# Research Papers — 2025-10 · bundle page 179 of 217

> Full-text markdown of papers #8901–8950 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/179
[← Previous bundle](https://aipod.dev/papers/archive/2025-10/page/178.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-10/page/180.md)

### SAEdit: Token-level control for continuous image editing via Sparse AutoEncoder

Authors: Ronen Kamenetsky, Sara Dorfman, Daniel Garibi, Roni Paiss, Or Patashnik, Daniel Cohen-Or
Published: 2025-10-06
Categories: cs.GR, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2510.05081v1

Large-scale text-to-image diffusion models have become the backbone of modern image editing, yet text prompts alone do not offer adequate control over the editing process. Two properties are especially desirable: disentanglement, where changing one attribute does not unintentionally alter others, and continuous control, where the strength of an edit can be smoothly adjusted. We introduce a method for disentangled and continuous editing through token-level manipulation of text embeddings. The edits are applied by manipulating the embeddings along carefully chosen directions, which control the strength of the target attribute. To identify such directions, we employ a Sparse Autoencoder (SAE), whose sparse latent space exposes semantically isolated dimensions. Our method operates directly on text embeddings without modifying the diffusion process, making it model agnostic and broadly applicable to various image synthesis backbones. Experiments show that it enables intuitive and efficient manipulations with continuous control across diverse attributes and domains.

---

### MICROTRIPS: MICRO-geography TRavel Intelligence and Pattern Synthesis

Authors: Yangyang Wang, Tayo Fabusuyi
Published: 2025-10-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.05080v1

This study presents a novel small-area estimation framework to enhance urban transportation planning through detailed characterization of travel behavior. Our approach improves on the four-step travel model by employing publicly available microdata files and machine learning methods to predict travel behavior for a representative, synthetic population at small geographic areas. This approach enables high-resolution estimation of trip generation, trip distribution, mode choice, and route assignment. Validation using ACS/PUMS work-commute datasets demonstrates that our framework achieves higher accuracy compared to conventional approaches. The resulting granular insights enable the tailoring of interventions to address localized situations and support a range of policy applications and targeted interventions, including the optimal placement of micro-fulfillment centers, effective curb-space management, and the design of more inclusive transportation solutions particularly for vulnerable communities.

---

### Slm-mux: Orchestrating small language models for reasoning

Authors: Chenyu Wang, Zishen Wan, Hao Kang, Emma Chen, Zhiqiang Xie, Tushar Krishna, Vijay Janapa Reddi, Yilun Du
Published: 2025-10-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.05077v2

With the rapid development of language models, the number of small language models (SLMs) has grown significantly. Although they do not achieve state-of-the-art accuracy, they are more efficient and often excel at specific tasks. This raises a natural question: can multiple SLMs be orchestrated into a system where each contributes effectively, achieving higher accuracy than any individual model? Existing orchestration methods have primarily targeted frontier models (e.g., GPT-4) and perform suboptimally when applied to SLMs. To address this gap, we propose a three-stage approach for orchestrating SLMs. First, we introduce SLM-MUX, a multi-model architecture that effectively coordinates multiple SLMs. Building on this, we develop two optimization strategies: (i) a model selection search that identifies the most complementary SLMs from a given pool, and (ii) test-time scaling tailored to SLM-MUX. Our approach delivers strong results: Compared to existing orchestration methods, our approach achieves up to 13.4% improvement on MATH, 8.8% on GPQA, and 7.0% on GSM8K. With just two SLMs, SLM-MUX outperforms Qwen 2.5 72B on GPQA and GSM8K, and matches its performance on MATH. We further provide theoretical analyses to substantiate the advantages of our method. Additional experiments show that the core principle of SLM-MUX extends to open-ended generation tasks (e.g., HumanEval) and benefits other model classes, including frontier LLMs and domain-specific fine-tuned SLMs. In summary, we demonstrate that SLMs can be effectively orchestrated into more accurate and efficient systems through the proposed approach. The project page is available at https://slm-mux.github.io/.

---

### Neuroplastic Modular Framework: Cross-Domain Image Classification of Garbage and Industrial Surfaces

Authors: Debojyoti Ghosh, Soumya K Ghosh, Adrijit Goswami
Published: 2025-10-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.05071v1

Efficient and accurate classification of waste and industrial surface defects is essential for ensuring sustainable waste management and maintaining high standards in quality control. This paper introduces the Neuroplastic Modular Classifier, a novel hybrid architecture designed for robust and adaptive image classification in dynamic environments. The model combines a ResNet-50 backbone for localized feature extraction with a Vision Transformer (ViT) to capture global semantic context. Additionally, FAISS-based similarity retrieval is incorporated to provide a memory-like reference to previously encountered data, enriching the model's feature space. A key innovation of our architecture is the neuroplastic modular design composed of expandable, learnable blocks that dynamically grow during training when performance plateaus. Inspired by biological learning systems, this mechanism allows the model to adapt to data complexity over time, improving generalization. Beyond garbage classification, we validate the model on the Kolektor Surface Defect Dataset 2 (KolektorSDD2), which involves industrial defect detection on metal surfaces. Experimental results across domains show that the proposed architecture outperforms traditional static models in both accuracy and adaptability. The Neuroplastic Modular Classifier offers a scalable, high-performance solution for real-world image classification, with strong applicability in both environmental and industrial domains.

---

### ResMimic: From General Motion Tracking to Humanoid Whole-body Loco-Manipulation via Residual Learning

Authors: Siheng Zhao, Yanjie Ze, Yue Wang, C. Karen Liu, Pieter Abbeel, Guanya Shi, Rocky Duan
Published: 2025-10-06
Categories: cs.RO, cs.LG
arXiv: https://arxiv.org/abs/2510.05070v2

Humanoid whole-body loco-manipulation promises transformative capabilities for daily service and warehouse tasks. While recent advances in general motion tracking (GMT) have enabled humanoids to reproduce diverse human motions, these policies lack the precision and object awareness required for loco-manipulation. To this end, we introduce ResMimic, a two-stage residual learning framework for precise and expressive humanoid control from human motion data. First, a GMT policy, trained on large-scale human-only motion, serves as a task-agnostic base for generating human-like whole-body movements. An efficient but precise residual policy is then learned to refine the GMT outputs to improve locomotion and incorporate object interaction. To further facilitate efficient training, we design (i) a point-cloud-based object tracking reward for smoother optimization, (ii) a contact reward that encourages accurate humanoid body-object interactions, and (iii) a curriculum-based virtual object controller to stabilize early training. We evaluate ResMimic in both simulation and on a real Unitree G1 humanoid. Results show substantial gains in task success, training efficiency, and robustness over strong baselines. Videos are available at https://resmimic.github.io/ .

---

### SwiReasoning: Switch-Thinking in Latent and Explicit for Pareto-Superior Reasoning LLMs

Authors: Dachuan Shi, Abedelkadir Asi, Keying Li, Xiangchi Yuan, Leyan Pan, Wenke Lee, Wen Xiao
Published: 2025-10-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.05069v3

Recent work shows that, beyond discrete reasoning through explicit chain-of-thought steps, which are limited by the boundaries of natural languages, large language models (LLMs) can also reason continuously in latent space, allowing richer information per step and thereby improving token efficiency. Despite this promise, latent reasoning still faces two challenges, especially in training-free settings: 1) purely latent reasoning broadens the search distribution by maintaining multiple implicit paths, which diffuses probability mass, introduces noise, and impedes convergence to a single high-confidence solution, thereby hurting accuracy; and 2) overthinking persists even without explicit text, wasting tokens and degrading efficiency. To address these issues, we introduce SwiReasoning, a training-free framework for LLM reasoning which features two key innovations: 1) SwiReasoning dynamically switches between explicit and latent reasoning, guided by block-wise confidence estimated from entropy trends in next-token distributions, to balance exploration and exploitation and promote timely convergence. 2) By limiting the maximum number of thinking-block switches, SwiReasoning curbs overthinking and improves token efficiency across varying problem difficulties. On widely used mathematics, STEM, coding, and general benchmarks, SwiReasoning consistently improves average accuracy by 1.8%-3.1% across reasoning LLMs of different model families and scales. Furthermore, under constrained budgets, SwiReasoning improves average token efficiency by 57%-79%, with larger gains as budgets tighten.

