# Research Papers — 2025-10 · bundle page 208 of 217

> Full-text markdown of papers #10351–10400 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/208
[← Previous bundle](https://aipod.dev/papers/archive/2025-10/page/207.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-10/page/209.md)

### Large-Scale Bayesian Causal Discovery with Interventional Data

Authors: Seong Woo Han, Daniel Duy Vo, Brielin C. Brown
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.01562v1

Inferring the causal relationships among a set of variables in the form of a directed acyclic graph (DAG) is an important but notoriously challenging problem. Recently, advancements in high-throughput genomic perturbation screens have inspired development of methods that leverage interventional data to improve model identification. However, existing methods still suffer poor performance on large-scale tasks and fail to quantify uncertainty. Here, we propose Interventional Bayesian Causal Discovery (IBCD), an empirical Bayesian framework for causal discovery with interventional data. Our approach models the likelihood of the matrix of total causal effects, which can be approximated by a matrix normal distribution, rather than the full data matrix. We place a spike-and-slab horseshoe prior on the edges and separately learn data-driven weights for scale-free and Erdős-Rényi structures from observational data, treating each edge as a latent variable to enable uncertainty-aware inference. Through extensive simulation, we show that IBCD achieves superior structure recovery compared to existing baselines. We apply IBCD to CRISPR perturbation (Perturb-seq) data on 521 genes, demonstrating that edge posterior inclusion probabilities enable identification of robust graph structures.

---

### AI Foundation Model for Time Series with Innovations Representation

Authors: Lang Tong, Xinyi Wang
Published: 2025-10-02
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2510.01560v1

This paper introduces an Artificial Intelligence (AI) foundation model for time series in engineering applications, where causal operations are required for real-time monitoring and control. Since engineering time series are governed by physical, rather than linguistic, laws, large-language-model-based AI foundation models may be ineffective or inefficient. Building on the classical innovations representation theory of Wiener, Kallianpur, and Rosenblatt, we propose Time Series GPT (TS-GPT) -- an innovations-representation-based Generative Pre-trained Transformer for engineering monitoring and control. As an example of foundation model adaptation, we consider Probabilistic Generative Forecasting, which produces future time series samples from conditional probability distributions given past realizations. We demonstrate the effectiveness of TS-GPT in forecasting real-time locational marginal prices using historical data from U.S. independent system operators.

---

### Consistent Assistant Domains Transformer for Source-free Domain Adaptation

Authors: Renrong Shao, Wei Zhang, Kangyang Luo, Qin Li, and Jun Wang
Published: 2025-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.01559v1

Source-free domain adaptation (SFDA) aims to address the challenge of adapting to a target domain without accessing the source domain directly. However, due to the inaccessibility of source domain data, deterministic invariable features cannot be obtained. Current mainstream methods primarily focus on evaluating invariant features in the target domain that closely resemble those in the source domain, subsequently aligning the target domain with the source domain. However, these methods are susceptible to hard samples and influenced by domain bias. In this paper, we propose a Consistent Assistant Domains Transformer for SFDA, abbreviated as CADTrans, which solves the issue by constructing invariable feature representations of domain consistency. Concretely, we develop an assistant domain module for CADTrans to obtain diversified representations from the intermediate aggregated global attentions, which addresses the limitation of existing methods in adequately representing diversity. Based on assistant and target domains, invariable feature representations are obtained by multiple consistent strategies, which can be used to distinguish easy and hard samples. Finally, to align the hard samples to the corresponding easy samples, we construct a conditional multi-kernel max mean discrepancy (CMK-MMD) strategy to distinguish between samples of the same category and those of different categories. Extensive experiments are conducted on various benchmarks such as Office-31, Office-Home, VISDA-C, and DomainNet-126, proving the significant performance improvements achieved by our proposed approaches. Code is available at https://github.com/RoryShao/CADTrans.git.

---

### CardioRAG: A Retrieval-Augmented Generation Framework for Multimodal Chagas Disease Detection

Authors: Zhengyang Shen, Xuehao Zhai, Hua Tu, Mayue Shi
Published: 2025-10-02
Categories: cs.CE, cs.LG, eess.SP
arXiv: https://arxiv.org/abs/2510.01558v1

Chagas disease affects nearly 6 million people worldwide, with Chagas cardiomyopathy representing its most severe complication. In regions where serological testing capacity is limited, AI-enhanced electrocardiogram (ECG) screening provides a critical diagnostic alternative. However, existing machine learning approaches face challenges such as limited accuracy, reliance on large labeled datasets, and more importantly, weak integration with evidence-based clinical diagnostic indicators. We propose a retrieval-augmented generation framework, CardioRAG, integrating large language models with interpretable ECG-based clinical features, including right bundle branch block, left anterior fascicular block, and heart rate variability metrics. The framework uses variational autoencoder-learned representations for semantic case retrieval, providing contextual cases to guide clinical reasoning. Evaluation demonstrated high recall performance of 89.80%, with a maximum F1 score of 0.68 for effective identification of positive cases requiring prioritized serological testing. CardioRAG provides an interpretable, clinical evidence-based approach particularly valuable for resource-limited settings, demonstrating a pathway for embedding clinical indicators into trustworthy medical AI systems.

---

### Rethinking KL Regularization in RLHF: From Value Estimation to Gradient Optimization

Authors: Kezhao Liu, Jason Klein Liu, Mingtao Chen, Yiming Liu
Published: 2025-10-02
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.01555v2

Reinforcement Learning from Human Feedback (RLHF) leverages a Kullback-Leibler (KL) divergence loss to stabilize training and prevent overfitting. However, in methods such as GRPO, its implementation may be guided by principles from numerical value estimation-a practice that overlooks the term's functional role as an optimization loss. To analyze this issue, we establish a unified framework that connects two seemingly distinct implementation styles: using the mathematical term $k_n$ as a detached coefficient for the policy's score function ('$k_n$ in reward') or as a direct loss function through which gradients are propagated ('$k_n$ as loss'). We show that the latter can always be analyzed via an equivalent gradient coefficient in the former, unifying the two perspectives. Through this framework, we prove that the conventional '$k_1$ in reward' (like in PPO) is the principled loss for Reverse KL (RKL) regularization. We further establish a key finding: under on-policy conditions, the '$k_2$ as loss' formulation is, in fact, gradient-equivalent to '$k_1$ in reward'. This equivalence, first proven in our work, identifies both as the theoretically sound implementations of the RKL objective. In contrast, we show that the recently adopted '$k_3$ as loss' (like in GRPO) is merely a first-order, biased approximation of the principled loss. Furthermore, we argue that common off-policy implementations of '$k_n$ as loss' methods are biased due to neglected importance sampling, and we propose a principled correction. Our findings provide a comprehensive, gradient-based rationale for choosing and correctly implementing KL regularization, paving the way for more robust and effective RLHF systems.

---

### POLAR: Automating Cyber Threat Prioritization through LLM-Powered Assessment

Authors: Luoxi Tang, Yuqiao Meng, Ankita Patra, Weicheng Ma, Muchao Ye, Zhaohan Xi
Published: 2025-10-02
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2510.01552v1

Large Language Models (LLMs) are intensively used to assist security analysts in counteracting the rapid exploitation of cyber threats, wherein LLMs offer cyber threat intelligence (CTI) to support vulnerability assessment and incident response. While recent work has shown that LLMs can support a wide range of CTI tasks such as threat analysis, vulnerability detection, and intrusion defense, significant performance gaps persist in practical deployments. In this paper, we investigate the intrinsic vulnerabilities of LLMs in CTI, focusing on challenges that arise from the nature of the threat landscape itself rather than the model architecture. Using large-scale evaluations across multiple CTI benchmarks and real-world threat reports, we introduce a novel categorization methodology that integrates stratification, autoregressive refinement, and human-in-the-loop supervision to reliably analyze failure instances. Through extensive experiments and human inspections, we reveal three fundamental vulnerabilities: spurious correlations, contradictory knowledge, and constrained generalization, that limit LLMs in effectively supporting CTI. Subsequently, we provide actionable insights for designing more robust LLM-powered CTI systems to facilitate future research.

