categories: cs.CL, cs.FL, cs.LG
arXiv: https://arxiv.org/abs/2510.02524v3

While language models achieve impressive results, their learning dynamics are far from understood. Many domains of interest -- such as natural language syntax, coding languages, arithmetic -- are captured by context-free grammars (CFGs). In this work, we extend prior work on neural language modeling of CFGs in a novel direction: how language modeling behaves with respect to CFG substructure, namely subgrammars. We define subgrammars, and prove a set of fundamental theorems connecting language modeling and subgrammars. We show that language modeling loss recurses linearly over its top-level subgrammars; applied recursively, the loss decomposes into losses for "irreducible" subgrammars. Under additional assumptions, and empirically, parametrized models learn subgrammars in parallel, unlike children who first master simple substructures. We find that subgrammar pretraining can improve final performance, but only for tiny models relative to the grammar, while alignment analyses show that pretraining consistently leads to internal representations that better reflect the grammar's substructure.

---

### Model-brain comparison using inter-animal transforms

Authors: Imran Thobani, Javier Sagastuy-Brena, Aran Nayebi, Jacob Prince, Rosa Cao, Daniel Yamins
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.02523v1

Artificial neural network models have emerged as promising mechanistic models of the brain. However, there is little consensus on the correct method for comparing model activations to brain responses. Drawing on recent work in philosophy of neuroscience, we propose a comparison methodology based on the Inter-Animal Transform Class (IATC) - the strictest set of functions needed to accurately map neural responses between subjects in an animal population. Using the IATC, we can map bidirectionally between a candidate model's responses and brain data, assessing how well the model can masquerade as a typical subject using the same kinds of transforms needed to map across real subjects. We identify the IATC in three settings: a simulated population of neural network models, a population of mouse subjects, and a population of human subjects. We find that the IATC resolves detailed aspects of the neural mechanism, such as the non-linear activation function. Most importantly, we find that the IATC enables accurate predictions of neural activity while also achieving high specificity in mechanism identification, evidenced by its ability to separate response patterns from different brain areas while strongly aligning same-brain-area responses between subjects. In other words, the IATC is a proof-by-existence that there is no inherent tradeoff between the neural engineering goal of high model-brain predictivity and the neuroscientific goal of identifying mechanistically accurate brain models. Using IATC-guided transforms, we obtain new evidence in favor of topographical deep neural networks (TDANNs) as models of the visual system. Overall, the IATC enables principled model-brain comparisons, contextualizing previous findings about the predictive success of deep learning models of the brain, while improving upon previous approaches to model-brain comparison.

---

### Graph Generation with Spectral Geodesic Flow Matching

Authors: Xikun Huang, Tianyu Ruan, Chihao Zhang, Shihua Zhang
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.02520v1

Graph generation is a fundamental task with wide applications in modeling complex systems. Although existing methods align the spectrum or degree profile of the target graph, they often ignore the geometry induced by eigenvectors and the global structure of the graph. In this work, we propose Spectral Geodesic Flow Matching (SFMG), a novel framework that uses spectral eigenmaps to embed both input and target graphs into continuous Riemannian manifolds. We then define geodesic flows between embeddings and match distributions along these flows to generate output graphs. Our method yields several advantages: (i) captures geometric structure beyond eigenvalues, (ii) supports flexible generation of diverse graphs, and (iii) scales efficiently. Empirically, SFMG matches the performance of state-of-the-art approaches on graphlet, degree, and spectral metrics across diverse benchmarks. In particular, it achieves up to 30$\times$ speedup over diffusion-based models, offering a substantial advantage in scalability and training efficiency. We also demonstrate its ability to generalize to unseen graph scales. Overall, SFMG provides a new approach to graph synthesis by integrating spectral geometry with flow matching.

---

### In-memory Training on Analog Devices with Limited Conductance States via Multi-tile Residual Learning

Authors: Jindan Li, Zhaoxian Wu, Gaowen Liu, Tayfun Gokmen, Tianyi Chen
Published: 2025-10-02
Categories: cs.LG, cs.AR, math.OC
arXiv: https://arxiv.org/abs/2510.02516v1

Analog in-memory computing (AIMC) accelerators enable efficient deep neural network computation directly within memory using resistive crossbar arrays, where model parameters are represented by the conductance states of memristive devices. However, effective in-memory training typically requires at least 8-bit conductance states to match digital baselines. Realizing such fine-grained states is costly and often requires complex noise mitigation techniques that increase circuit complexity and energy consumption. In practice, many promising memristive devices such as ReRAM offer only about 4-bit resolution due to fabrication constraints, and this limited update precision substantially degrades training accuracy. To enable on-chip training with these limited-state devices, this paper proposes a \emph{residual learning} framework that sequentially learns on multiple crossbar tiles to compensate the residual errors from low-precision weight updates. Our theoretical analysis shows that the optimality gap shrinks with the number of tiles and achieves a linear convergence rate. Experiments on standard image classification benchmarks demonstrate that our method consistently outperforms state-of-the-art in-memory analog training strategies under limited-state settings, while incurring only moderate hardware overhead as confirmed by our cost analysis.