---

### Boomerang Distillation Enables Zero-Shot Model Size Interpolation

Authors: Sara Kangaslahti, Nihal V. Nayak, Jonathan Geuter, Marco Fumero, Francesco Locatello, David Alvarez-Melis
Published: 2025-10-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.05064v2

Large language models (LLMs) are typically deployed under diverse memory and compute constraints. Existing approaches build model families by training each size independently, which is prohibitively expensive and provides only coarse-grained size options. In this work, we identify a novel phenomenon that we call boomerang distillation: starting from a large base model (the teacher), one first distills down to a small student and then progressively reconstructs intermediate-sized models by re-incorporating blocks of teacher layers into the student without any additional training. This process produces zero-shot interpolated models of many intermediate sizes whose performance scales smoothly between the student and teacher, often matching or surpassing pretrained or distilled models of the same size. We further analyze when this type of interpolation succeeds, showing that alignment between teacher and student through pruning and distillation is essential. Boomerang distillation thus provides a simple and efficient way to generate fine-grained model families, dramatically reducing training cost while enabling flexible adaptation across deployment environments. The code and models are available at https://github.com/dcml-lab/boomerang-distillation.

---

### ResCP: Reservoir Conformal Prediction for Time Series Forecasting

Authors: Roberto Neglia, Andrea Cini, Michael M. Bronstein, Filippo Maria Bianchi
Published: 2025-10-06
Categories: cs.LG, math.ST, stat.ML
arXiv: https://arxiv.org/abs/2510.05060v3

Conformal prediction offers a powerful framework for building distribution-free prediction intervals for exchangeable data. Existing methods that extend conformal prediction to sequential data rely on fitting a relatively complex model to capture temporal dependencies. However, these methods can fail if the sample size is small and often require expensive retraining when the underlying data distribution changes. To overcome these limitations, we propose Reservoir Conformal Prediction (ResCP), a novel training-free conformal prediction method for time series. Our approach leverages the efficiency and representation learning capabilities of reservoir computing to dynamically reweight conformity scores. In particular, we compute similarity scores among reservoir states and use them to adaptively reweight the observed residuals at each step. With this approach, ResCP enables us to account for local temporal dynamics when modeling the error distribution without compromising computational scalability. We prove that, under reasonable assumptions, ResCP achieves asymptotic conditional coverage, and we empirically demonstrate its effectiveness across diverse forecasting tasks.

---

### Staircase Streaming for Low-Latency Multi-Agent Inference

Authors: Junlin Wang, Jue Wang, Zhen, Xu, Ben Athiwaratkun, Bhuwan Dhingra, Ce Zhang, James Zou
Published: 2025-10-06
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.05059v1

Recent advances in large language models (LLMs) opened up new directions for leveraging the collective expertise of multiple LLMs. These methods, such as Mixture-of-Agents, typically employ additional inference steps to generate intermediate outputs, which are then used to produce the final response. While multi-agent inference can enhance response quality, it can significantly increase the time to first token (TTFT), posing a challenge for latency-sensitive applications and hurting user experience. To address this issue, we propose staircase streaming for low-latency multi-agent inference. Instead of waiting for the complete intermediate outputs from previous steps, we begin generating the final response as soon as we receive partial outputs from these steps. Experimental results demonstrate that staircase streaming reduces TTFT by up to 93% while maintaining response quality.

---

### StaMo: Unsupervised Learning of Generalizable Robot Motion from Compact State Representation

Authors: Mingyu Liu, Jiuhe Shu, Hui Chen, Zeju Li, Canyu Zhao, Jiange Yang, Shenyuan Gao, Hao Chen, Chunhua Shen
Published: 2025-10-06
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2510.05057v2

A fundamental challenge in embodied intelligence is developing expressive and compact state representations for efficient world modeling and decision making. However, existing methods often fail to achieve this balance, yielding representations that are either overly redundant or lacking in task-critical information. We propose an unsupervised approach that learns a highly compressed two-token state representation using a lightweight encoder and a pre-trained Diffusion Transformer (DiT) decoder, capitalizing on its strong generative prior. Our representation is efficient, interpretable, and integrates seamlessly into existing VLA-based models, improving performance by 14.3% on LIBERO and 30% in real-world task success with minimal inference overhead. More importantly, we find that the difference between these tokens, obtained via latent interpolation, naturally serves as a highly effective latent action, which can be further decoded into executable robot actions. This emergent capability reveals that our representation captures structured dynamics without explicit supervision. We name our method StaMo for its ability to learn generalizable robotic Motion from compact State representation, which is encoded from static images, challenging the prevalent dependence to learning latent action on complex architectures and video data. The resulting latent actions also enhance policy co-training, outperforming prior methods by 10.4% with improved interpretability. Moreover, our approach scales effectively across diverse data sources, including real-world robot data, simulation, and human egocentric video.

---

### Modeling Student Learning with 3.8 Million Program Traces

Authors: Alexis Ross, Megha Srivastava, Jeremiah Blanchard, Jacob Andreas
Published: 2025-10-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.05056v2

As programmers write code, they often edit and retry multiple times, creating rich "interaction traces" that reveal how they approach coding tasks and provide clues about their level of skill development. For novice programmers in particular, these traces reflect the diverse reasoning processes they employ to code, such as exploratory behavior to understand how a programming concept works, re-strategizing in response to bugs, and personalizing stylistic choices. In this work, we explore what can be learned from training language models on such reasoning traces: not just about code, but about coders, and particularly students learning to program. We introduce a dataset of over 3.8 million programming reasoning traces from users of Pencil Code, a free online educational platform used by students to learn simple programming concepts. Compared to models trained only on final programs or synthetically-generated traces, we find that models trained on real traces are stronger at modeling diverse student behavior. Through both behavioral and probing analyses, we also find that many properties of code traces, such as goal backtracking or number of comments, can be predicted from learned representations of the students who write them. Building on this result, we show that we can help students recover from mistakes by steering code generation models to identify a sequence of edits that will results in more correct code while remaining close to the original student's style. Together, our results suggest that many properties of code are properties of individual students and that training on edit traces can lead to models that are more steerable, more predictive of student behavior while programming, and better at generating programs in their final states. Code and data is available at https://github.com/meghabyte/pencilcode-public

---

### HybridFlow: Quantification of Aleatoric and Epistemic Uncertainty with a Single Hybrid Model

Authors: Peter Van Katwyk, Karianne J. Bergen
Published: 2025-10-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.05054v2

Uncertainty quantification is critical for ensuring robustness in high-stakes machine learning applications. We introduce HybridFlow, a modular hybrid architecture that unifies the modeling of aleatoric and epistemic uncertainty by combining a Conditional Masked Autoregressive normalizing flow for estimating aleatoric uncertainty with a flexible probabilistic predictor for epistemic uncertainty. The framework supports integration with any probabilistic model class, allowing users to easily adapt HybridFlow to existing architectures without sacrificing predictive performance. HybridFlow improves upon previous uncertainty quantification frameworks across a range of regression tasks, such as depth estimation, a collection of regression benchmarks, and a scientific case study of ice sheet emulation. We also provide empirical results of the quantified uncertainty, showing that the uncertainty quantified by HybridFlow is calibrated and better aligns with model error than existing methods for quantifying aleatoric and epistemic uncertainty. HybridFlow addresses a key challenge in Bayesian deep learning, unifying aleatoric and epistemic uncertainty modeling in a single robust framework.