---

### MIRA: Towards Mitigating Reward Hacking in Inference-Time Alignment of T2I Diffusion Models

Authors: Kevin Zhai, Utsav Singh, Anirudh Thatipelli, Souradip Chakraborty, Anit Kumar Sahu, Furong Huang, Amrit Singh Bedi, Mubarak Shah
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.01549v1

Diffusion models excel at generating images conditioned on text prompts, but the resulting images often do not satisfy user-specific criteria measured by scalar rewards such as Aesthetic Scores. This alignment typically requires fine-tuning, which is computationally demanding. Recently, inference-time alignment via noise optimization has emerged as an efficient alternative, modifying initial input noise to steer the diffusion denoising process towards generating high-reward images. However, this approach suffers from reward hacking, where the model produces images that score highly, yet deviate significantly from the original prompt. We show that noise-space regularization is insufficient and that preventing reward hacking requires an explicit image-space constraint. To this end, we propose MIRA (MItigating Reward hAcking), a training-free, inference-time alignment method. MIRA introduces an image-space, score-based KL surrogate that regularizes the sampling trajectory with a frozen backbone, constraining the output distribution so reward can increase without off-distribution drift (reward hacking). We derive a tractable approximation to KL using diffusion scores. Across SDv1.5 and SDXL, multiple rewards (Aesthetic, HPSv2, PickScore), and public datasets (e.g., Animal-Animal, HPDv2), MIRA achieves >60\% win rate vs. strong baselines while preserving prompt adherence; mechanism plots show reward gains with near-zero drift, whereas DNO drifts as compute increases. We further introduce MIRA-DPO, mapping preference optimization to inference time with a frozen backbone, extending MIRA to non-differentiable rewards without fine-tuning.

---

### Robust Classification of Oral Cancer with Limited Training Data

Authors: Akshay Bhagwan Sonawane, Lena D. Swamikannan, Lakshman Tamil
Published: 2025-10-02
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.01547v1

Oral cancer ranks among the most prevalent cancers globally, with a particularly high mortality rate in regions lacking adequate healthcare access. Early diagnosis is crucial for reducing mortality; however, challenges persist due to limited oral health programs, inadequate infrastructure, and a shortage of healthcare practitioners. Conventional deep learning models, while promising, often rely on point estimates, leading to overconfidence and reduced reliability. Critically, these models require large datasets to mitigate overfitting and ensure generalizability, an unrealistic demand in settings with limited training data. To address these issues, we propose a hybrid model that combines a convolutional neural network (CNN) with Bayesian deep learning for oral cancer classification using small training sets. This approach employs variational inference to enhance reliability through uncertainty quantification. The model was trained on photographic color images captured by smartphones and evaluated on three distinct test datasets. The proposed method achieved 94% accuracy on a test dataset with a distribution similar to that of the training data, comparable to traditional CNN performance. Notably, for real-world photographic image data, despite limitations and variations differing from the training dataset, the proposed model demonstrated superior generalizability, achieving 88% accuracy on diverse datasets compared to 72.94% for traditional CNNs, even with a smaller dataset. Confidence analysis revealed that the model exhibits low uncertainty (high confidence) for correctly classified samples and high uncertainty (low confidence) for misclassified samples. These results underscore the effectiveness of Bayesian inference in data-scarce environments in enhancing early oral cancer diagnosis by improving model reliability and generalizability.

---

### Growing Visual Generative Capacity for Pre-Trained MLLMs

Authors: Hanyu Wang, Jiaming Han, Ziyan Yang, Qi Zhao, Shanchuan Lin, Xiangyu Yue, Abhinav Shrivastava, Zhenheng Yang, Hao Chen
Published: 2025-10-02
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.01546v1

Multimodal large language models (MLLMs) extend the success of language models to visual understanding, and recent efforts have sought to build unified MLLMs that support both understanding and generation. However, constructing such models remains challenging: hybrid approaches combine continuous embeddings with diffusion or flow-based objectives, producing high-quality images but breaking the autoregressive paradigm, while pure autoregressive approaches unify text and image prediction over discrete visual tokens but often face trade-offs between semantic alignment and pixel-level fidelity. In this work, we present Bridge, a pure autoregressive unified MLLM that augments pre-trained visual understanding models with generative ability through a Mixture-of-Transformers architecture, enabling both image understanding and generation within a single next-token prediction framework. To further improve visual generation fidelity, we propose a semantic-to-pixel discrete representation that integrates compact semantic tokens with fine-grained pixel tokens, achieving strong language alignment and precise description of visual details with only a 7.9% increase in sequence length. Extensive experiments across diverse multimodal benchmarks demonstrate that Bridge achieves competitive or superior results in both understanding and generation benchmarks, while requiring less training data and reduced training time compared to prior unified MLLMs.

---

### Predictive Preference Learning from Human Interventions

Authors: Haoyuan Cai, Zhenghao Peng, Bolei Zhou
Published: 2025-10-02
Categories: cs.LG, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2510.01545v2

Learning from human involvement aims to incorporate the human subject to monitor and correct agent behavior errors. Although most interactive imitation learning methods focus on correcting the agent's action at the current state, they do not adjust its actions in future states, which may be potentially more hazardous. To address this, we introduce Predictive Preference Learning from Human Interventions (PPL), which leverages the implicit preference signals contained in human interventions to inform predictions of future rollouts. The key idea of PPL is to bootstrap each human intervention into L future time steps, called the preference horizon, with the assumption that the agent follows the same action and the human makes the same intervention in the preference horizon. By applying preference optimization on these future states, expert corrections are propagated into the safety-critical regions where the agent is expected to explore, significantly improving learning efficiency and reducing human demonstrations needed. We evaluate our approach with experiments on both autonomous driving and robotic manipulation benchmarks and demonstrate its efficiency and generality. Our theoretical analysis further shows that selecting an appropriate preference horizon L balances coverage of risky states with label correctness, thereby bounding the algorithmic optimality gap. Demo and code are available at: https://metadriverse.github.io/ppl

---

### Advancing Reasoning in Diffusion Language Models with Denoising Process Rewards

Authors: Shaoan Xie, Lingjing Kong, Xiangchen Song, Xinshuai Dong, Guangyi Chen, Eric P. Xing, Kun Zhang
Published: 2025-10-02
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.01544v2

Diffusion-based large language models offer a non-autoregressive alternative for text generation, but enabling them to perform complex reasoning remains challenging. Reinforcement learning has recently emerged as an effective post-training strategy for improving their performance; however, existing methods rely primarily on outcome-based rewards, which provide no direct supervision over the denoising process and often result in poorly structured reasoning that is difficult to interpret and inconsistently supports the final prediction. To address this limitation, we introduce \emph{denoising process reward}, a process-level reinforcement signal defined over the denoising trajectory of diffusion language models. This reward is obtained by estimating the contribution of intermediate denoising intervals to the final task outcome, encouraging the model to favor reasoning trajectories that consistently guide generation toward correct predictions. We further propose an efficient stochastic estimator that reuses standard training rollouts, enabling practical process-level supervision at scale. Experiments on challenging reasoning benchmarks demonstrate that our approach yields consistent improvements in reasoning stability, interpretability, and overall task performance.

---

### Towards Better Optimization For Listwise Preference in Diffusion Models

Authors: Jiamu Bai, Xin Yu, Meilong Xu, Weitao Lu, Xin Pan, Kiwan Maeng, Daniel Kifer, Jian Wang, Yu Wang
Published: 2025-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.01540v1

Reinforcement learning from human feedback (RLHF) has proven effectiveness for aligning text-to-image (T2I) diffusion models with human preferences. Although Direct Preference Optimization (DPO) is widely adopted for its computational efficiency and avoidance of explicit reward modeling, its applications to diffusion models have primarily relied on pairwise preferences. The precise optimization of listwise preferences remains largely unaddressed. In practice, human feedback on image preferences often contains implicit ranked information, which conveys more precise human preferences than pairwise comparisons. In this work, we propose Diffusion-LPO, a simple and effective framework for Listwise Preference Optimization in diffusion models with listwise data. Given a caption, we aggregate user feedback into a ranked list of images and derive a listwise extension of the DPO objective under the Plackett-Luce model. Diffusion-LPO enforces consistency across the entire ranking by encouraging each sample to be preferred over all of its lower-ranked alternatives. We empirically demonstrate the effectiveness of Diffusion-LPO across various tasks, including text-to-image generation, image editing, and personalized preference alignment. Diffusion-LPO consistently outperforms pairwise DPO baselines on visual quality and preference alignment.