---

### Learning a distance measure from the information-estimation geometry of data

Authors: Guy Ohayon, Pierre-Etienne H. Fiquet, Florentin Guth, Jona Ballé, Eero P. Simoncelli
Published: 2025-10-02
Categories: eess.IV, cs.CV, cs.IT, eess.SP, stat.ML
arXiv: https://arxiv.org/abs/2510.02514v2

We introduce the Information-Estimation Metric (IEM), a novel form of distance function derived from an underlying continuous probability density over a domain of signals. The IEM is rooted in a fundamental relationship between information theory and estimation theory, which links the log-probability of a signal with the errors of an optimal denoiser, applied to noisy observations of the signal. In particular, the IEM between a pair of signals is obtained by comparing their denoising error vectors over a range of noise amplitudes. Geometrically, this amounts to comparing the score vector fields of the blurred density around the signals over a range of blur levels. We prove that the IEM is a valid global distance metric and derive a closed-form expression for its local second-order approximation, which yields a Riemannian metric. For Gaussian-distributed signals, the IEM coincides with the Mahalanobis distance. But for more complex distributions, it adapts, both locally and globally, to the geometry of the distribution. In practice, the IEM can be computed using a learned denoiser (analogous to generative diffusion models) and solving a one-dimensional integral. To demonstrate the value of our framework, we learn an IEM on the ImageNet database. Experiments show that this IEM is competitive with or outperforms state-of-the-art supervised image quality metrics in predicting human perceptual judgments.

---

### Adaptive randomized pivoting and volume sampling

Authors: Ethan N. Epperly
Published: 2025-10-02
Categories: stat.ML, cs.DS, cs.LG, math.NA, stat.CO
arXiv: https://arxiv.org/abs/2510.02513v2

Adaptive randomized pivoting (ARP) is a recently proposed and highly effective algorithm for column subset selection. This paper reinterprets the ARP algorithm by drawing connections to the volume sampling distribution and active learning algorithms for linear regression. As consequences, this paper presents new analysis for the ARP algorithm and faster implementations using rejection sampling.

---

### Beyond Linear Diffusions: Improved Representations for Rare Conditional Generative Modeling

Authors: Kulunu Dharmakeerthi, Yousef El-Laham, Henry H. Wong, Vamsi K. Potluru, Changhong He, Taosong He
Published: 2025-10-02
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2510.02499v1

Diffusion models have emerged as powerful generative frameworks with widespread applications across machine learning and artificial intelligence systems. While current research has predominantly focused on linear diffusions, these approaches can face significant challenges when modeling a conditional distribution, $P(Y|X=x)$, when $P(X=x)$ is small. In these regions, few samples, if any, are available for training, thus modeling the corresponding conditional density may be difficult. Recognizing this, we show it is possible to adapt the data representation and forward scheme so that the sample complexity of learning a score-based generative model is small in low probability regions of the conditioning space. Drawing inspiration from conditional extreme value theory we characterize this method precisely in the special case in the tail regions of the conditioning variable, $X$. We show how diffusion with a data-driven choice of nonlinear drift term is best suited to model tail events under an appropriate representation of the data. Through empirical validation on two synthetic datasets and a real-world financial dataset, we demonstrate that our tail-adaptive approach significantly outperforms standard diffusion models in accurately capturing response distributions at the extreme tail conditions.

---

### Beyond Imitation: Recovering Dense Rewards from Demonstrations

Authors: Jiangnan Li, Thuy-Trang Vu, Ehsan Abbasnejad, Gholamreza Haffari
Published: 2025-10-02
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2510.02493v1

Conventionally, supervised fine-tuning (SFT) is treated as a simple imitation learning process that only trains a policy to imitate expert behavior on demonstration datasets. In this work, we challenge this view by establishing a fundamental equivalence between SFT and Inverse Reinforcement Learning. We prove that the SFT objective is a special case of Inverse Q-Learning, which implies that the SFT process does not just learn a policy, but also an implicit, dense, token-level reward model that explains the expert demonstrations. We then show how to recover this dense reward signal directly from the SFT model by formulating a baseline-relative reward function. The availability of such a dense reward model offers numerous benefits, providing granular credit assignment for each token generated. We demonstrate one key application by using these recovered rewards to further improve the policy with reinforcement learning. Our method, Dense-Path REINFORCE, consistently outperforms the original SFT models on instruction-following benchmarks. This work reframes SFT not merely as policy imitation but as a powerful reward learning mechanism, opening new possibilities for leveraging expert demonstrations.

---

### Improved Robustness of Deep Reinforcement Learning for Control of Time-Varying Systems by Bounded Extremum Seeking

Authors: Shaifalee Saxena, Alan Williams, Rafael Fierro, Alexander Scheinker
Published: 2025-10-02
Categories: cs.LG, eess.SY
arXiv: https://arxiv.org/abs/2510.02490v2

In this paper, we study the use of robust model independent bounded extremum seeking (ES) feedback control to improve the robustness of deep reinforcement learning (DRL) controllers for a class of nonlinear time-varying systems. DRL has the potential to learn from large datasets to quickly control or optimize the outputs of many-parameter systems, but its performance degrades catastrophically when the system model changes rapidly over time. Bounded ES can handle time-varying systems with unknown control directions, but its convergence speed slows down as the number of tuned parameters increases and, like all local adaptive methods, it can get stuck in local minima. We demonstrate that together, DRL and bounded ES result in a hybrid controller whose performance exceeds the sum of its parts with DRL taking advantage of historical data to learn how to quickly control a many-parameter system to a desired setpoint while bounded ES ensures its robustness to time variations. We present a numerical study of a general time-varying system and a combined ES-DRL controller for automatic tuning of the Low Energy Beam Transport section at the Los Alamos Neutron Science Center linear particle accelerator.