---

### No-reference Quality Assessment of Contrast-distorted Images using Contrast-enhanced Pseudo Reference

Authors: Mohammad-Ali Mahmoudpour, Saeed Mahmoudpour
Published: 2025-10-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.05053v1

Contrast change is an important factor that affects the quality of images. During image capturing, unfavorable lighting conditions can cause contrast change and visual quality loss. While various methods have been proposed to assess the quality of images under different distortions such as blur and noise, contrast distortion has been largely overlooked as its visual impact and properties are different from other conventional types of distortions. In this paper, we propose a no-reference image quality assessment (NR-IQA) metric for contrast-distorted images. Using a set of contrast enhancement algorithms, we aim to generate pseudo-reference images that are visually close to the actual reference image, such that the NR problem is transformed to a Full-reference (FR) assessment with higher accuracy. To this end, a large dataset of contrast-enhanced images is produced to train a classification network that can select the most suitable contrast enhancement algorithm based on image content and distortion for pseudo-reference image generation. Finally, the evaluation is performed in the FR manner to assess the quality difference between the contrast-enhanced (pseudoreference) and degraded images. Performance evaluation of the proposed method on three databases containing contrast distortions (CCID2014, TID2013, and CSIQ), indicates the promising performance of the proposed method.

---

### Proactive defense against LLM Jailbreak

Authors: Weiliang Zhao, Jinjun Peng, Daniel Ben-Levi, Zhou Yu, Junfeng Yang
Published: 2025-10-06
Categories: cs.CR, cs.CL
arXiv: https://arxiv.org/abs/2510.05052v2

The proliferation of powerful large language models (LLMs) has necessitated robust safety alignment, yet these models remain vulnerable to evolving adversarial attacks, including multi-turn jailbreaks that iteratively search for successful queries. Current defenses, which are primarily reactive and static, often fail to handle these iterative attacks. In this paper, we introduce ProAct, a novel proactive defense framework designed to disrupt and mislead these iterative search jailbreak methods. Our core idea is to intentionally mislead these jailbreak methods into thinking that the model has been jailbroken with "spurious responses". These misleading responses provide false signals to the attacker's internal optimization loop, causing the adversarial search to terminate prematurely and effectively jailbreaking the jailbreak. By conducting extensive experiments across state-of-the-art LLMs, jailbreaking frameworks, and safety benchmarks, we demonstrate that our method consistently and significantly reduces attack success rates by up to 94% without affecting utility. When combined with other defense fraeworks, it further reduces the latest attack strategies' success rate to 0%. ProActrepresents an orthogonal defense strategy that serves as an additional guardrail to enhance LLM safety against the most effective jailbreaking attacks.

---

### SegMASt3R: Geometry Grounded Segment Matching

Authors: Rohit Jayanti, Swayam Agrawal, Vansh Garg, Siddharth Tourani, Muhammad Haris Khan, Sourav Garg, Madhava Krishna
Published: 2025-10-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.05051v2

Segment matching is an important intermediate task in computer vision that establishes correspondences between semantically or geometrically coherent regions across images. Unlike keypoint matching, which focuses on localized features, segment matching captures structured regions, offering greater robustness to occlusions, lighting variations, and viewpoint changes. In this paper, we leverage the spatial understanding of 3D foundation models to tackle wide-baseline segment matching, a challenging setting involving extreme viewpoint shifts. We propose an architecture that uses the inductive bias of these 3D foundation models to match segments across image pairs with up to 180 degree view-point change rotation. Extensive experiments show that our approach outperforms state-of-the-art methods, including the SAM2 video propagator and local feature matching methods, by up to 30% on the AUPRC metric, on ScanNet++ and Replica datasets. We further demonstrate benefits of the proposed model on relevant downstream tasks, including 3D instance mapping and object-relative navigation. Project Page: https://segmast3r.github.io/

---

### KEEP: Integrating Medical Ontologies with Clinical Data for Robust Code Embeddings

Authors: Ahmed Elhussein, Paul Meddeb, Abigail Newbury, Jeanne Mirone, Martin Stoll, Gamze Gursoy
Published: 2025-10-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.05049v1

Machine learning in healthcare requires effective representation of structured medical codes, but current methods face a trade off: knowledge graph based approaches capture formal relationships but miss real world patterns, while data driven methods learn empirical associations but often overlook structured knowledge in medical terminologies. We present KEEP (Knowledge preserving and Empirically refined Embedding Process), an efficient framework that bridges this gap by combining knowledge graph embeddings with adaptive learning from clinical data. KEEP first generates embeddings from knowledge graphs, then employs regularized training on patient records to adaptively integrate empirical patterns while preserving ontological relationships. Importantly, KEEP produces final embeddings without task specific auxiliary or end to end training enabling KEEP to support multiple downstream applications and model architectures. Evaluations on structured EHR from UK Biobank and MIMIC IV demonstrate that KEEP outperforms both traditional and Language Model based approaches in capturing semantic relationships and predicting clinical outcomes. Moreover, KEEP's minimal computational requirements make it particularly suitable for resource constrained environments.

---

### Look-ahead Reasoning with a Learned Model in Imperfect Information Games

Authors: Ondřej Kubíček, Viliam Lisý
Published: 2025-10-06
Categories: cs.AI, cs.GT
arXiv: https://arxiv.org/abs/2510.05048v1

Test-time reasoning significantly enhances pre-trained AI agents' performance. However, it requires an explicit environment model, often unavailable or overly complex in real-world scenarios. While MuZero enables effective model learning for search in perfect information games, extending this paradigm to imperfect information games presents substantial challenges due to more nuanced look-ahead reasoning techniques and large number of states relevant for individual decisions. This paper introduces an algorithm LAMIR that learns an abstracted model of an imperfect information game directly from the agent-environment interaction. During test time, this trained model is used to perform look-ahead reasoning. The learned abstraction limits the size of each subgame to a manageable size, making theoretically principled look-ahead reasoning tractable even in games where previous methods could not scale. We empirically demonstrate that with sufficient capacity, LAMIR learns the exact underlying game structure, and with limited capacity, it still learns a valuable abstraction, which improves game playing performance of the pre-trained agents even in large games.

---

### A Unified Optimization Framework for Multiclass Classification with Structured Hyperplane Arrangements

Authors: Víctor Blanco, Harshit Kothari, James Luedtke
Published: 2025-10-06
Categories: math.OC, cs.LG
arXiv: https://arxiv.org/abs/2510.05047v1

In this paper, we propose a new mathematical optimization model for multiclass classification based on arrangements of hyperplanes. Our approach preserves the core support vector machine (SVM) paradigm of maximizing class separation while minimizing misclassification errors, and it is computationally more efficient than a previous formulation. We present a kernel-based extension that allows it to construct nonlinear decision boundaries. Furthermore, we show how the framework can naturally incorporate alternative geometric structures, including classification trees, $\ell_p$-SVMs, and models with discrete feature selection. To address large-scale instances, we develop a dynamic clustering matheuristic that leverages the proposed MIP formulation. Extensive computational experiments demonstrate the efficiency of the proposed model and dynamic clustering heuristic, and we report competitive classification performance on both synthetic datasets and real-world benchmarks from the UCI Machine Learning Repository, comparing our method with state-of-the-art implementations available in scikit-learn.

---

### COLE: a Comprehensive Benchmark for French Language Understanding Evaluation

Authors: David Beauchemin, Yan Tremblay, Mohamed Amine Youssef, Richard Khoury
Published: 2025-10-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05046v2

To address the need for a more comprehensive evaluation of French Natural Language Understanding (NLU), we introduce COLE, a new benchmark composed of 23 diverse task covering a broad range of NLU capabilities, including sentiment analysis, paraphrase detection, grammatical judgment, and reasoning, with a particular focus on linguistic phenomena relevant to the French language. We benchmark 94 large language models (LLM), providing an extensive analysis of the current state of French NLU. Our results highlight a significant performance gap between closed- and open-weights models and identify key challenging frontiers for current LLMs, such as zero-shot extractive question-answering (QA), fine-grained word sense disambiguation, and understanding of regional language variations. We release COLE as a public resource to foster further progress in French language modelling.