---

### Executable Counterfactuals: Improving LLMs' Causal Reasoning Through Code

Authors: Aniket Vashishtha, Qirun Dai, Hongyuan Mei, Amit Sharma, Chenhao Tan, Hao Peng
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.01539v1

Counterfactual reasoning, a hallmark of intelligence, consists of three steps: inferring latent variables from observations (abduction), constructing alternatives (interventions), and predicting their outcomes (prediction). This skill is essential for advancing LLMs' causal understanding and expanding their applications in high-stakes domains such as scientific research. However, existing efforts in assessing LLM's counterfactual reasoning capabilities tend to skip the abduction step, effectively reducing to interventional reasoning and leading to overestimation of LLM performance. To address this, we introduce executable counterfactuals, a novel framework that operationalizes causal reasoning through code and math problems. Our framework explicitly requires all three steps of counterfactual reasoning and enables scalable synthetic data creation with varying difficulty, creating a frontier for evaluating and improving LLM's reasoning. Our results reveal substantial drop in accuracy (25-40%) from interventional to counterfactual reasoning for SOTA models like o4-mini and Claude-4-Sonnet. To address this gap, we construct a training set comprising counterfactual code problems having if-else condition and test on out-of-domain code structures (e.g. having while-loop); we also test whether a model trained on code would generalize to counterfactual math word problems. While supervised finetuning on stronger models' reasoning traces improves in-domain performance of Qwen models, it leads to a decrease in accuracy on OOD tasks such as counterfactual math problems. In contrast, reinforcement learning induces the core cognitive behaviors and generalizes to new domains, yielding gains over the base model on both code (improvement of 1.5x-2x) and math problems. Analysis of the reasoning traces reinforces these findings and highlights the promise of RL for improving LLMs' counterfactual reasoning.

---

### TimeSeriesScientist: A General-Purpose AI Agent for Time Series Analysis

Authors: Haokun Zhao, Xiang Zhang, Jiaqi Wei, Yiwei Xu, Yuting He, Siqi Sun, Chenyu You
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.01538v2

Time series forecasting is central to decision-making in domains as diverse as energy, finance, climate, and public health. In practice, forecasters face thousands of short, noisy series that vary in frequency, quality, and horizon, where the dominant cost lies not in model fitting, but in the labor-intensive preprocessing, validation, and ensembling required to obtain reliable predictions. Prevailing statistical and deep learning models are tailored to specific datasets or domains and generalize poorly. A general, domain-agnostic framework that minimizes human intervention is urgently in demand. In this paper, we introduce TimeSeriesScientist (TSci), the first LLM-driven agentic framework for general time series forecasting. The framework comprises four specialized agents: Curator performs LLM-guided diagnostics augmented by external tools that reason over data statistics to choose targeted preprocessing; Planner narrows the hypothesis space of model choice by leveraging multi-modal diagnostics and self-planning over the input; Forecaster performs model fitting and validation and, based on the results, adaptively selects the best model configuration as well as ensemble strategy to make final predictions; and Reporter synthesizes the whole process into a comprehensive, transparent report. With transparent natural-language rationales and comprehensive reports, TSci transforms the forecasting workflow into a white-box system that is both interpretable and extensible across tasks. Empirical results on eight established benchmarks demonstrate that TSci consistently outperforms both statistical and LLM-based baselines, reducing forecast error by an average of 10.4% and 38.2%, respectively. Moreover, TSci produces a clear and rigorous report that makes the forecasting workflow more transparent and interpretable.

---

### NVIDIA AI Aerial: AI-Native Wireless Communications

Authors: Kobi Cohen-Arazi, Michael Roe, Zhen Hu, Rohan Chavan, Anna Ptasznik, Joanna Lin, Joao Morais, Joseph Boccuzzi, Tommaso Balercia
Published: 2025-10-02
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.01533v1

6G brings a paradigm shift towards AI-native wireless systems, necessitating the seamless integration of digital signal processing (DSP) and machine learning (ML) within the software stacks of cellular networks. This transformation brings the life cycle of modern networks closer to AI systems, where models and algorithms are iteratively trained, simulated, and deployed across adjacent environments. In this work, we propose a robust framework that compiles Python-based algorithms into GPU-runnable blobs. The result is a unified approach that ensures efficiency, flexibility, and the highest possible performance on NVIDIA GPUs. As an example of the capabilities of the framework, we demonstrate the efficacy of performing the channel estimation function in the PUSCH receiver through a convolutional neural network (CNN) trained in Python. This is done in a digital twin first, and subsequently in a real-time testbed. Our proposed methodology, realized in the NVIDIA AI Aerial platform, lays the foundation for scalable integration of AI/ML models into next-generation cellular systems, and is essential for realizing the vision of natively intelligent 6G networks.

---

### MATCH: Multi-faceted Adaptive Topo-Consistency for Semi-Supervised Histopathology Segmentation

Authors: Meilong Xu, Xiaoling Hu, Shahira Abousamra, Chen Li, Chao Chen
Published: 2025-10-02
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.01532v2

In semi-supervised segmentation, capturing meaningful semantic structures from unlabeled data is essential. This is particularly challenging in histopathology image analysis, where objects are densely distributed. To address this issue, we propose a semi-supervised segmentation framework designed to robustly identify and preserve relevant topological features. Our method leverages multiple perturbed predictions obtained through stochastic dropouts and temporal training snapshots, enforcing topological consistency across these varied outputs. This consistency mechanism helps distinguish biologically meaningful structures from transient and noisy artifacts. A key challenge in this process is to accurately match the corresponding topological features across the predictions in the absence of ground truth. To overcome this, we introduce a novel matching strategy that integrates spatial overlap with global structural alignment, minimizing discrepancies among predictions. Extensive experiments demonstrate that our approach effectively reduces topological errors, resulting in more robust and accurate segmentations essential for reliable downstream analysis. Code is available at https://github.com/Melon-Xu/MATCH.

---

### Information Seeking for Robust Decision Making under Partial Observability

Authors: Djengo Cyun-Jyun Fang, Tsung-Wei Ke
Published: 2025-10-02
Categories: cs.AI, cs.CL, cs.RO
arXiv: https://arxiv.org/abs/2510.01531v1

Explicit information seeking is essential to human problem-solving in practical environments characterized by incomplete information and noisy dynamics. When the true environmental state is not directly observable, humans seek information to update their internal dynamics and inform future decision-making. Although existing Large Language Model (LLM) planning agents have addressed observational uncertainty, they often overlook discrepancies between their internal dynamics and the actual environment. We introduce Information Seeking Decision Planner (InfoSeeker), an LLM decision-making framework that integrates task-oriented planning with information seeking to align internal dynamics and make optimal decisions under uncertainty in both agent observations and environmental dynamics. InfoSeeker prompts an LLM to actively gather information by planning actions to validate its understanding, detect environmental changes, or test hypotheses before generating or revising task-oriented plans. To evaluate InfoSeeker, we introduce a novel benchmark suite featuring partially observable environments with incomplete observations and uncertain dynamics. Experiments demonstrate that InfoSeeker achieves a 74% absolute performance gain over prior methods without sacrificing sample efficiency. Moreover, InfoSeeker generalizes across LLMs and outperforms baselines on established benchmarks such as robotic manipulation and web navigation. These findings underscore the importance of tightly integrating planning and information seeking for robust behavior in partially observable environments. The project page is available at https://infoseekerllm.github.io

---

### LOGicalThought: Logic-Based Ontological Grounding of LLMs for High-Assurance Reasoning

Authors: Navapat Nananukul, Yue Zhang, Ryan Lee, Eric Boxer, Jonathan May, Vibhav Giridhar Gogate, Jay Pujara, Mayank Kejriwal
Published: 2025-10-02
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.01530v1

High-assurance reasoning, particularly in critical domains such as law and medicine, requires conclusions that are accurate, verifiable, and explicitly grounded in evidence. This reasoning relies on premises codified from rules, statutes, and contracts, inherently involving defeasible or non-monotonic logic due to numerous exceptions, where the introduction of a single fact can invalidate general rules, posing significant challenges. While large language models (LLMs) excel at processing natural language, their capabilities in standard inference tasks do not translate to the rigorous reasoning required over high-assurance text guidelines. Core reasoning challenges within such texts often manifest specific logical structures involving negation, implication, and, most critically, defeasible rules and exceptions. In this paper, we propose a novel neurosymbolically-grounded architecture called LOGicalThought (LogT) that uses an advanced logical language and reasoner in conjunction with an LLM to construct a dual symbolic graph context and logic-based context. These two context representations transform the problem from inference over long-form guidelines into a compact grounded evaluation. Evaluated on four multi-domain benchmarks against four baselines, LogT improves overall performance by 11.84% across all LLMs. Performance improves significantly across all three modes of reasoning: by up to +10.2% on negation, +13.2% on implication, and +5.5% on defeasible reasoning compared to the strongest baseline.