---

### From Pixels to Factors: Learning Independently Controllable State Variables for Reinforcement Learning

Authors: Rafael Rodriguez-Sanchez, Cameron Allen, George Konidaris
Published: 2025-10-02
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.02484v1

Algorithms that exploit factored Markov decision processes are far more sample-efficient than factor-agnostic methods, yet they assume a factored representation is known a priori -- a requirement that breaks down when the agent sees only high-dimensional observations. Conversely, deep reinforcement learning handles such inputs but cannot benefit from factored structure. We address this representation problem with Action-Controllable Factorization (ACF), a contrastive learning approach that uncovers independently controllable latent variables -- state components each action can influence separately. ACF leverages sparsity: actions typically affect only a subset of variables, while the rest evolve under the environment's dynamics, yielding informative data for contrastive training. ACF recovers the ground truth controllable factors directly from pixel observations on three benchmarks with known factored structure -- Taxi, FourRooms, and MiniGrid-DoorKey -- consistently outperforming baseline disentanglement algorithms.

---

### Litespark Technical Report: High-Throughput, Energy-Efficient LLM Training Framework

Authors: Nii Osae Osae Dade, Moinul Hossain Rahat
Published: 2025-10-02
Categories: cs.LG, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2510.02483v1

Training Large Language Models (LLMs) is plagued by long training times and massive energy consumption, with modern models requiring months of computation and gigawatt-hours of electricity. In light of these challenges,we introduce Litespark, a novel pre-training framework that addresses these inefficiencies through targeted optimizations to transformer attention and MLP layers. Our approach combines architectural improvements with algorithmic enhancements to maximize Model FLOPs Utilization (MFU) while maintaining compatibility with standard transformer implementations. Comprehensive benchmarking on 3B and 30B parameter Llama models using the SlimPajama-627B dataset demonstrates substantial performance gains: 2x-6x training throughput improvement and $55\%-83$% energy consumption reduction across multi-node H200 GPU clusters. These optimizations are model- and hardware-agnostic, enabling broad applicability across transformer architectures and extending to post-training phases including supervised fine-tuning and direct preference optimization.

---

### Controlling the Risk of Corrupted Contexts for Language Models via Early-Exiting

Authors: Andrea Wynn, Metod Jazbec, Charith Peris, Rinat Khaziev, Anqi Liu, Daniel Khashabi, Eric Nalisnick
Published: 2025-10-02
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.02480v3

Large language models (LLMs) can be influenced by harmful or irrelevant context, which can significantly harm model performance on downstream tasks. This motivates principled designs in which LLM systems include built-in mechanisms to guard against such "garbage in, garbage out" scenarios. We propose a novel approach to limit the degree to which harmful context can degrade model performance. First, we define a baseline "safe" behavior for the model -- the model's performance given no context at all (zero-shot). Next, we apply distribution-free risk control (DFRC) to control the extent to which the user-provided context can decay performance below this safe zero-shot baseline. We achieve this by leveraging dynamic early exit prediction, ignoring later attention heads that attend the most to the unsafe inputs. Finally, we propose modifications to DFRC that allow it to both control risk for harmful inputs \textit{and} leverage performance and efficiency gains on helpful inputs. We present both theoretical and empirical results across 9 tasks spanning in-context learning and open-ended question answering, showing that our approach can effectively control risk for harmful context and simultaneously achieve substantial computational efficiency gains with helpful context.

---

### Uncertainty-Guided Model Selection for Tabular Foundation Models in Biomolecule Efficacy Prediction

Authors: Jie Li, Andrew McCarthy, Zhizhuo Zhang, Stephen Young
Published: 2025-10-02
Categories: cs.LG, q-bio.QM
arXiv: https://arxiv.org/abs/2510.02476v2

In-context learners like TabPFN are promising for biomolecule efficacy prediction, where established molecular feature sets and relevant experimental results can serve as powerful contextual examples. However, their performance is highly sensitive to the provided context, making strategies like post-hoc ensembling of models trained on different data subsets a viable approach. An open question is how to select the best models for the ensemble without access to ground truth labels. In this study, we investigate an uncertainty-guided strategy for model selection. We demonstrate on an siRNA knockdown efficacy task that a TabPFN model using straightforward sequence-based features can surpass specialized state-of-the-art predictors. We also show that the model's predicted inter-quantile range (IQR), a measure of its uncertainty, has a negative correlation with true prediction error. We developed the OligoICP method, which selects and averages an ensemble of models with the lowest mean IQR for siRNA efficacy prediction, achieving superior performance compared to naive ensembling or using a single model trained on all available data. This finding highlights model uncertainty as a powerful, label-free heuristic for optimizing biomolecule efficacy predictions.