---

### Test-Time Scaling in Diffusion LLMs via Hidden Semi-Autoregressive Experts

Authors: Jihoon Lee, Hoyeon Moon, Kevin Zhai, Arun Kumar Chithanar, Anit Kumar Sahu, Soummya Kar, Chul Lee, Souradip Chakraborty, Amrit Singh Bedi
Published: 2025-10-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.05040v1

Diffusion-based large language models (dLLMs) are trained flexibly to model extreme dependence in the data distribution; however, how to best utilize this information at inference time remains an open problem. In this work, we uncover an interesting property of these models: dLLMs trained on textual data implicitly learn a mixture of semi-autoregressive experts, where different generation orders reveal different specialized behaviors. We show that committing to any single, fixed inference time schedule, a common practice, collapses performance by failing to leverage this latent ensemble. To address this, we introduce HEX (Hidden semiautoregressive EXperts for test-time scaling), a training-free inference method that ensembles across heterogeneous block schedules. By doing a majority vote over diverse block-sized generation paths, HEX robustly avoids failure modes associated with any single fixed schedule. On reasoning benchmarks such as GSM8K, it boosts accuracy by up to 3.56X (from 24.72% to 88.10%), outperforming top-K margin inference and specialized fine-tuned methods like GRPO, without additional training. HEX even yields significant gains on MATH benchmark from 16.40% to 40.00%, scientific reasoning on ARC-C from 54.18% to 87.80%, and TruthfulQA from 28.36% to 57.46%. Our results establish a new paradigm for test-time scaling in diffusion-based LLMs (dLLMs), revealing that the sequence in which masking is performed plays a critical role in determining performance during inference.

---

### Guided Query Refinement: Multimodal Hybrid Retrieval with Test-Time Optimization

Authors: Omri Uzan, Asaf Yehudai, Roi pony, Eyal Shnarch, Ariel Gera
Published: 2025-10-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05038v3

Multimodal encoders have pushed the boundaries of visual document retrieval, matching textual query tokens directly to image patches and achieving state-of-the-art performance on public benchmarks. Recent models relying on this paradigm have massively scaled the sizes of their query and document representations, presenting obstacles to deployment and scalability in real-world pipelines. Furthermore, purely vision-centric approaches may be constrained by the inherent modality gap still exhibited by modern vision-language models. In this work, we connect these challenges to the paradigm of hybrid retrieval, investigating whether a lightweight dense text retriever can enhance a stronger vision-centric model. Existing hybrid methods, which rely on coarse-grained fusion of ranks or scores, fail to exploit the rich interactions within each model's representation space. To address this, we introduce Guided Query Refinement (GQR), a novel test-time optimization method that refines a primary retriever's query embedding using guidance from a complementary retriever's scores. Through extensive experiments on visual document retrieval benchmarks, we demonstrate that GQR allows vision-centric models to match the performance of models with significantly larger representations, while being up to 14x faster and requiring 54x less memory. Our findings show that GQR effectively pushes the Pareto frontier for performance and efficiency in multimodal retrieval. We release our code at https://github.com/IBM/test-time-hybrid-retrieval

---

### Graph-Aware Diffusion for Signal Generation

Authors: Sergio Rozada, Vimal K. B., Andrea Cavallo, Antonio G. Marques, Hadi Jamali-Rad, Elvin Isufi
Published: 2025-10-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.05036v1

We study the problem of generating graph signals from unknown distributions defined over given graphs, relevant to domains such as recommender systems or sensor networks. Our approach builds on generative diffusion models, which are well established in vision and graph generation but remain underexplored for graph signals. Existing methods lack generality, either ignoring the graph structure in the forward process or designing graph-aware mechanisms tailored to specific domains. We adopt a forward process that incorporates the graph through the heat equation. Rather than relying on the standard formulation, we consider a time-warped coefficient to mitigate the exponential decay of the drift term, yielding a graph-aware generative diffusion model (GAD). We analyze its forward dynamics, proving convergence to a Gaussian Markov random field with covariance parametrized by the graph Laplacian, and interpret the backward dynamics as a sequence of graph-signal denoising problems. Finally, we demonstrate the advantages of GAD on synthetic data, real traffic speed measurements, and a temperature sensor network.

---

### Video-LMM Post-Training: A Deep Dive into Video Reasoning with Large Multimodal Models

Authors: Yolo Y. Tang, Jing Bi, Pinxin Liu, Zhenyu Pan, Zhangyun Tan, Qianxiang Shen, Jiani Liu, Hang Hua, Junjia Guo, Yunzhong Xiao, Chao Huang, Zhiyuan Wang et al.
Published: 2025-10-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.05034v6

Video understanding represents the most challenging frontier in computer vision, requiring models to reason about complex spatiotemporal relationships, long-term dependencies, and multimodal evidence. The recent emergence of Video-Large Multimodal Models (Video-LMMs), which integrate visual encoders with powerful decoder-based language models, has demonstrated remarkable capabilities in video understanding tasks. However, the critical phase that transforms these models from basic perception systems into sophisticated reasoning engines, post-training, remains fragmented across the literature. This survey provides the first comprehensive examination of post-training methodologies for Video-LMMs, encompassing three fundamental pillars: supervised fine-tuning (SFT) with chain-of-thought, reinforcement learning (RL) from verifiable objectives, and test-time scaling (TTS) through enhanced inference computation. We present a structured taxonomy that clarifies the roles, interconnections, and video-specific adaptations of these techniques, addressing unique challenges such as temporal localization, spatiotemporal grounding, long video efficiency, and multimodal evidence integration. Through systematic analysis of representative methods, we synthesize key design principles, insights, and evaluation protocols while identifying critical open challenges in reward design, scalability, and cost-performance optimization. We further curate essential benchmarks, datasets, and metrics to facilitate rigorous assessment of post-training effectiveness. This survey aims to provide researchers and practitioners with a unified framework for advancing Video-LMM capabilities. Additional resources and updates are maintained at: https://github.com/yunlong10/Awesome-Video-LMM-Post-Training

---

### Causal Abstractions, Categorically Unified

Authors: Markus Englberger, Devendra Singh Dhami
Published: 2025-10-06
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2510.05033v1

We present a categorical framework for relating causal models that represent the same system at different levels of abstraction. We define a causal abstraction as natural transformations between appropriate Markov functors, which concisely consolidate desirable properties a causal abstraction should exhibit. Our approach unifies and generalizes previously considered causal abstractions, and we obtain categorical proofs and generalizations of existing results on causal abstractions. Using string diagrammatical tools, we can explicitly describe the graphs that serve as consistent abstractions of a low-level graph under interventions. We discuss how methods from mechanistic interpretability, such as circuit analysis and sparse autoencoders, fit within our categorical framework. We also show how applying do-calculus on a high-level graphical abstraction of an acyclic-directed mixed graph (ADMG), when unobserved confounders are present, gives valid results on the low-level graph, thus generalizing an earlier statement by Anand et al. (2023). We argue that our framework is more suitable for modeling causal abstractions compared to existing categorical frameworks. Finally, we discuss how notions such as $τ$-consistency and constructive $τ$-abstractions can be recovered with our framework.

---

### Exploration-Exploitation-Evaluation (EEE): A Framework for Metaheuristic Algorithms in Combinatorial Optimization

Authors: Ethan Davis
Published: 2025-10-06
Categories: cs.NE
arXiv: https://arxiv.org/abs/2510.05027v1

We introduce a framework for applying metaheuristic algorithms, such as ant colony optimization (ACO), to combinatorial optimization problems (COPs) like the traveling salesman problem (TSP). The framework consists of three sequential stages: broad exploration of the parameter space, exploitation of top-performing parameters, and uncertainty quantification (UQ) to assess the reliability of results. As a case study, we apply ACO to the TSPLIB berlin52 dataset, which has a known optimal tour length of 7542. Using our framework, we calculate that the probability of ACO finding the global optimum is approximately 1/40 in a single run and improves to 1/5 when aggregated over ten runs.