---

### Bypassing Prompt Guards in Production with Controlled-Release Prompting

Authors: Jaiden Fairoze, Sanjam Garg, Keewoo Lee, Mingyuan Wang
Published: 2025-10-02
Categories: cs.LG, cs.CR
arXiv: https://arxiv.org/abs/2510.01529v4

Ball et al. recently established that prompt filtering for AI alignment faces a fundamental barrier: under standard cryptographic assumptions, no filter running significantly faster than the protected model can universally distinguish adversarial prompts from benign ones. We investigate whether this impossibility result translates to real-world vulnerabilities in deployed large language model (LLM) systems. We answer affirmatively by introducing controlled-release prompting, a practical instantiation of the theoretical framework that exploits the resource asymmetry between lightweight input filters and the main models they protect. Unlike the theoretical construction, our attack does not require model modification: it generates malicious prompts that are indecipherable by any bounded filter yet remain tractable to the target LLM. We find our attack to be successful on four major chat platforms (Google Gemini, DeepSeek Chat, xAI Grok, and Mistral Le Chat) where baseline methods fail. Additionally, we apply our attack to extract copyrighted data from Gemini. Finally, we provide a systematic evaluation of 14 open-weight prompt guard models, revealing that even reasoning-capable filters cannot reliably detect our attack without incurring prohibitive resource overhead.

---

### Towards Interpretable and Inference-Optimal COT Reasoning with Sparse Autoencoder-Guided Generation

Authors: Daniel Zhao, Abhilash Shankarampeta, Lanxiang Hu, Tajana Rosing, Hao Zhang
Published: 2025-10-02
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.01528v1

We propose a novel method that leverages sparse autoencoders (SAEs) and clustering techniques to analyze the internal token representations of large language models (LLMs) and guide generations in mathematical reasoning tasks. Our approach first trains an SAE to generate sparse vector representations for training tokens, then applies k-means clustering to construct a graph where vertices represent token clusters and weighted edges capture sequential token transitions. Using this graph, we define an edge-weight based reward function to quantify adherence to established reasoning traces, thereby identifying exploitative reasoning trajectories. Additionally, we measure generation diversity from clustering to assess the extent of exploration. Our findings indicate that balancing both exploitation and exploration is crucial for achieving high accuracy in mathematical reasoning tasks. During generation, the SAE can serve as a scalable reward model to guide generations, ensuring a balanced trade-off between exploitation and exploration. This prevents extreme behaviors in either direction, ultimately fostering a higher-quality reasoning process in LLMs.

---

### Causally-Informed Reinforcement Learning for Adaptive Emotion-Aware Social Media Recommendation

Authors: Bhavika Jain, Robert Pitsko, Ananya Drishti, Mahfuza Farooque
Published: 2025-10-01
Categories: cs.IR, cs.AI
arXiv: https://arxiv.org/abs/2511.14768v1

Social media recommendation systems play a central role in shaping users' emotional experiences. However, most systems are optimized solely for engagement metrics, such as click rate, viewing time, or scrolling, without accounting for users' emotional states. Repeated exposure to emotionally charged content has been shown to negatively affect users' emotional well-being over time. We propose an Emotion-aware Social Media Recommendation (ESMR) framework that personalizes content based on users' evolving emotional trajectories. ESMR integrates a Transformer-based emotion predictor with a hybrid recommendation policy: a LightGBM model for engagement during stable periods and a reinforcement learning agent with causally informed rewards when negative emotional states persist. Through behaviorally grounded evaluation over 30-day interaction traces, ESMR demonstrates improved emotional recovery, reduced volatility, and strong engagement retention. ESMR offers a path toward emotionally aware recommendations without compromising engagement performance.

---

### Social and Physical Attributes-Defined Trust Evaluation for Effective Collaborator Selection in Human-Device Coexistence Systems

Authors: Botao Zhu, Xianbin Wang
Published: 2025-10-01
Categories: cs.HC, cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2511.11578v1

In human-device coexistence systems, collaborations among devices are determined by not only physical attributes such as network topology but also social attributes among human users. Consequently, trust evaluation of potential collaborators based on these multifaceted attributes becomes critical for ensuring the eventual outcome. However, due to the high heterogeneity and complexity of physical and social attributes, efficiently integrating them for accurate trust evaluation remains challenging. To overcome this difficulty, a canonical correlation analysis-enhanced hypergraph self-supervised learning (HSLCCA) method is proposed in this research. First, by treating all attributes as relationships among connected devices, a relationship hypergraph is constructed to comprehensively capture inter-device relationships across three dimensions: spatial attribute-related, device attribute-related, and social attribute-related. Next, a self-supervised learning framework is developed to integrate these multi-dimensional relationships and generate device embeddings enriched with relational semantics. In this learning framework, the relationship hypergraph is augmented into two distinct views to enhance semantic information. A parameter-sharing hypergraph neural network is then utilized to learn device embeddings from both views. To further enhance embedding quality, a CCA approach is applied, allowing the comparison of data between the two views. Finally, the trustworthiness of devices is calculated based on the learned device embeddings. Extensive experiments demonstrate that the proposed HSLCCA method significantly outperforms the baseline algorithm in effectively identifying trusted devices.

---

### Conversational Agents for Building Energy Efficiency -- Advising Housing Cooperatives in Stockholm on Reducing Energy Consumption

Authors: Shadaab Ghani, Anne Håkansson, Oleksii Pasichnyi, Hossein Shahrokni
Published: 2025-10-01
Categories: cs.CY, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2511.08587v1

Housing cooperative is a common type of multifamily building ownership in Sweden. Although this ownership structure grants decision-making autonomy, it places a burden of responsibility on cooperative's board members. Most board members lack the resources or expertise to manage properties and their energy consumption. This ignorance presents a unique challenge, especially given the EU directives that prohibit buildings rated as energy classes F and G by 2033. Conversational agents (CAs) enable human-like interactions with computer systems, facilitating human-computer interaction across various domains. In our case, CAs can be implemented to support cooperative members in making informed energy retrofitting and usage decisions. This paper introduces a Conversational agent system, called SPARA, designed to advise cooperatives on energy efficiency. SPARA functions as an energy efficiency advisor by leveraging the Retrieval-Augmented Generation (RAG) framework with a Language Model(LM). The LM generates targeted recommendations based on a knowledge base composed of email communications between professional energy advisors and cooperatives' representatives in Stockholm. The preliminary results indicate that SPARA can provide energy efficiency advice with precision 80\%, comparable to that of municipal energy efficiency (EE) experts. A pilot implementation is currently underway, where municipal EE experts are evaluating SPARA performance based on questions posed to EE experts by BRF members. Our findings suggest that LMs can significantly improve outreach by supporting stakeholders in their energy transition. For future work, more research is needed to evaluate this technology, particularly limitations to the stability and trustworthiness of its energy efficiency advice.

---

### Analyse comparative d'algorithmes de restauration en architecture dépliée pour des signaux chromatographiques parcimonieux

Authors: Mouna Gharbi, Silvia Villa, Emilie Chouzenoux, Jean-Christophe Pesquet, Laurent Duval
Published: 2025-10-01
Categories: eess.SP, cs.LG, physics.chem-ph
arXiv: https://arxiv.org/abs/2510.18760v1

Data restoration from degraded observations, of sparsity hypotheses, is an active field of study. Traditional iterative optimization methods are now complemented by deep learning techniques. The development of unfolded methods benefits from both families. We carry out a comparative study of three architectures on parameterized chromatographic signal databases, highlighting the performance of these approaches, especially when employing metrics adapted to physico-chemical peak signal characterization.