---

### Heterogeneous Graph Representation of Stiffened Panels with Non-Uniform Boundary Conditions and Loads

Authors: Yuecheng Cai, Jasmin Jelovica
Published: 2025-10-02
Categories: cs.CE, cs.LG
arXiv: https://arxiv.org/abs/2510.02472v1

Surrogate models are essential in structural analysis and optimization. We propose a heterogeneous graph representation of stiffened panels that accounts for geometrical variability, non-uniform boundary conditions, and diverse loading scenarios, using heterogeneous graph neural networks (HGNNs). The structure is partitioned into multiple structural units, such as stiffeners and the plates between them, with each unit represented by three distinct node types: geometry, boundary, and loading nodes. Edge heterogeneity is introduced by incorporating local orientations and spatial relationships of the connecting nodes. Several heterogeneous graph representations, each with varying degrees of heterogeneity, are proposed and analyzed. These representations are implemented into a heterogeneous graph transformer (HGT) to predict von Mises stress and displacement fields across stiffened panels, based on loading and degrees of freedom at their boundaries. To assess the efficacy of our approach, we conducted numerical tests on panels subjected to patch loads and box beams composed of stiffened panels under various loading conditions. The heterogeneous graph representation was compared with a homogeneous counterpart, demonstrating superior performance. Additionally, an ablation analysis was performed to evaluate the impact of graph heterogeneity on HGT performance. The results show strong predictive accuracy for both displacement and von Mises stress, effectively capturing structural behavior patterns and maximum values.

---

### Predictive inference for time series: why is split conformal effective despite temporal dependence?

Authors: Rina Foygel Barber, Ashwin Pananjady
Published: 2025-10-02
Categories: stat.ML, cs.LG, math.ST
arXiv: https://arxiv.org/abs/2510.02471v2

We consider the problem of uncertainty quantification for prediction in a time series: if we use past data to forecast the next time point, can we provide valid prediction intervals around our forecasts? To avoid placing distributional assumptions on the data, in recent years the conformal prediction method has been a popular approach for predictive inference, since it provides distribution-free coverage for any iid or exchangeable data distribution. However, in the time series setting, the strong empirical performance of conformal prediction methods is not well understood, since even short-range temporal dependence is a strong violation of the exchangeability assumption. Using predictors with "memory" -- i.e., predictors that utilize past observations, such as autoregressive models -- further exacerbates this problem. In this work, we examine the theoretical properties of split conformal prediction in the time series setting, including the case where predictors may have memory. Our results bound the loss of coverage of these methods in terms of a new "switch coefficient", measuring the extent to which temporal dependence within the time series creates violations of exchangeability. Our characterization of the coverage probability is sharp over the class of stationary, $β$-mixing processes. Along the way, we introduce tools that may prove useful in analyzing other predictive inference methods for dependent data.

---

### SAGE: Streaming Agreement-Driven Gradient Sketches for Representative Subset Selection

Authors: Ashish Jha, Salman Ahmadi-Asl
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.02470v2

Training modern neural networks on large datasets is computationally and energy intensive. We present SAGE, a streaming data-subset selection method that maintains a compact Frequent Directions (FD) sketch of gradient geometry in $O(\ell D)$ memory and prioritizes examples whose sketched gradients align with a consensus direction. The approach eliminates $N \times N$ pairwise similarities and explicit $N \times \ell$ gradient stores, yielding a simple two-pass, GPU-friendly pipeline. Leveraging FD's deterministic approximation guarantees, we analyze how agreement scoring preserves gradient energy within the principal sketched subspace. Across multiple benchmarks, SAGE trains with small kept-rate budgets while retaining competitive accuracy relative to full-data training and recent subset-selection baselines, and reduces end-to-end compute and peak memory. Overall, SAGE offers a practical, constant-memory alternative that complements pruning and model compression for efficient training.

---

### SIMSplat: Language-Aligned 4D Gaussian Splatting for Driving Scenario Generation

Authors: Sung-Yeon Park, Adam Lee, Juanwu Lu, Can Cui, Luyang Jiang, Rohit Gupta, Kyungtae Han, Ahmadreza Moradipari, Ziran Wang
Published: 2025-10-02
Categories: cs.RO, cs.AI, cs.CL, cs.CV
arXiv: https://arxiv.org/abs/2510.02469v2

Driving scene manipulation using real-world sensor data has emerged as a promising alternative to traditional driving simulators. Despite advances in language control and neural scene representations, existing methods treat grounding, editing, and simulation as loosely connected stages, relying on heuristic object localization, manual guidance, and single-agent validation, thereby constraining semantic expressiveness and hindering scalable, reactive scenario generation. We introduce SIMSplat, a driving scene editor built on scene-graph-based 4D Gaussian Splatting augmented with language-aligned features. By embedding appearance, motion, and location semantics directly into Gaussian scene-graph nodes, SIMSplat makes reconstructed scenes queryable through free-form natural language, bridging language understanding to object-level editing and multi-agent simulation within a single framework. Building on this language-grounded scene graph, SIMSplat supports diverse edits including fine-grained pedestrian manipulation, while a multi-agent path refinement module propagates changes across all agents to ensure reactive, physically plausible simulations. The pipeline further integrates with Vision-Language Models for automated scenario mining. Experiments show that SIMSplat more than doubles baseline grounding accuracy, achieves the highest task completion rate, and produces the lowest failure rates across diverse driving scenarios.