---

### Idiom Understanding as a Tool to Measure the Dialect Gap

Authors: David Beauchemin, Yan Tremblay, Mohamed Amine Youssef, Richard Khoury
Published: 2025-10-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05026v4

The tasks of idiom understanding and dialect understanding are both well-established benchmarks in natural language processing. In this paper, we propose combining them, and using regional idioms as a test of dialect understanding. Towards this end, we propose three new benchmark datasets for the Quebec dialect of French: QFrCoRE, which contains 4,633 instances of idiomatic phrases, and QFrCoRT, which comprises 171 regional instances of idiomatic words, and a new benchmark for French Metropolitan expressions, MFrCoE, which comprises 4,938 phrases. We explain how to construct these corpora, so that our methodology can be replicated for other dialects. Our experiments with 111 LLMs reveal a critical disparity in dialectal competence: while models perform well on French Metropolitan, 65.77% of them perform significantly worse on Quebec idioms, with only 9.0% favoring the regional dialect. These results confirm that our benchmarks are a reliable tool for quantifying the dialect gap and that prestige-language proficiency does not guarantee regional dialect understanding.

---

### Imperceptible Jailbreaking against Large Language Models

Authors: Kuofeng Gao, Yiming Li, Chao Du, Xin Wang, Xingjun Ma, Shu-Tao Xia, Tianyu Pang
Published: 2025-10-06
Categories: cs.CL, cs.AI, cs.CR
arXiv: https://arxiv.org/abs/2510.05025v1

Jailbreaking attacks on the vision modality typically rely on imperceptible adversarial perturbations, whereas attacks on the textual modality are generally assumed to require visible modifications (e.g., non-semantic suffixes). In this paper, we introduce imperceptible jailbreaks that exploit a class of Unicode characters called variation selectors. By appending invisible variation selectors to malicious questions, the jailbreak prompts appear visually identical to original malicious questions on screen, while their tokenization is "secretly" altered. We propose a chain-of-search pipeline to generate such adversarial suffixes to induce harmful responses. Our experiments show that our imperceptible jailbreaks achieve high attack success rates against four aligned LLMs and generalize to prompt injection attacks, all without producing any visible modifications in the written prompt. Our code is available at https://github.com/sail-sg/imperceptible-jailbreaks.

---

### Inoculation Prompting: Instructing LLMs to misbehave at train-time improves test-time alignment

Authors: Nevan Wichers, Aram Ebtekar, Ariana Azarbal, Victor Gillioz, Christine Ye, Emil Ryd, Neil Rathi, Henry Sleight, Alex Mallen, Fabien Roger, Samuel Marks
Published: 2025-10-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.05024v3

Large language models are sometimes trained with imperfect oversight signals, leading to undesired behaviors such as reward hacking and sycophancy. Improving oversight quality can be expensive or infeasible, motivating methods that improve learned behavior despite an imperfect training signal. We introduce Inoculation Prompting (IP), a simple but counterintuitive technique that prevents learning of an undesired behavior by modifying training prompts to explicitly request it. For example, to inoculate against reward hacking, we modify the prompts used in supervised fine-tuning to request code that only works on provided test cases but fails on other inputs. Across four settings we find that IP reduces the learning of undesired behavior without substantially reducing the learning of desired capabilities. We also show that prompts which more strongly elicit the undesired behavior prior to fine-tuning more effectively inoculate against the behavior when used during training; this serves as a heuristic to identify promising inoculation prompts. Overall, IP is a simple yet effective way to control how models generalize from fine-tuning, preventing learning of undesired behaviors without substantially disrupting desired capabilities.

---

### Rethinking Langevin Thompson Sampling from A Stochastic Approximation Perspective

Authors: Weixin Wang, Haoyang Zheng, Guang Lin, Wei Deng, Pan Xu
Published: 2025-10-06
Categories: cs.LG, cs.AI, stat.ML
arXiv: https://arxiv.org/abs/2510.05023v2

Langevin-based Thompson Sampling (TS) methods for multi-armed bandits use likelihood gradients to approximate posterior samples when exact sampling is intractable. As new rewards arrive, the posterior changes, and these methods run an inner Markov chain to approximate the updated target. To simplify this repeated sampling procedure, we introduce TS-SA, which incorporates stochastic approximation (SA) within the TS framework. In each round, TS-SA takes one Langevin step using the newly observed reward and applies an SA update to combine the noisy proposal with the current parameter estimate. By retaining this estimate across rounds, TS-SA maintains one continuing Markov chain for each arm, preserving the influence of past observations without storing or sampling them again. Within this recursion, the proposal step size remains fixed across rounds, while the SA weights decrease as observations accumulate to control the effect of noise on the estimate. This continuing-chain formulation simplifies the theoretical analysis, allowing us to establish a near-optimal worst-case regret bound of $\widetilde{\mathcal O}(\sqrt{KT})$ under log-concavity and smoothness assumptions with one update per new reward. Our experiments further demonstrate that TS-SA, using a single-step Langevin update after an initial warm-up, achieves strong performance against standard baselines.

---

### Large Language Models Achieve Gold Medal Performance at the International Olympiad on Astronomy & Astrophysics (IOAA)

Authors: Lucas Carrit Delgado Pinheiro, Ziru Chen, Bruno Caixeta Piazza, Ness Shroff, Yingbin Liang, Yuan-Sen Ting, Huan Sun
Published: 2025-10-06
Categories: astro-ph.IM, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2510.05016v2

While task-specific demonstrations show early success in applying large language models (LLMs) to automate some astronomical research tasks, they only provide incomplete views of all necessary capabilities in solving astronomy problems, calling for more thorough understanding of LLMs' strengths and limitations. So far, existing benchmarks and evaluations focus on simple question-answering that primarily tests astronomical knowledge and fails to evaluate the complex reasoning required for real-world research in the discipline. Here, we address this gap by systematically benchmarking five state-of-the-art LLMs on the International Olympiad on Astronomy and Astrophysics (IOAA) exams, which are designed to examine deep conceptual understanding, multi-step derivations, and multimodal analysis. With average scores of 85.6% and 84.2%, Gemini 2.5 Pro and GPT-5 (the two top-performing models) not only achieve gold medal level performance but also rank in the top two among ~200-300 participants in all four IOAA theory exams evaluated (2022-2025). In comparison, results on the data analysis exams show more divergence. GPT-5 still excels in the exams with an 88.5% average score, ranking top 10 among the participants in the four most recent IOAAs, while other models' performances drop to 48-76%. Furthermore, our in-depth error analysis underscores conceptual reasoning, geometric reasoning, and spatial visualization (52-79% accuracy) as consistent weaknesses among all LLMs. Hence, although LLMs approach peak human performance in theory exams, critical gaps must be addressed before they can serve as autonomous research agents in astronomy.

---

### Exploring the Efficacy of Modified Transfer Learning in Identifying Parkinson's Disease Through Drawn Image Patterns

Authors: Nabil Daiyan, Md Rakibul Haque
Published: 2025-10-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.05015v1

Parkinson's disease (PD) is a progressive neurodegenerative condition characterized by the death of dopaminergic neurons, leading to various movement disorder symptoms. Early diagnosis of PD is crucial to prevent adverse effects, yet traditional diagnostic methods are often cumbersome and costly. In this study, a machine learning-based approach is proposed using hand-drawn spiral and wave images as potential biomarkers for PD detection. Our methodology leverages convolutional neural networks (CNNs), transfer learning, and attention mechanisms to improve model performance and resilience against overfitting. To enhance the diversity and richness of both spiral and wave categories, the training dataset undergoes augmentation to increase the number of images. The proposed architecture comprises three phases: utilizing pre-trained CNNs, incorporating custom convolutional layers, and ensemble voting. Employing hard voting further enhances performance by aggregating predictions from multiple models. Experimental results show promising accuracy rates. For spiral images, weighted average precision, recall, and F1-score are 90%, and for wave images, they are 96.67%. After combining the predictions through ensemble hard voting, the overall accuracy is 93.3%. These findings underscore the potential of machine learning in early PD diagnosis, offering a non-invasive and cost-effective solution to improve patient outcomes.