---

### Speak to a Protein: An Interactive Multimodal Co-Scientist

Authors: Carles Navarro, Mariona Torrens, Philipp Thölke, Stefan Doerr, Gianni De Fabritiis
Published: 2025-10-01
Categories: q-bio.BM, cs.AI
arXiv: https://arxiv.org/abs/2510.17826v2

Building a working mental model of a protein typically requires weeks of reading, cross-referencing crystal and predicted structures, and inspecting ligand complexes, an effort that is slow, unevenly accessible, and often requires specialized computational skills. We introduce \emph{Speak to a Protein}, a new capability that turns protein analysis into an interactive, multimodal dialogue with an expert co-scientist. The AI system retrieves and synthesizes relevant literature, structures, and ligand data; grounds answers in a live 3D scene; and can highlight, annotate, manipulate and see the visualization. It also generates and runs code when needed, explaining results in both text and graphics. We demonstrate these capabilities on relevant proteins, posing questions about binding pockets, conformational changes, or structure-activity relationships to test ideas in real time. \emph{Speak to a Protein} reduces the time from question to evidence, lowers the barrier to advanced structural analysis, and enables hypothesis generation by tightly coupling language, code, and 3D structures. \emph{Speak to a Protein} is freely accessible at https://open.playmolecule.org.

---

### Carbon-Aware Orchestration of Integrated Satellite Aerial Terrestrial Networks via Digital Twin

Authors: Shumaila Javaid, Nasir Saeed
Published: 2025-10-01
Categories: eess.SP, cs.AI, eess.SY
arXiv: https://arxiv.org/abs/2510.17825v1

Integrated Satellite Aerial Terrestrial Networks (ISATNs) are envisioned as key enablers of 6G, providing global connectivity for applications such as autonomous transportation, Industrial IoT, and disaster response. Their large-scale deployment, however, risks unsustainable energy use and carbon emissions. This work advances prior energy-aware studies by proposing a carbon-aware orchestration framework for ISATNs that leverages Digital Twin (DT) technology. The framework adopts grams of CO$_2$-equivalent per bit (gCO$_2$/bit) as a primary sustainability metric and implements a multi timescale Plan Do Check Act (PDCA) loop that combines day-ahead forecasting with real-time adaptive optimization. ISATN-specific control knobs, including carbon-aware handovers, UAV duty cycling, and renewable-aware edge placement, are exploited to reduce emissions. Simulation results with real carbon intensity data show up to 29\% lower gCO$_2$/bit than QoS-only orchestration, while improving renewable utilization and resilience under adverse events.

---

### Spiking Neural Network for Cross-Market Portfolio Optimization in Financial Markets: A Neuromorphic Computing Approach

Authors: Amarendra Mohan, Ameer Tamoor Khan, Shuai Li, Xinwei Cao, Zhibin Li
Published: 2025-10-01
Categories: q-fin.PM, cs.NE, q-fin.CP
arXiv: https://arxiv.org/abs/2510.15921v1

Cross-market portfolio optimization has become increasingly complex with the globalization of financial markets and the growth of high-frequency, multi-dimensional datasets. Traditional artificial neural networks, while effective in certain portfolio management tasks, often incur substantial computational overhead and lack the temporal processing capabilities required for large-scale, multi-market data. This study investigates the application of Spiking Neural Networks (SNNs) for cross-market portfolio optimization, leveraging neuromorphic computing principles to process equity data from both the Indian (Nifty 500) and US (S&P 500) markets. A five-year dataset comprising approximately 1,250 trading days of daily stock prices was systematically collected via the Yahoo Finance API. The proposed framework integrates Leaky Integrate-andFire neuron dynamics with adaptive thresholding, spike-timingdependent plasticity, and lateral inhibition to enable event-driven processing of financial time series. Dimensionality reduction is achieved through hierarchical clustering, while populationbased spike encoding and multiple decoding strategies support robust portfolio construction under realistic trading constraints, including cardinality limits, transaction costs, and adaptive risk aversion. Experimental evaluation demonstrates that the SNN-based framework delivers superior risk-adjusted returns and reduced volatility compared to ANN benchmarks, while substantially improving computational efficiency. These findings highlight the promise of neuromorphic computation for scalable, efficient, and robust portfolio optimization across global financial markets.

---

### Constrained Diffusion for Protein Design with Hard Structural Constraints

Authors: Jacob K. Christopher, Austin Seamann, Jingyi Cui, Sagar Khare, Ferdinando Fioretto
Published: 2025-10-01
Categories: q-bio.BM, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.14989v2

Diffusion models offer a powerful means of capturing the manifold of realistic protein structures, enabling rapid design for protein engineering tasks. However, existing approaches observe critical failure modes when precise constraints are necessary for functional design. To this end, we present a constrained diffusion framework for structure-guided protein design, ensuring strict adherence to functional requirements while maintaining precise stereochemical and geometric feasibility. The approach integrates proximal feasibility updates with ADMM decomposition into the generative process, scaling effectively to the complex constraint sets of this domain. We evaluate on challenging protein design tasks, including motif scaffolding and vacancy-constrained pocket design, while introducing a novel curated benchmark dataset for motif scaffolding in the PDZ domain. Our approach achieves state-of-the-art, providing perfect satisfaction of bonding and geometric constraints with no degradation in structural diversity.

---

### Articulation-Informed ASR: Integrating Articulatory Features into ASR via Auxiliary Speech Inversion and Cross-Attention Fusion

Authors: Ahmed Adel Attia, Jing Liu, Carol Espy Wilson
Published: 2025-10-01
Categories: eess.AS, cs.AI, cs.CL, cs.SD
arXiv: https://arxiv.org/abs/2510.08585v1

Prior works have investigated the use of articulatory features as complementary representations for automatic speech recognition (ASR), but their use was largely confined to shallow acoustic models. In this work, we revisit articulatory information in the era of deep learning and propose a framework that leverages articulatory representations both as an auxiliary task and as a pseudo-input to the recognition model. Specifically, we employ speech inversion as an auxiliary prediction task, and the predicted articulatory features are injected into the model as a query stream in a cross-attention module with acoustic embeddings as keys and values. Experiments on LibriSpeech demonstrate that our approach yields consistent improvements over strong transformer-based baselines, particularly under low-resource conditions. These findings suggest that articulatory features, once sidelined in ASR research, can provide meaningful benefits when reintroduced with modern architectures.

---

### Neu-RadBERT for Enhanced Diagnosis of Brain Injuries and Conditions

Authors: Manpreet Singh, Sean Macrae, Pierre-Marc Williams, Nicole Hung, Sabrina Araujo de Franca, Laurent Letourneau-Guillon, François-Martin Carrier, Bang Liu, Yiorgos Alexandros Cavayas
Published: 2025-10-01
Categories: q-bio.TO, cs.LG
arXiv: https://arxiv.org/abs/2510.06232v1

Objective: We sought to develop a classification algorithm to extract diagnoses from free-text radiology reports of brain imaging performed in patients with acute respiratory failure (ARF) undergoing invasive mechanical ventilation. Methods: We developed and fine-tuned Neu-RadBERT, a BERT-based model, to classify unstructured radiology reports. We extracted all the brain imaging reports (computed tomography and magnetic resonance imaging) from MIMIC-IV database, performed in patients with ARF. Initial manual labelling was performed on a subset of reports for various brain abnormalities, followed by fine-tuning Neu-RadBERT using three strategies: 1) baseline RadBERT, 2) Neu-RadBERT with Masked Language Modeling (MLM) pretraining, and 3) Neu-RadBERT with MLM pretraining and oversampling to address data skewness. We compared the performance of this model to Llama-2-13B, an autoregressive LLM. Results: The Neu-RadBERT model, particularly with oversampling, demonstrated significant improvements in diagnostic accuracy compared to baseline RadBERT for brain abnormalities, achieving up to 98.0% accuracy for acute brain injuries. Llama-2-13B exhibited relatively lower performance, peaking at 67.5% binary classification accuracy. This result highlights potential limitations of current autoregressive LLMs for this specific classification task, though it remains possible that larger models or further fine-tuning could improve performance. Conclusion: Neu-RadBERT, enhanced through target domain pretraining and oversampling techniques, offered a robust tool for accurate and reliable diagnosis of neurological conditions from radiology reports. This study underscores the potential of transformer-based NLP models in automatically extracting diagnoses from free text reports with potential applications to both research and patient care.