---

### CLARITY: Clinical Assistant for Routing, Inference, and Triage

Authors: Vladimir Shaposhnikov, Aleksandr Nesterov, Ilia Kopanichuk, Ivan Bakulin, Egor Zhelvakov, Ruslan Abramov, Ekaterina Tsapieva, Iaroslav Bespalov, Dmitry V. Dylov, Ivan Oseledets
Published: 2025-10-02
Categories: cs.CL, cs.AI, cs.MA
arXiv: https://arxiv.org/abs/2510.02463v2

We present CLARITY (Clinical Assistant for Routing, Inference and Triage), an AI-driven platform designed to facilitate patient-to-specialist routing, clinical consultations, and severity assessment of patient conditions. Its hybrid architecture combines a Finite State Machine (FSM) for structured dialogue flows with collaborative agents that employ Large Language Model (LLM) to analyze symptoms and prioritize referrals to appropriate specialists. Built on a modular microservices framework, CLARITY ensures safe, efficient, and robust performance, flexible and readily scalable to meet the demands of existing workflows and IT solutions in healthcare. We report integration of our clinical assistant into a large-scale national interhospital platform, with more than 55,000 content-rich user dialogues completed within the two months of deployment, 2,500 of which were expert-annotated for subsequent validation. The validation results show that CLARITY surpasses human-level performance in terms of the first-attempt routing precision, naturally requiring up to 3 times shorter duration of the consultation than with a human.

---

### Assessing the Potential for Catastrophic Failure in Dynamic Post-Training Quantization

Authors: Logan Frank, Paul Ardis
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.02457v1

Post-training quantization (PTQ) has recently emerged as an effective tool for reducing the computational complexity and memory usage of a neural network by representing its weights and activations with lower precision. While this paradigm has shown great success in lowering compute and storage costs, there is the potential for drastic performance reduction depending upon the distribution of inputs experienced in inference. When considering possible deployment in safety-critical environments, it is important to investigate the extent of potential performance reduction, and what characteristics of input distributions may give rise to this reduction. In this work, we explore the idea of extreme failure stemming from dynamic PTQ and formulate a knowledge distillation and reinforcement learning task to learn a network and bit-width policy pair such that catastrophic failure under quantization is analyzed in terms of worst case potential. Our results confirm the existence of this "detrimental" network-policy pair, with several instances demonstrating performance reductions in the range of 10-65% in accuracy, compared to their "robust" counterparts encountering a <2% decrease. From systematic experimentation and analyses, we also provide an initial exploration into points at highest vulnerability. While our results represent an initial step toward understanding failure cases introduced by PTQ, our findings ultimately emphasize the need for caution in real-world deployment scenarios. We hope this work encourages more rigorous examinations of robustness and a greater emphasis on safety considerations for future works within the broader field of deep learning.

---

### Market-Driven Subset Selection for Budgeted Training

Authors: Ashish Jha, Valentin Leplat, AH Phan
Published: 2025-10-02
Categories: cs.LG, cs.AI, math.NA
arXiv: https://arxiv.org/abs/2510.02456v2

Training large language models on massive datasets is computationally expensive, yet empirical evidence suggests that substantial portions of training examples contribute minimally to final performance. Data subset selection addresses this inefficiency by identifying small, high-utility subsets under resource constraints. However, example utility is inherently multi-faceted, encompassing uncertainty, distributional rarity, and diversity signals that are heterogeneous and typically combined through ad hoc weighted sums lacking theoretical grounding. We propose a market-based framework that treats each training example as a tradeable contract and employs the Logarithmic Market Scoring Rule to aggregate multiple utility signals into coherent prices. Heterogeneous signals act as traders, a single liquidity parameter controls concentration versus smoothing, and topic-wise normalization ensures calibrated aggregation. Token budgets are handled explicitly through a price-per-token decision rule with an interpretable length-bias parameter. We establish theoretical connections to maximum-entropy aggregation and provide utility recovery guarantees under noisy but monotone signals. On GSM8K mathematical reasoning under strict 60k-token budgets, our selector achieves parity with strong single-signal baselines while exhibiting lower variance and incurring less than 0.1 GPU-hour overhead. On AGNews classification at 5-25\% retention rates, the market formulation delivers competitive accuracy with improved stability. Our framework unifies multi-signal data curation under fixed computational budgets for prompt-level reasoning and classification tasks.