---

### Think Then Embed: Generative Context Improves Multimodal Embedding

Authors: Xuanming Cui, Jianpeng Cheng, Hong-you Chen, Satya Narayan Shukla, Abhijeet Awasthi, Xichen Pan, Chaitanya Ahuja, Shlok Kumar Mishra, Yonghuan Yang, Jun Xiao, Qi Guo, Ser-Nam Lim et al.
Published: 2025-10-06
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.05014v4

There is a growing interest in Universal Multimodal Embeddings (UME), where models are required to generate task-specific representations. While recent studies show that Multimodal Large Language Models (MLLMs) perform well on such tasks, they treat MLLMs solely as encoders, overlooking their generative capacity. However, such an encoding paradigm becomes less effective as instructions become more complex and require compositional reasoning. Inspired by the proven effectiveness of chain-of-thought reasoning, we propose a general Think-Then-Embed (TTE) framework for UME, composed of a reasoner and an embedder. The reasoner MLLM first generates reasoning traces that explain complex queries, followed by an embedder that produces representations conditioned on both the original query and the intermediate reasoning. This explicit reasoning step enables more nuanced understanding of complex multimodal instructions. Our contributions are threefold. First, by leveraging a powerful MLLM reasoner, we achieve state-of-the-art performance on the MMEB-V2 benchmark, surpassing proprietary models trained on massive in-house datasets. Second, to reduce the dependency on large MLLM reasoners, we finetune a smaller MLLM reasoner using high-quality embedding-centric reasoning traces, achieving the best performance among open-source models with a 7% absolute gain over recently proposed models. Third, we investigate strategies for integrating the reasoner and embedder into a unified model for improved efficiency without sacrificing performance.

---

### Curiosity-Driven Development of Action and Language in Robots Through Self-Exploration

Authors: Theodore Jerome Tinker, Kenji Doya, Jun Tani
Published: 2025-10-06
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2510.05013v7

Infants acquire language with generalization from minimal experience, whereas large language models require billions of training tokens. What underlies efficient development in humans? We investigated this problem through experiments wherein robotic agents learn to perform actions associated with imperative sentences (e.g., push red cube) via curiosity-driven self-exploration. Our approach amortizes active inference using Q-learning, enabling intrinsically motivated developmental learning. The simulations reveal key findings corresponding to observations in developmental psychology. i) Generalization improves drastically as the scale of compositional elements increases. ii) Curiosity-driven exploration enables faster learning. iii) Rote pairing of sentences and actions precedes compositional generalization. iv) Exception-handling induces U-shaped developmental performance, a pattern like representational redescription in child language learning. These results suggest that curiosity-driven active inference accounts for how intrinsically motivated sensorimotor-linguistic learning supports scalable compositional generalization and exception handling in humans and artificial agents.

---

### Latent Uncertainty Representations for Video-based Driver Action and Intention Recognition

Authors: Koen Vellenga, H. Joe Steinhauer, Jonas Andersson, Anders Sjögren
Published: 2025-10-06
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.05006v1

Deep neural networks (DNNs) are increasingly applied to safety-critical tasks in resource-constrained environments, such as video-based driver action and intention recognition. While last layer probabilistic deep learning (LL-PDL) methods can detect out-of-distribution (OOD) instances, their performance varies. As an alternative to last layer approaches, we propose extending pre-trained DNNs with transformation layers to produce multiple latent representations to estimate the uncertainty. We evaluate our latent uncertainty representation (LUR) and repulsively trained LUR (RLUR) approaches against eight PDL methods across four video-based driver action and intention recognition datasets, comparing classification performance, calibration, and uncertainty-based OOD detection. We also contribute 28,000 frame-level action labels and 1,194 video-level intention labels for the NuScenes dataset. Our results show that LUR and RLUR achieve comparable in-distribution classification performance to other LL-PDL approaches. For uncertainty-based OOD detection, LUR matches top-performing PDL methods while being more efficient to train and easier to tune than approaches that require Markov-Chain Monte Carlo sampling or repulsive training procedures.

---

### Resource-Efficient Fine-Tuning of LLaMA-3.2-3B for Medical Chain-of-Thought Reasoning

Authors: Imran Mansha
Published: 2025-10-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.05003v1

Large Language Models (LLMs) such as GPT-4 and LLaMA have demonstrated remarkable reasoning abilities but require significant computational resources for fine-tuning. This paper presents a resource-efficient fine-tuning approach for LLaMA-3.2-3B to enhance medical chain-of-thought reasoning while operating under constrained GPU and memory settings. Using parameter-efficient tuning techniques such as LoRA and QLoRA, we adapt the base model on publicly available medical reasoning datasets. The model achieves improved reasoning coherence and factual accuracy while reducing memory usage by up to 60% compared to standard full fine-tuning. Experimental evaluation demonstrates that lightweight adaptations can retain strong reasoning capability in medical question-answering tasks. This work highlights practical strategies for deploying LLMs in low-resource research environments and provides insights into balancing efficiency and domain specialization for medical AI systems.

---

### Bridging Text and Video Generation: A Survey

Authors: Nilay Kumar, Priyansh Bhandari, G. Maragatham
Published: 2025-10-06
Categories: cs.GR, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2510.04999v1

Text-to-video (T2V) generation technology holds potential to transform multiple domains such as education, marketing, entertainment, and assistive technologies for individuals with visual or reading comprehension challenges, by creating coherent visual content from natural language prompts. From its inception, the field has advanced from adversarial models to diffusion-based models, yielding higher-fidelity, temporally consistent outputs. Yet challenges persist, such as alignment, long-range coherence, and computational efficiency. Addressing this evolving landscape, we present a comprehensive survey of text-to-video generative models, tracing their development from early GANs and VAEs to hybrid Diffusion-Transformer (DiT) architectures, detailing how these models work, what limitations they addressed in their predecessors, and why shifts toward new architectural paradigms were necessary to overcome challenges in quality, coherence, and control. We provide a systematic account of the datasets, which the surveyed text-to-video models were trained and evaluated on, and, to support reproducibility and assess the accessibility of training such models, we detail their training configurations, including their hardware specifications, GPU counts, batch sizes, learning rates, optimizers, epochs, and other key hyperparameters. Further, we outline the evaluation metrics commonly used for evaluating such models and present their performance across standard benchmarks, while also discussing the limitations of these metrics and the emerging shift toward more holistic, perception-aligned evaluation strategies. Finally, drawing from our analysis, we outline the current open challenges and propose a few promising future directions, laying out a perspective for future researchers to explore and build upon in advancing T2V research and applications.

---

### AutoEmpirical: LLM-Based Automated Research for Empirical Software Fault Analysis

Authors: Jiongchi Yu, Weipeng Jiang, Xiaoyu Zhang, Qiang Hu, Xiaofei Xie, Chao Shen
Published: 2025-10-06
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2510.04997v1

Understanding software faults is essential for empirical research in software development and maintenance. However, traditional fault analysis, while valuable, typically involves multiple expert-driven steps such as collecting potential faults, filtering, and manual investigation. These processes are both labor-intensive and time-consuming, creating bottlenecks that hinder large-scale fault studies in complex yet critical software systems and slow the pace of iterative empirical research. In this paper, we decompose the process of empirical software fault study into three key phases: (1) research objective definition, (2) data preparation, and (3) fault analysis, and we conduct an initial exploration study of applying Large Language Models (LLMs) for fault analysis of open-source software. Specifically, we perform the evaluation on 3,829 software faults drawn from a high-quality empirical study. Our results show that LLMs can substantially improve efficiency in fault analysis, with an average processing time of about two hours, compared to the weeks of manual effort typically required. We conclude by outlining a detailed research plan that highlights both the potential of LLMs for advancing empirical fault studies and the open challenges that required be addressed to achieve fully automated, end-to-end software fault analysis.