---

### CML-Bench: A Framework for Evaluating and Enhancing LLM-Powered Movie Scripts Generation

Authors: Mingzhe Zheng, Dingjie Song, Guanyu Zhou, Jun You, Jiahao Zhan, Xuran Ma, Xinyuan Song, Ser-Nam Lim, Qifeng Chen, Harry Yang
Published: 2025-10-01
Categories: cs.CV, cs.CL
arXiv: https://arxiv.org/abs/2510.06231v1

Large Language Models (LLMs) have demonstrated remarkable proficiency in generating highly structured texts. However, while exhibiting a high degree of structural organization, movie scripts demand an additional layer of nuanced storytelling and emotional depth-the 'soul' of compelling cinema-that LLMs often fail to capture. To investigate this deficiency, we first curated CML-Dataset, a dataset comprising (summary, content) pairs for Cinematic Markup Language (CML), where 'content' consists of segments from esteemed, high-quality movie scripts and 'summary' is a concise description of the content. Through an in-depth analysis of the intrinsic multi-shot continuity and narrative structures within these authentic scripts, we identified three pivotal dimensions for quality assessment: Dialogue Coherence (DC), Character Consistency (CC), and Plot Reasonableness (PR). Informed by these findings, we propose the CML-Bench, featuring quantitative metrics across these dimensions. CML-Bench effectively assigns high scores to well-crafted, human-written scripts while concurrently pinpointing the weaknesses in screenplays generated by LLMs. To further validate our benchmark, we introduce CML-Instruction, a prompting strategy with detailed instructions on character dialogue and event logic, to guide LLMs to generate more structured and cinematically sound scripts. Extensive experiments validate the effectiveness of our benchmark and demonstrate that LLMs guided by CML-Instruction generate higher-quality screenplays, with results aligned with human preferences.

---

### SynCED-EnDe 2025: A Synthetic and Curated English - German Dataset for Critical Error Detection in Machine Translation

Authors: Muskaan Chopra, Lorenz Sparrenberg, Rafet Sifa
Published: 2025-10-01
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.05144v1

Critical Error Detection (CED) in machine translation aims to determine whether a translation is safe to use or contains unacceptable deviations in meaning. While the WMT21 English-German CED dataset provided the first benchmark, it is limited in scale, label balance, domain coverage, and temporal freshness. We present SynCED-EnDe, a new resource consisting of 1,000 gold-labeled and 8,000 silver-labeled sentence pairs, balanced 50/50 between error and non-error cases. SynCED-EnDe draws from diverse 2024-2025 sources (StackExchange, GOV.UK) and introduces explicit error subclasses, structured trigger flags, and fine-grained auxiliary judgments (obviousness, severity, localization complexity, contextual dependency, adequacy deviation). These enrichments enable systematic analyses of error risk and intricacy beyond binary detection. The dataset is permanently hosted on GitHub and Hugging Face, accompanied by documentation, annotation guidelines, and baseline scripts. Benchmark experiments with XLM-R and related encoders show substantial performance gains over WMT21 due to balanced labels and refined annotations. We envision SynCED-EnDe as a community resource to advance safe deployment of MT in information retrieval and conversational assistants, particularly in emerging contexts such as wearable AI devices.

---

### Reliable End-to-End Material Information Extraction from the Literature with Source-Tracked Multi-Stage Large Language Models

Authors: Xin Wang, Anshu Raj, Matthew Luebbe, Haiming Wen, Shuozhi Xu, Kun Lu
Published: 2025-10-01
Categories: cs.CL, cond-mat.mtrl-sci
arXiv: https://arxiv.org/abs/2510.05142v1

Data-driven materials discovery requires large-scale experimental datasets, yet most of the information remains trapped in unstructured literature. Existing extraction efforts often focus on a limited set of features and have not addressed the integrated composition-processing-microstructure-property relationships essential for understanding materials behavior, thereby posing challenges for building comprehensive databases. To address this gap, we propose a multi-stage information extraction pipeline powered by large language models, which captures 47 features spanning composition, processing, microstructure, and properties exclusively from experimentally reported materials. The pipeline integrates iterative extraction with source tracking to enhance both accuracy and reliability. Evaluations at the feature level (independent attributes) and tuple level (interdependent features) yielded F1 scores around 0.96. Compared with single-pass extraction without source tracking, our approach improved F1 scores of microstructure category by 10.0% (feature level) and 13.7% (tuple level), and reduced missed materials from 49 to 13 out of 396 materials in 100 articles on precipitate-containing multi-principal element alloys (miss rate reduced from 12.4% to 3.3%). The pipeline enables scalable and efficient literature mining, producing databases with high precision, minimal omissions, and zero false positives. These datasets provide trustworthy inputs for machine learning and materials informatics, while the modular design generalizes to diverse material classes, enabling comprehensive materials information extraction.

---

### To model human linguistic prediction, make LLMs less superhuman

Authors: Byung-Doh Oh, Tal Linzen
Published: 2025-10-01
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05141v2

When we read, we make predictions about upcoming words; these predictions influence our reading behavior. The success of large language models (LLMs), which, like humans, make predictions about upcoming words, has motivated their use as models of human linguistic prediction. Surprisingly, in the last few years, as LLMs' ability to predict the next word has improved, their ability to explain reading behavior has declined. We argue this is because current LLMs can predict upcoming words much better than human readers can. This 'superhumanness' is driven by LLMs' extensive training data, stronger long-term memory of training examples, and stronger short-term memory. We advocate for LLMs with human-like memory and for new experiments to measure the alignment between humans and LLMs, and outline directions towards achieving these goals.

---

### Auditing Algorithmic Bias in Transformer-Based Trading

Authors: Armin Gerami, Ramani Duraiswami
Published: 2025-10-01
Categories: cs.LG, cs.CE
arXiv: https://arxiv.org/abs/2510.05140v2

Transformer models have become increasingly popular in financial applications, yet their potential risk making and biases remain under-explored. The purpose of this work is to audit the reliance of the model on volatile data for decision-making, and quantify how the frequency of price movements affects the model's prediction confidence. We employ a transformer model for prediction, and introduce a metric based on Partial Information Decomposition (PID) to measure the influence of each asset on the model's decision making. Our analysis reveals two key observations: first, the model disregards data volatility entirely, and second, it is biased toward data with lower-frequency price movements.

---

### NLD-LLM: A systematic framework for evaluating small language transformer models on natural language description

Authors: Hamed Jelodar, Mohammad Meymani, Parisa Hamedi, Tochukwu Emmanuel Nwankwo, Samita Bai, Roozbeh Razavi-Far, Ali A. Ghorbani
Published: 2025-10-01
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05139v1

Natural Language Description (NLD) is a Natural Language Processing (NLP) task that requires models to generate structured and meaningful outputs from natural language inputs. In this work, we propose NLD-LLM, a systematic NLP framework to evaluate the performance of language models to generate accurate and concise source code descriptions. This framework incorporates a diverse set of transformer models, including Qwen, DeepSeek, Phi, LLaMA, and Mistral, spanning various sizes, architectures, and training approaches. Central to NLD-LLM is a comprehensive prompt design strategy that includes standardized formatting, clear task guidance, and NLD prompting, ensuring fair and consistent evaluation. Additionally, we apply an iterative refinement process to improve output's quality and assess the model's adaptability. Using semantic and structural metrics, our analysis demonstrates that prompt engineering significantly impacts the effectiveness of the model such that smaller models often performing competitively when supported by well-crafted prompts.

---

### LiRA: A Multi-Agent Framework for Reliable and Readable Literature Review Generation

Authors: Gregory Hok Tjoan Go, Khang Ly, Anders Søgaard, Amin Tabatabaei, Maarten de Rijke, Xinyi Chen
Published: 2025-10-01
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05138v4

The rapid growth of scientific publications has made it increasingly difficult to keep literature reviews comprehensive and up-to-date. Though prior work has focused on automating retrieval and screening, the writing phase of systematic reviews remains largely under-explored, especially with regard to readability and factual accuracy. To address this, we present LiRA (Literature Review Agents), a multi-agent collaborative workflow which emulates the human literature review process. LiRA utilizes specialized agents for content outlining, subsection writing, editing, and reviewing, producing cohesive and comprehensive review articles. Evaluated on SciReviewGen and a proprietary ScienceDirect dataset, LiRA outperforms current baselines such as AutoSurvey and MASS-Survey in writing and citation quality, while maintaining competitive similarity to human-written reviews. We further evaluate LiRA in real-world scenarios using document retrieval and assess its robustness to reviewer model variation. Our findings highlight the potential of agentic LLM workflows, even without domain-specific tuning, to improve the reliability and usability of automated scientific writing.