---

### How to Train Your Advisor: Steering Black-Box LLMs with Advisor Models

Authors: Parth Asawa, Alan Zhu, Abigail O'Neill, Matei Zaharia, Alexandros G. Dimakis, Joseph E. Gonzalez
Published: 2025-10-02
Categories: cs.LG, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2510.02453v3

Frontier language models are deployed as black-box services, where model weights cannot be modified and customization is limited to prompting. We introduce Advisor Models, a method to train small open-weight models to generate dynamic, per-instance natural language advice that improves the capabilities of black-box frontier models. Advisor Models improve GPT-5.2's performance on RuleArena (Taxes) by 27.4%, reduce Gemini 3 Pro's steps taken in SWE agent tasks by 24.6%, and outperform static prompt optimizers in personalizing GPT-5 to user preferences (85-100% vs. 40-60%). We also find that advisors are transferable: an advisor trained with a low-cost student model still transfers improvements to a frontier model. Moreover, Advisor Models are robust: we observe no degradation on other benchmarks than the pipeline is trained on. Our method shows how to perform parametric optimization for black-box frontier models in a practical and cost-effective way.

---

### Words That Make Language Models Perceive

Authors: Sophie L. Wang, Phillip Isola, Brian Cheung
Published: 2025-10-02
Categories: cs.CL, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.02425v1

Large language models (LLMs) trained purely on text ostensibly lack any direct perceptual experience, yet their internal representations are implicitly shaped by multimodal regularities encoded in language. We test the hypothesis that explicit sensory prompting can surface this latent structure, bringing a text-only LLM into closer representational alignment with specialist vision and audio encoders. When a sensory prompt tells the model to 'see' or 'hear', it cues the model to resolve its next-token predictions as if they were conditioned on latent visual or auditory evidence that is never actually supplied. Our findings reveal that lightweight prompt engineering can reliably activate modality-appropriate representations in purely text-trained LLMs.

---

### Adaptive Deception Framework with Behavioral Analysis for Enhanced Cybersecurity Defense

Authors: Basil Abdullah AL-Zahrani
Published: 2025-10-02
Categories: cs.CR, cs.LG
arXiv: https://arxiv.org/abs/2510.02424v1

This paper presents CADL (Cognitive-Adaptive Deception Layer), an adaptive deception framework achieving 99.88% detection rate with 0.13% false positive rate on the CICIDS2017 dataset. The framework employs ensemble machine learning (Random Forest, XGBoost, Neural Networks) combined with behavioral profiling to identify and adapt responses to network intrusions. Through a coordinated signal bus architecture, security components share real-time intelligence, enabling collective decision-making. The system profiles attackers based on temporal patterns and deploys customized deception strategies across five escalation levels. Evaluation on 50,000 CICIDS2017 test samples demonstrates that CADL significantly outperforms traditional intrusion detection systems (Snort: 71.2%, Suricata: 68.5%) while maintaining production-ready false positive rates. The framework's behavioral analysis achieves 89% accuracy in classifying attacker profiles. We provide open-source implementation and transparent performance metrics, offering an accessible alternative to commercial deception platforms costing $150-400 per host annually.

---

### RefineShot: Rethinking Cinematography Understanding with Foundational Skill Evaluation

Authors: Hang Wu, Yujun Cai, Haonan Ge, Hongkai Chen, Ming-Hsuan Yang, Yiwei Wang
Published: 2025-10-02
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.02423v1

Cinematography understanding refers to the ability to recognize not only the visual content of a scene but also the cinematic techniques that shape narrative meaning. This capability is attracting increasing attention, as it enhances multimodal understanding in real-world applications and underpins coherent content creation in film and media. As the most comprehensive benchmark for this task, ShotBench spans a wide range of cinematic concepts and VQA-style evaluations, with ShotVL achieving state-of-the-art results on it. However, our analysis reveals that ambiguous option design in ShotBench and ShotVL's shortcomings in reasoning consistency and instruction adherence undermine evaluation reliability, limiting fair comparison and hindering future progress. To overcome these issues, we systematically refine ShotBench through consistent option restructuring, conduct the first critical analysis of ShotVL's reasoning behavior, and introduce an extended evaluation protocol that jointly assesses task accuracy and core model competencies. These efforts lead to RefineShot, a refined and expanded benchmark that enables more reliable assessment and fosters future advances in cinematography understanding.

---

### Dynamic Jailbreaking Attack

Authors: Kedong Xiu, Yunhan Yang, Churui Zeng, Tianhang Zheng, Xinzhe Huang, Di Wang, Puning Zhao, Zhan Qin, Kui Ren
Published: 2025-10-02
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2510.02422v4

Existing gradient-based jailbreak attacks typically optimize a fixed-length adversarial suffix toward a predefined target response with a static optimization strategy. However, this fully static formulation undermines the effectiveness, efficiency and flexibility of gradient-based jailbreaking because (i) A predefined target usually lies in the low-probability region of a safety-aligned LLM's conditional output distribution, forcing the optimization to pursue an unlikely response pattern; (ii) Simple affirmative targets may even mislead LLMs to generate affirmative responses that are not highly relevant to the prompts; (iii) Fixed optimization strategy and suffix length treat all prompts equally, leading to limited attack capability for hard prompts and redundant capacity for easy ones. To address these limitations, we propose Dynamic Jailbreaking Attack (DJA), a parameter-free gradient-based jailbreak framework using dynamic candidate exploration, dynamic relevant targets and dynamic optimization strategy to craft adversarial prompts. In each optimization round, DJA samples multiple candidate target responses directly from the LLM's distribution conditioned on the current adversarial prompt. Among these candidates, DJA employs a multi-objective scorer to select an optimal target that satisfies multi-dimensional criteria such as harmfulness, relevance, and usefulness. Moreover, DJA introduces a parameter-free dynamic optimization strategy that allocates adversarial effort based on real-time feedback, adapting suffix length, candidate sampling capacity, and optimization iterations according to the difficulty of each harmful prompt. In an extensive evaluation of 40 safety-aligned LLMs (12 model families, scaling from 0.5B to 32B), DJA achieves a 100% ASR across all LLMs, requiring only 13.68 optimization rounds on average (10 iterations per round).