---

### Reinforce-Ada: An Adaptive Sampling Framework under Non-linear RL Objectives

Authors: Wei Xiong, Chenlu Ye, Baohao Liao, Hanze Dong, Xinxing Xu, Christof Monz, Jiang Bian, Nan Jiang, Tong Zhang
Published: 2025-10-06
Categories: cs.LG, cs.AI, cs.CL, stat.ML
arXiv: https://arxiv.org/abs/2510.04996v3

Reinforcement learning (RL) for large language model reasoning is frequently hindered by signal loss, a phenomenon where standard uniform sampling with small group sizes fails to uncover informative learning signals for difficult prompts. We demonstrate that this collapse is a statistical artifact of undersampling rather than an inherent model limitation. To address this systematically, we introduce a theoretical framework based on optimizing a non-linear RL objective (e.g., log-likelihood). We show that this objective naturally induces a weighted gradient estimator that prioritizes difficult prompts, which can be robustly realized through adaptive sampling. Guided by this framework, we propose Reinforce-Ada, a family of algorithms that dynamically allocates inference budgets based on prompt difficulty, effectively scaling up RL compute to where it is needed most. Unlike passive filtering methods that discard low-signal prompts, Reinforce-Ada actively invests compute to recover them. We introduce two efficient realizations: an estimation-based approach and a model-free sequential sampling approach. Extensive experiments across multiple benchmarks show that Reinforce-Ada significantly outperforms uniform baselines like GRPO, recovering lost signals and accelerating convergence by up to $2\times$ while maintaining the same total inference budget. Code is available at https://github.com/RLHFlow/Reinforce-Ada.

---

### Power Transform Revisited: Numerically Stable, and Federated

Authors: Xuefeng Xu, Graham Cormode
Published: 2025-10-06
Categories: cs.LG, math.NA
arXiv: https://arxiv.org/abs/2510.04995v3

Power transforms are popular parametric methods for making data more Gaussian-like, and are widely used as preprocessing steps in statistical analysis and machine learning. However, we find that direct implementations of power transforms suffer from severe numerical instabilities, which can lead to incorrect results or even crashes. In this paper, we provide a comprehensive analysis of the sources of these instabilities and propose effective remedies. We further extend power transforms to the federated learning setting, addressing both numerical and distributional challenges that arise in this context. Experiments on real-world datasets demonstrate that our methods are both effective and robust, substantially improving stability compared to existing approaches.

---

### Adaptive Memory Momentum via a Model-Based Framework for Deep Learning Optimization

Authors: Kristi Topollai, Anna Choromanska
Published: 2025-10-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.04988v3

The vast majority of modern deep learning models are trained with momentum-based first-order optimizers. The momentum term governs the optimizer's memory by determining how much each past gradient contributes to the current convergence direction. Fundamental momentum methods, such as Nesterov Accelerated Gradient and the Heavy Ball method, as well as more recent optimizers such as AdamW and Lion, all rely on the momentum coefficient that is customarily set to $β= 0.9$ and kept constant during model training, a strategy widely used by practitioners, yet suboptimal. In this paper, we introduce an \textit{adaptive memory} mechanism that replaces constant momentum with a dynamic momentum coefficient that is adjusted online during optimization. We derive our method by approximating the objective function using two planes: one derived from the gradient at the current iterate and the other obtained from the accumulated memory of the past gradients. To the best of our knowledge, such a proximal framework was never used for momentum-based optimization. Our proposed approach is novel, extremely simple to use, and does not rely on extra assumptions or hyperparameter tuning. We implement adaptive memory variants of both SGD and AdamW across a wide range of learning tasks, from simple convex problems to large-scale deep learning scenarios, demonstrating that our approach can outperform standard SGD and Adam with hand-tuned momentum coefficients. Finally, our work opens doors for new ways of inducing adaptivity in optimization.

---

### What your brain activity says about you: A review of neuropsychiatric disorders identified in resting-state and sleep EEG data

Authors: J. E. M. Scanlon, A. Pelzer, M. Gharleghi, K. C. Fuhrmeister, T. Köllmer, P. Aichroth, R. Göder, C. Hansen, K. I. Wolf
Published: 2025-10-06
Categories: cs.NE, cs.CR, cs.CY, q-bio.NC
arXiv: https://arxiv.org/abs/2510.04984v2

Electroencephalogram monitoring devices and online data repositories hold large amounts of data from individuals participating in research and medical studies without direct reference to personal identifiers. This paper explores what types of personal and health information have been detected and classified within task-free EEG data. Additionally, we investigate key characteristics of the collected resting-state and sleep data, in order to determine the privacy risks involved with openly available EEG data. We used Google Scholar, Web of Science and searched relevant journals to find studies which classified or detected the presence of various disorders and personal information in resting state and sleep EEG. Only English full-text peer-reviewed journal articles or conference papers about classifying the presence of medical disorders between individuals were included. A quality analysis carried out by 3 reviewers determined general paper quality based on specified evaluation criteria. In resting state EEG, various disorders including Autism Spectrum Disorder, Parkinson's disease, and alcohol use disorder have been classified with high classification accuracy, often requiring only 5 mins of data or less. Sleep EEG tends to hold classifiable information about sleep disorders such as sleep apnea, insomnia, and REM sleep disorder, but usually involve longer recordings or data from multiple sleep stages. Many classification methods are still developing but even today, access to a person's EEG can reveal sensitive personal health information. With an increasing ability of machine learning methods to re-identify individuals from their EEG data, this review demonstrates the importance of anonymization, and the development of improved tools for keeping study participants and medical EEG users' privacy safe.

---

### AWARE, Beyond Sentence Boundaries: A Contextual Transformer Framework for Identifying Cultural Capital in STEM Narratives

Authors: Khalid Mehtab Khan, Anagha Kulkarni
Published: 2025-10-06
Categories: cs.CL, cs.AI, cs.CY, cs.LG
arXiv: https://arxiv.org/abs/2510.04983v3

Identifying cultural capital (CC) themes in student reflections can offer valuable insights that help foster equitable learning environments in classrooms. However, themes such as aspirational goals or family support are often woven into narratives, rather than appearing as direct keywords. This makes them difficult to detect for standard NLP models that process sentences in isolation. The core challenge stems from a lack of awareness, as standard models are pre-trained on general corpora, leaving them blind to the domain-specific language and narrative context inherent to the data. To address this, we introduce AWARE, a framework that systematically attempts to improve a transformer model's awareness for this nuanced task. AWARE has three core components: 1) Domain Awareness, adapting the model's vocabulary to the linguistic style of student reflections; 2) Context Awareness, generating sentence embeddings that are aware of the full essay context; and 3) Class Overlap Awareness, employing a multi-label strategy to recognize the coexistence of themes in a single sentence. Our results show that by making the model explicitly aware of the properties of the input, AWARE outperforms a strong baseline by 2.1 percentage points in Macro-F1 and shows considerable improvements across all themes. This work provides a robust and generalizable methodology for any text classification task in which meaning depends on the context of the narrative.

---

### LLM-Hanabi: Evaluating Multi-Agent Gameplays with Theory-of-Mind and Rationale Inference in Imperfect Information Collaboration Game

Authors: Fangzhou Liang, Tianshi Zheng, Chunkit Chan, Yauwai Yim, Yangqiu Song
Published: 2025-10-06
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2510.04980v1

Effective multi-agent collaboration requires agents to infer the rationale behind others' actions, a capability rooted in Theory-of-Mind (ToM). While recent Large Language Models (LLMs) excel at logical inference, their ability to infer rationale in dynamic, collaborative settings remains under-explored. This study introduces LLM-Hanabi, a novel benchmark that uses the cooperative game Hanabi to evaluate the rationale inference and ToM of LLMs. Our framework features an automated evaluation system that measures both game performance and ToM proficiency. Across a range of models, we find a significant positive correlation between ToM and in-game success. Notably, first-order ToM (interpreting others' intent) correlates more strongly with performance than second-order ToM (predicting others' interpretations). These findings highlight that for effective AI collaboration, the ability to accurately interpret a partner's rationale is more critical than higher-order reasoning. We conclude that prioritizing first-order ToM is a promising direction for enhancing the collaborative capabilities of future models.