---

### Demystifying deep search: a holistic evaluation with hint-free multi-hop questions and factorised metrics

Authors: Maojia Song, Renhang Liu, Xinyu Wang, Yong Jiang, Pengjun Xie, Fei Huang, Jingren Zhou, Dorien Herremans, Soujanya Poria
Published: 2025-10-01
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05137v3

RAG (Retrieval-Augmented Generation) systems and web agents are increasingly evaluated on multi-hop deep search tasks, yet current practice suffers from two major limitations. First, most benchmarks leak the reasoning path in the question text, allowing models to follow surface cues rather than discover reasoning chains autonomously. Second, evaluation is typically reduced to a single pass rate, which collapses diverse behaviours into one score and obscures whether failures stem from inadequate search, poor knowledge use, or inappropriate refusal. To address these issues, we present WebDetective, a benchmark of hint-free multi-hop questions paired with a controlled Wikipedia sandbox that ensures full traceability of model actions, and a holistic evaluation framework that separates search sufficiency, knowledge utilisation, and refusal behaviour. Our evaluation of 25 state-of-the-art models reveals systematic weaknesses across all architectures: models struggle with knowledge utilisation despite having sufficient evidence and demonstrate near-absent appropriate refusal when evidence is lacking. These patterns expose a fundamental gap: today's systems excel at executing given reasoning paths but fail when required to discover them. We develop an agentic workflow, EvidenceLoop, that explicitly targets the challenges our benchmark identifies, incorporating verification loops and systematic evidence tracking that improve both search and synthesis capabilities. This baseline demonstrates that WebDetective's diagnostic framework can guide concrete architectural improvements, establishing our benchmark as a critical tool for developing genuinely autonomous reasoning systems rather than pattern-following agents.

---

### Linguistic Characteristics of AI-Generated Text: A Survey

Authors: Luka Terčon, Kaja Dobrovoljc
Published: 2025-10-01
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.05136v1

Large language models (LLMs) are solidifying their position in the modern world as effective tools for the automatic generation of text. Their use is quickly becoming commonplace in fields such as education, healthcare, and scientific research. There is a growing need to study the linguistic features present in AI-generated text, as the increasing presence of such texts has profound implications in various disciplines such as corpus linguistics, computational linguistics, and natural language processing. Many observations have already been made, however a broader synthesis of the findings made so far is required to provide a better understanding of the topic. The present survey paper aims to provide such a synthesis of extant research. We categorize the existing works along several dimensions, including the levels of linguistic description, the models included, the genres analyzed, the languages analyzed, and the approach to prompting. Additionally, the same scheme is used to present the findings made so far and expose the current trends followed by researchers. Among the most-often reported findings is the observation that AI-generated text is more likely to contain a more formal and impersonal style, signaled by the increased presence of nouns, determiners, and adpositions and the lower reliance on adjectives and adverbs. AI-generated text is also more likely to feature a lower lexical diversity, a smaller vocabulary size, and repetitive text. Current research, however, remains heavily concentrated on English data and mostly on text generated by the GPT model family, highlighting the need for broader cross-linguistic and cross-model investigation. In most cases authors also fail to address the issue of prompt sensitivity, leaving much room for future studies that employ multiple prompt wordings in the text generation phase.

---

### Curiosity-Driven LLM-as-a-judge for Personalized Creative Judgment

Authors: Vanya Bannihatti Kumar, Divyanshu Goyal, Akhil Eppa, Neel Bhandari
Published: 2025-10-01
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2510.05135v1

Modern large language models (LLMs) excel at objective tasks such as evaluating mathematical reasoning and factual accuracy, yet they falter when faced with the nuanced, subjective nature of assessing creativity. In this work, we propose a novel curiosity-driven LLM-as-a-judge for evaluating creative writing which is personlized to each individual's creative judgments. We use the Torrance Test of Creative Thinking(TTCW) benchmark introduced in Chakrabarty et al. (2024), which has stories annotated by expert humans across various subjective dimensions like Originality, to test our hypothesis. We show that our method enables models across various sizes, to learn the nuanced creative judgments of different individuals, by showing improvements over baseline supervised finetuning(SFT) method across various evaluation metrics like Pearson correlation, Cohen's and F1 values. Our method is especially useful in subjective evaluations where not all the annotators agree with each other.

---

### Structuring Reasoning for Complex Rules Beyond Flat Representations

Authors: Zhihao Yang, Ancheng Xu, Jingpeng Li, Liang Yan, Jiehui Zhou, Zhen Qin, Hengyu Chang, Yukun Chen, Longze Chen, Ahmadreza Argha, Hamid Alinejad-Rokny, Minghuan Tan et al.
Published: 2025-10-01
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.05134v2

Large language models (LLMs) face significant challenges when processing complex rule systems, as they typically treat interdependent rules as unstructured textual data rather than as logically organized frameworks. This limitation results in reasoning divergence, where models often overlook critical rule dependencies essential for accurate interpretation. Although existing approaches such as Chain-of-Thought (CoT) reasoning have shown promise, they lack systematic methodologies for structured rule processing and are particularly susceptible to error propagation through sequential reasoning chains. To address these limitations, we propose the Dynamic Adjudication Template (DAT), a novel framework inspired by expert human reasoning processes. DAT structures the inference mechanism into three methodical stages: qualitative analysis, evidence gathering, and adjudication. During the qualitative analysis phase, the model comprehensively evaluates the contextual landscape. The subsequent evidence gathering phase involves the targeted extraction of pertinent information based on predefined template elements ((placeholder)), followed by systematic verification against applicable rules. Finally, in the adjudication phase, the model synthesizes these validated components to formulate a comprehensive judgment. Empirical results demonstrate that DAT consistently outperforms conventional CoT approaches in complex rule-based tasks. Notably, DAT enables smaller language models to match, and in some cases exceed, the performance of significantly larger LLMs, highlighting its efficiency and effectiveness in managing intricate rule systems.

---

### Characterizing Model Behavior Under Synthetic Data Training: An Empirical Study Across Scales and Mixing Ratios

Authors: Y. Du, G. Wu, G. Tang, W. Wang, Q. Fan
Published: 2025-10-01
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.05133v1

Synthetic data generated by large language models has become integral to modern NLP training pipelines, from bootstrapping reasoning capabilities to augmenting instruction-following datasets. While recent work demonstrates successful applications maintaining high external data ratios, systematic understanding of how synthetic data proportion affects model behavior across different scales remains limited. This paper presents a controlled empirical study examining model performance, calibration, and output characteristics when trained on varying synthetic-to-external data ratios. Using the Pythia model suite (410M-12B parameters) across five diverse tasks, we evaluate models after one to three training iterations with synthetic data proportions ranging from 0-50\%. Our key findings include: models maintain stable performance with up to 20\% synthetic data, but degradation accelerates beyond 30\%; larger models (6.9B-12B) show greater robustness to synthetic data than smaller models (410M-1.4B); calibration degradation precedes accuracy loss, providing an early warning signal; and task characteristics matter, with reasoning tasks degrading faster than retrieval tasks under synthetic data training. Importantly, we find that current best practices, such as those employed in STaR and Self-Instruct systems that maintain greater than 80\% external data, operate well within safe regimes identified by our experiments. We provide practical guidance for practitioners on synthetic data budgets based on model scale and task requirements, alongside detailed comparison with concurrent work including Shumailov et al.'s model collapse findings.

---

### Training Large Language Models To Reason In Parallel With Global Forking Tokens

Authors: Sheng Jia, Xiao Wang, Shiva Prasad Kasiviswanathan
Published: 2025-10-01
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.05132v3

Although LLMs have demonstrated improved performance by scaling parallel test-time compute, doing so relies on generating reasoning paths that are both diverse and accurate. For challenging problems, the forking tokens that trigger diverse yet correct reasoning modes are typically deep in the sampling tree. Consequently, common strategies to encourage diversity, such as temperature scaling, encounter a worsened trade-off between diversity and accuracy. Motivated by this challenge, we treat parallel reasoning as a set-of-next-token-prediction problem and incorporate a set-based global loss into Supervised Fine-Tuning (SFT) using bipartite matching between global forking tokens and unique reasoning traces. We observe that whereas naive fine-tuning with multiple reasoning traces collapses these unique reasoning modes, our proposed method, Set Supervised Fine-Tuning (SSFT), preserves these modes and produces emergent global forking tokens. Global Forking Policy Optimization (GFPO) leverages these maximally steerable tokens to incentivize complex reasoning, and the resulting models consistently outperform their SFT counterparts with GRPO on both math reasoning and execution-based code generation benchmarks.