---

### Higher-arity PAC learning, VC dimension and packing lemma

Authors: Artem Chernikov, Henry Towsner
Published: 2025-10-02
Categories: stat.ML, cs.DM, cs.LG, math.CO, math.LO, math.ST
arXiv: https://arxiv.org/abs/2510.02420v2

The aim of this note is to overview some of our work in Chernikov, Towsner'20 (arXiv:2010.00726) developing higher arity VC theory (VC$_n$ dimension), including a generalization of Haussler packing lemma, and an associated tame (slice-wise) hypergraph regularity lemma; and to demonstrate that it characterizes higher arity PAC learning (PAC$_n$ learning) in $n$-fold product spaces with respect to product measures introduced by Kobayashi, Kuriyama and Takeuchi'15. We also point out how some of the recent results in arXiv:2402.14294, arXiv:2505.15688, arXiv:2509.20404 follow from our work in arXiv:2010.00726.

---

### BrowserArena: Evaluating LLM Agents on Real-World Web Navigation Tasks

Authors: Sagnik Anupam, Davis Brown, Shuo Li, Eric Wong, Hamed Hassani, Osbert Bastani
Published: 2025-10-02
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.02418v2

LLM web agents now browse and take actions on the open web, yet current agent evaluations are constrained to sandboxed environments or artificial tasks. We introduce BrowserArena, a live open-web agent evaluation platform that collects user-submitted tasks, runs Arena-style head-to-head comparisons, and uses step-level human feedback to surface failure modes. Collecting and analyzing step-level annotations on the agent traces, we identify three consistent failure modes: captcha resolution, pop-up banner removal, and direct navigation to URLs. By constructing targeted datasets to further study these tasks, we discover variations in how different language models navigate these failure modes. We find, for example, that o4-mini deploys a wider variety of strategies to circumvent captcha resolution than other models and DeepSeek-R1 consistently misleads users about pop-up banner closure. Our findings surface both the diversity and brittleness of current web agents. More broadly, our benchmarking methodology provides an approach to evaluating and understanding web agent failure modes at scale.

---

### NEURODNAAI: Neural pipeline approaches for the advancing dna-based information storage as a sustainable digital medium using deep learning framework

Authors: Rakesh Thakur, Lavanya Singh, Yashika, Manomay Bundawala, Aruna Kumar
Published: 2025-10-02
Categories: cs.ET, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.02417v1

DNA is a promising medium for digital information storage for its exceptional density and durability. While prior studies advanced coding theory, workflow design, and simulation tools, challenges such as synthesis costs, sequencing errors, and biological constraints (GC-content imbalance, homopolymers) limit practical deployment. To address this, our framework draws from quantum parallelism concepts to enhance encoding diversity and resilience, integrating biologically informed constraints with deep learning to enhance error mitigation in DNA storage. NeuroDNAAI encodes binary data streams into symbolic DNA sequences, transmits them through a noisy channel with substitutions, insertions, and deletions, and reconstructs them with high fidelity. Our results show that traditional prompting or rule-based schemes fail to adapt effectively to realistic noise, whereas NeuroDNAAI achieves superior accuracy. Experiments on benchmark datasets demonstrate low bit error rates for both text and images. By unifying theory, workflow, and simulation into one pipeline, NeuroDNAAI enables scalable, biologically valid archival DNA storage

---

### Cross-Platform DNA Methylation Classifier for the Eight Molecular Subtypes of Group 3 & 4 Medulloblastoma

Authors: Omer Abid, Gholamreza Rafiee
Published: 2025-10-02
Categories: q-bio.GN, cs.AI
arXiv: https://arxiv.org/abs/2510.02416v1

Medulloblastoma is a malignant pediatric brain cancer, and the discovery of molecular subgroups is enabling personalized treatment strategies. In 2019, a consensus identified eight novel subtypes within Groups 3 and 4, each displaying heterogeneous characteristics. Classifiers are essential for translating these findings into clinical practice by supporting clinical trials, personalized therapy development and application, and patient monitoring. This study presents a DNA methylation-based, cross-platform machine learning classifier capable of distinguishing these subtypes on both HM450 and EPIC methylation array samples. Across two independent test sets, the model achieved weighted F1 = 0.95 and balanced accuracy = 0.957, consistent across platforms. As the first cross-platform solution, it provides backward compatibility while extending applicability to a newer platform, also enhancing accessibility. It also has the potential to become the first publicly available classifier for these subtypes once deployed through a web application, as planned in the future. This work overall takes steps in the direction of advancing precision medicine and improving clinical outcomes for patients within the majority prevalence medulloblastoma subgroups, groups 3 and 4.