---

### Federated Computation of ROC and PR Curves

Authors: Xuefeng Xu, Graham Cormode
Published: 2025-10-06
Categories: cs.LG, cs.CR
arXiv: https://arxiv.org/abs/2510.04979v2

Receiver Operating Characteristic (ROC) and Precision-Recall (PR) curves are fundamental analytical tools for evaluating classification models, providing a comprehensive view of performance across decision thresholds. In modern data management settings, such evaluation must often be performed over data that is distributed across multiple parties and subject to strict privacy constraints. This setting arises naturally in federated learning (FL) and collaborative data platforms, where raw prediction scores and labels cannot be centrally collected. We study the problem of computing ROC and PR curves over distributed data under formal privacy guarantees. The key challenge is that exact computation requires access to all prediction scores and labels, leading to linear communication costs and privacy risks. To address this, we propose a novel framework for approximating ROC and PR curves using quantiles of the score distribution, which can be computed efficiently under secure aggregation and distributed differential privacy. Our approach requires only $O(Q)$ communication, where $Q$ is the number of quantiles, and avoids sharing raw data entirely. We provide theoretical guarantees on the approximation quality by bounding the Area Error (AE) between the true and estimated curves, demonstrating principled trade-offs among approximation accuracy, privacy protection, and communication cost. Our method is robust to data heterogeneity and skew, making it suitable for real-world distributed data management scenarios. Experiments on real-world datasets show that our approach achieves high accuracy with low communication overhead under strong privacy constraints. Overall, this work provides a practical and theoretically grounded solution for privacy-preserving evaluation of classification models over distributed data.

---

### Aligning Perception, Reasoning, Modeling and Interaction: A Survey on Physical AI

Authors: Kun Xiang, Terry Jingchen Zhang, Yinya Huang, Jixi He, Zirong Liu, Yueling Tang, Ruizhe Zhou, Lijing Luo, Youpeng Wen, Xiuwei Chen, Bingqian Lin, Jianhua Han et al.
Published: 2025-10-06
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.04978v5

The rapid advancement of embodied intelligence and world models has intensified efforts to integrate physical laws into AI systems, yet physical perception and symbolic physics reasoning have developed along separate trajectories without a unified bridging framework. This work provides a comprehensive overview of physical AI, establishing clear distinctions between theoretical physics reasoning and applied physical understanding while systematically examining how physics-grounded methods enhance AI's real-world comprehension across structured symbolic reasoning, embodied systems, and generative models. Through rigorous analysis of recent advances, we advocate for intelligent systems that ground learning in both physical principles and embodied reasoning processes, transcending pattern recognition toward genuine understanding of physical laws. Our synthesis envisions next-generation world models capable of explaining physical phenomena and predicting future states, advancing safe, generalizable, and interpretable AI systems. We maintain a continuously updated resource at https://github.com/AI4Phys/Awesome-AI-for-Physics.

---

### StructuralDecompose: A Modular Framework for Robust Time Series Decomposition in R

Authors: Allen Daniel Sunny
Published: 2025-10-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.04974v1

We present StructuralDecompose, an R package for modular and interpretable time series decomposition. Unlike existing approaches that treat decomposition as a monolithic process, StructuralDecompose separates the analysis into distinct components: changepoint detection, anomaly detection, smoothing, and decomposition. This design provides flexibility and robust- ness, allowing users to tailor methods to specific time series characteristics. We demonstrate the package on simulated and real-world datasets, benchmark its performance against state-of-the- art tools such as Rbeast and autostsm, and discuss its role in interpretable machine learning workflows.

---

### Pivotal CLTs for Pseudolikelihood via Conditional Centering in Dependent Random Fields

Authors: Nabarun Deb
Published: 2025-10-06
Categories: math.ST, cs.LG, math.PR
arXiv: https://arxiv.org/abs/2510.04972v1

In this paper, we study fluctuations of conditionally centered statistics of the form $$N^{-1/2}\sum_{i=1}^N c_i(g(σ_i)-\mathbb{E}_N(g(σ_i)|σ_j,j\neq i))$$ where $(σ_1,\ldots ,σ_N)$ are sampled from a dependent random field, and $g$ is some bounded function. Our first main result shows that under weak smoothness assumptions on the conditional means (which cover both sparse and dense interactions), the above statistic converges to a Gaussian \emph{scale mixture} with a random scale determined by a \emph{quadratic variance} and an \emph{interaction component}. We also show that under appropriate studentization, the limit becomes a pivotal Gaussian. We leverage this theory to develop a general asymptotic framework for maximum pseudolikelihood (MPLE) inference in dependent random fields. We apply our results to Ising models with pairwise as well as higher-order interactions and exponential random graph models (ERGMs). In particular, we obtain a joint central limit theorem for the inverse temperature and magnetization parameters via the joint MPLE (to our knowledge, the first such result in dense, irregular regimes), and we derive conditionally centered edge CLTs and marginal MPLE CLTs for ERGMs without restricting to the ``sub-critical" region. Our proof is based on a method of moments approach via combinatorial decision-tree pruning, which may be of independent interest.

---

### Embracing Discrete Search: A Reasonable Approach to Causal Structure Learning

Authors: Marcel Wienöbst, Leonard Henckel, Sebastian Weichwald
Published: 2025-10-06
Categories: stat.ML, cs.AI, cs.LG, stat.ME
arXiv: https://arxiv.org/abs/2510.04970v2

We present FLOP (Fast Learning of Order and Parents), a score-based causal discovery algorithm for linear models. It pairs fast parent selection with iterative Cholesky-based score updates, cutting run-times over prior algorithms. This makes it feasible to fully embrace discrete search, enabling iterated local search with principled order initialization to find graphs with scores at or close to the global optimum. The resulting structures are highly accurate across benchmarks, with near-perfect recovery in standard settings. This performance calls for revisiting discrete search over graphs as a reasonable approach to causal discovery.

---

### ActiveMark: on watermarking of visual foundation models via massive activations

Authors: Anna Chistyakova, Mikhail Pautov
Published: 2025-10-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2510.04966v1

Being trained on large and vast datasets, visual foundation models (VFMs) can be fine-tuned for diverse downstream tasks, achieving remarkable performance and efficiency in various computer vision applications. The high computation cost of data collection and training motivates the owners of some VFMs to distribute them alongside the license to protect their intellectual property rights. However, a dishonest user of the protected model's copy may illegally redistribute it, for example, to make a profit. As a consequence, the development of reliable ownership verification tools is of great importance today, since such methods can be used to differentiate between a redistributed copy of the protected model and an independent model. In this paper, we propose an approach to ownership verification of visual foundation models by fine-tuning a small set of expressive layers of a VFM along with a small encoder-decoder network to embed digital watermarks into an internal representation of a hold-out set of input images. Importantly, the watermarks embedded remain detectable in the functional copies of the protected model, obtained, for example, by fine-tuning the VFM for a particular downstream task. Theoretically and experimentally, we demonstrate that the proposed method yields a low probability of false detection of a non-watermarked model and a low probability of false misdetection of a watermarked model.

---

### SSDD: Single-Step Diffusion Decoder for Efficient Image Tokenization

Authors: Théophane Vallaeys, Jakob Verbeek, Matthieu Cord
Published: 2025-10-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.04961v2

Tokenizers are a key component of state-of-the-art generative image models, extracting the most important features from the signal while reducing data dimension and redundancy. Most current tokenizers are based on KL-regularized variational autoencoders (KL-VAE), trained with reconstruction, perceptual and adversarial losses. Diffusion decoders have been proposed as a more principled alternative to model the distribution over images conditioned on the latent. However, matching the performance of KL-VAE still requires adversarial losses, as well as a higher decoding time due to iterative sampling. To address these limitations, we introduce a new pixel diffusion decoder architecture for improved scaling and training stability, benefiting from transformer components and GAN-free training. We use distillation to replicate the performance of the diffusion decoder in an efficient single-step decoder. This makes SSDD the first diffusion decoder optimized for single-step reconstruction trained without adversarial losses, reaching higher reconstruction quality and faster sampling than KL-VAE. In particular, SSDD improves reconstruction FID from $0.87$ to $0.46$ with $1.4\times$ higher throughput and preserve generation quality of DiTs with $3.8\times$ faster sampling. As such, SSDD can be used as a drop-in replacement for KL-VAE, and for building higher-quality and faster generative models.

---