---

### Rationale-Augmented Retrieval with Constrained LLM Re-Ranking for Task Discovery

Authors: Bowen Wei
Published: 2025-10-01
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.05131v1

Head Start programs utilizing GoEngage face significant challenges when new or rotating staff attempt to locate appropriate Tasks (modules) on the platform homepage. These difficulties arise from domain-specific jargon (e.g., IFPA, DRDP), system-specific nomenclature (e.g., Application Pool), and the inherent limitations of lexical search in handling typos and varied word ordering. We propose a pragmatic hybrid semantic search system that synergistically combines lightweight typo-tolerant lexical retrieval, embedding-based vector similarity, and constrained large language model (LLM) re-ranking. Our approach leverages the organization's existing Task Repository and Knowledge Base infrastructure while ensuring trustworthiness through low false-positive rates, evolvability to accommodate terminological changes, and economic efficiency via intelligent caching, shortlist generation, and graceful degradation mechanisms. We provide a comprehensive framework detailing required resources, a phased implementation strategy with concrete milestones, an offline evaluation protocol utilizing curated test cases (Hit@K, Precision@K, Recall@K, MRR), and an online measurement methodology incorporating query success metrics, zero-result rates, and dwell-time proxies.

---

### Intelligent Healthcare Ecosystems: Optimizing the Iron Triangle of Healthcare (Access, Cost, Quality)

Authors: Vivek Acharya
Published: 2025-10-01
Categories: cs.CY, cs.AI
arXiv: https://arxiv.org/abs/2510.03331v1

The United States spends nearly 17% of GDP on healthcare yet continues to face uneven access and outcomes. This well-known trade-off among cost, quality, and access - the "iron triangle" - motivates a system-level redesign. This paper proposes an Intelligent Healthcare Ecosystem (iHE): an integrated, data-driven framework that uses generative AI and large language models, federated learning, interoperability standards (FHIR, TEFCA), and digital twins to improve access and quality while lowering cost. We review historical spending trends, waste, and international comparisons; introduce a value equation that jointly optimizes access, quality, and cost; and synthesize evidence on the enabling technologies and operating model for iHE. Methods follow a narrative review of recent literature and policy reports. Results outline core components (AI decision support, interoperability, telehealth, automation) and show how iHE can reduce waste, personalize care, and support value-based payment while addressing privacy, bias, and adoption challenges. We argue that a coordinated iHE can bend - if not break - the iron triangle, moving the system toward care that is more accessible, affordable, and high quality.

---

### Constant in an Ever-Changing World

Authors: Andy Wu, Chun-Cheng Lin, Yuehua Huang, Rung-Tzuo Liaw
Published: 2025-10-01
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.03330v1

The training process of reinforcement learning often suffers from severe oscillations, leading to instability and degraded performance. In this paper, we propose a Constant in an Ever-Changing World (CIC) framework that enhances algorithmic stability to improve performance. CIC maintains both a representative policy and a current policy. Instead of updating the representative policy blindly, CIC selectively updates it only when the current policy demonstrates superiority. Furthermore, CIC employs an adaptive adjustment mechanism, enabling the representative and current policies to jointly facilitate critic training. We evaluate CIC on five MuJoCo environments, and the results show that CIC improves the performance of conventional algorithms without incurring additional computational cost.

---

### DECOR: Deep Embedding Clustering with Orientation Robustness

Authors: Fiona Victoria Stanley Jothiraj, Arunaggiri Pandian Karunanidhi, Seth A. Eichmeyer
Published: 2025-10-01
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.03328v2

In semiconductor manufacturing, early detection of wafer defects is critical for product yield optimization. However, raw wafer data from wafer quality tests are often complex, unlabeled, imbalanced and can contain multiple defects on a single wafer, making it crucial to design clustering methods that remain reliable under such imperfect data conditions. We introduce DECOR, a deep clustering with orientation robustness framework that groups complex defect patterns from wafer maps into consistent clusters. We evaluate our method on the open source MixedWM38 dataset, demonstrating its ability to discover clusters without manual tuning. DECOR explicitly accounts for orientation variations in wafer maps, ensuring that spatially similar defects are consistently clustered regardless of its rotation or alignment. Experiments indicate that our method outperforms existing clustering baseline methods, thus providing a reliable and scalable solution in automated visual inspection systems.

---

### NS-Pep: De novo Peptide Design with Non-Standard Amino Acids

Authors: Tao Guo, Junbo Yin, Yu Wang, Xin Gao
Published: 2025-10-01
Categories: q-bio.BM, cs.AI
arXiv: https://arxiv.org/abs/2510.03326v1

Peptide drugs incorporating non-standard amino acids (NSAAs) offer improved binding affinity and improved pharmacological properties. However, existing peptide design methods are limited to standard amino acids, leaving NSAA-aware design largely unexplored. We introduce NS-Pep, a unified framework for co-designing peptide sequences and structures with NSAAs. The main challenge is that NSAAs are extremely underrepresented-even the most frequent one, SEP, accounts for less than 0.4% of residues-resulting in a severe long-tailed distribution. To improve generalization to rare amino acids, we propose Residue Frequency-Guided Modification (RFGM), which mitigates over-penalization through frequency-aware logit calibration, supported by both theoretical and empirical analysis. Furthermore, we identify that insufficient side-chain modeling limits geometric representation of NSAAs. To address this, we introduce Progressive Side-chain Perception (PSP) for coarse-to-fine torsion and location prediction, and Interaction-Aware Weighting (IAW) to emphasize pocket-proximal residues. Moreover, NS-Pep generalizes naturally to the peptide folding task with NSAAs, addressing a major limitation of current tools. Experiments show that NS-Pep improves sequence recovery rate and binding affinity by 6.23% and 5.12%, respectively, and outperforms AlphaFold3 by 17.76% in peptide folding success rate.

---

### Fast frequency reconstruction using Deep Learning for event recognition in ring laser data

Authors: Giuseppe Di Somma, Giorgio Carelli, Angela D. V. Di Virgilio, Francesco Fuso, Enrico Maccioni, Paolo Marsili
Published: 2025-10-01
Categories: cs.LG, physics.comp-ph, physics.data-an, physics.geo-ph
arXiv: https://arxiv.org/abs/2510.03325v1

The reconstruction of a frequency with minimal delay from a sinusoidal signal is a common task in several fields; for example Ring Laser Gyroscopes, since their output signal is a beat frequency. While conventional methods require several seconds of data, we present a neural network approach capable of reconstructing frequencies of several hundred Hertz within approximately 10 milliseconds. This enables rapid trigger generation. The method outperforms standard Fourier-based techniques, improving frequency estimation precision by a factor of 2 in the operational range of GINGERINO, our Ring Laser Gyroscope.\\ In addition to fast frequency estimation, we introduce an automated classification framework to identify physical disturbances in the signal, such as laser instabilities and seismic events, achieving accuracy rates between 99\% and 100\% on independent test datasets for the seismic class. These results mark a step forward in integrating artificial intelligence into signal analysis for geophysical applications.

---

### Enhancing Agentic Textual Graph Retrieval with Synthetic Stepwise Supervision

Authors: Ge Chang, Jinbo Su, Jiacheng Liu, Pengfei Yang, Yuhao Shang, Huiwen Zheng, Hongli Ma, Yan Liang, Yuanchun Li, Yunxin Liu
Published: 2025-10-01
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.03323v3

Integrating textual graphs into Large Language Models (LLMs) is promising for complex graph-based QA. However, a key bottleneck is retrieving informative yet compact subgraphs that fit the LLM context. Existing retrievers often struggle, relying either on shallow embedding similarity or costly interactive policies that require excessive supervision. To address these challenges, we introduce an agentic textual graph reasoning framework featuring an LLM-based retriever trained with synthetic stepwise supervision. Rather than relying on final answer rewards which often yield sparse and unstable signals, we optimize the retriever by evaluating each step against offline-extracted golden subgraphs. Our approach distills golden subgraphs via a specialized data synthesis pipeline to formulate dense rewards, facilitating a two-stage training scheme that effectively learns the interactive graph exploration policy. Based on extensive experiments on three common datasets in comparison with seven strong baselines, our approach achieves an average improvement of 15.6% in accuracy and 17.2% in F1 score. The advantage is even higher in more complicated multi-hop reasoning tasks.

---