---

### The Equilibrium Response of Atmospheric Machine-Learning Models to Uniform Sea Surface Temperature Warming

Authors: Bosong Zhang, Timothy M. Merlis
Published: 2025-10-02
Categories: physics.ao-ph, cs.LG
arXiv: https://arxiv.org/abs/2510.02415v3

Machine learning models for the global atmosphere that are capable of producing stable, multi-year simulations of Earth's climate have recently been developed. However, the ability of these ML models to generalize beyond the training distribution remains an open question. In this study, we evaluate the climate response of several state-of-the-art ML models (ACE2-ERA5, NeuralGCM, and cBottle) to a uniform sea surface temperature warming, a widely used benchmark for evaluating climate change. We assess each ML model's performance relative to a physics-based general circulation model (NOAA's Geophysical Fluid Dynamics Laboratory AM4) across key diagnostics, including surface air temperature, precipitation, temperature and wind profiles, and top-of-atmosphere radiation. While the ML models reproduce key aspects of the physical model response, particularly the response of precipitation, some exhibit notable departures from robust physical responses, including radiative responses and land region warming. Our results highlight the promise and current limitations of ML models for climate change applications and suggest that further improvements are needed for robust out-of-sample generalization.

---

### RainSeer: Fine-Grained Rainfall Reconstruction via Physics-Guided Modeling

Authors: Lin Chen, Jun Chen, Minghui Qiu, Shuxin Zhong, Binghong Chen, Kaishun Wu
Published: 2025-10-02
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.02414v2

Reconstructing high-resolution rainfall fields is essential for flood forecasting, hydrological modeling, and climate analysis. However, existing spatial interpolation methods-whether based on automatic weather station (AWS) measurements or enhanced with satellite/radar observations often over-smooth critical structures, failing to capture sharp transitions and localized extremes. We introduce RainSeer, a structure-aware reconstruction framework that reinterprets radar reflectivity as a physically grounded structural prior-capturing when, where, and how rain develops. This shift, however, introduces two fundamental challenges: (i) translating high-resolution volumetric radar fields into sparse point-wise rainfall observations, and (ii) bridging the physical disconnect between aloft hydro-meteors and ground-level precipitation. RainSeer addresses these through a physics-informed two-stage architecture: a Structure-to-Point Mapper performs spatial alignment by projecting mesoscale radar structures into localized ground-level rainfall, through a bidirectional mapping, and a Geo-Aware Rain Decoder captures the semantic transformation of hydro-meteors through descent, melting, and evaporation via a causal spatiotemporal attention mechanism. We evaluate RainSeer on two public datasets-RAIN-F (Korea, 2017-2019) and MeteoNet (France, 2016-2018)-and observe consistent improvements over state-of-the-art baselines, reducing MAE by over 13.31% and significantly enhancing structural fidelity in reconstructed rainfall fields.

---

### OpenTSLM: Time-Series Language Models for Reasoning over Multivariate Medical Text- and Time-Series Data

Authors: Patrick Langer, Thomas Kaar, Max Rosenblattl, Maxwell A. Xu, Winnie Chow, Martin Maritsch, Robert Jakob, Ning Wang, Juncheng Liu, Aradhana Verma, Brian Han, Daniel Seung Kim et al.
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.02410v3

LLMs have emerged as powerful tools for interpreting multimodal data. In medicine, they hold particular promise for synthesizing large volumes of clinical information into actionable insights and digital health applications. Yet, a major limitation remains their inability to handle time series. To overcome this gap, we present OpenTSLM, a family of Time Series Language Models (TSLMs) created by integrating time series as a native modality to pretrained LLMs, enabling reasoning over multiple time series of any length. We investigate two architectures for OpenTSLM. The first, OpenTSLM-SoftPrompt, models time series implicitly by concatenating learnable time series tokens with text tokens via soft prompting. Although parameter-efficient, we hypothesize that explicit time series modeling scales better and outperforms implicit approaches. We thus introduce OpenTSLM-Flamingo, which integrates time series with text via cross-attention. We benchmark both variants against baselines that treat time series as text tokens or plots, across a suite of text-time-series Chain-of-Thought (CoT) reasoning tasks. We introduce three datasets: HAR-CoT, Sleep-CoT, and ECG-QA-CoT. Across all, OpenTSLM models outperform baselines, reaching 69.9 F1 in sleep staging and 65.4 in HAR, compared to 9.05 and 52.2 for finetuned text-only models. Notably, even 1B-parameter OpenTSLM models surpass GPT-4o (15.47 and 2.95). OpenTSLM-Flamingo matches OpenTSLM-SoftPrompt in performance and outperforms on longer sequences, while maintaining stable memory requirements. By contrast, SoftPrompt grows exponentially in memory with sequence length, requiring around 110 GB compared to 40 GB VRAM when training on ECG-QA with LLaMA-3B. Expert reviews by clinicians find strong reasoning capabilities exhibited by OpenTSLMs on ECG-QA. To facilitate further research, we provide all code, datasets, and models open-source.

---