# Research Papers — 2025-10 · bundle page 120 of 217

> Full-text markdown of papers #5951–6000 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/120
[← Previous bundle](https://aipod.dev/papers/archive/2025-10/page/119.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-10/page/121.md)

### LiteVPNet: A Lightweight Network for Video Encoding Control in Quality-Critical Applications

Authors: Vibhoothi Vibhoothi, François Pitié, Anil Kokaram
Published: 2025-10-14
Categories: eess.IV, cs.AI, cs.LG, cs.MM
arXiv: https://arxiv.org/abs/2510.12379v1

In the last decade, video workflows in the cinema production ecosystem have presented new use cases for video streaming technology. These new workflows, e.g. in On-set Virtual Production, present the challenge of requiring precise quality control and energy efficiency. Existing approaches to transcoding often fall short of these requirements, either due to a lack of quality control or computational overhead. To fill this gap, we present a lightweight neural network (LiteVPNet) for accurately predicting Quantisation Parameters for NVENC AV1 encoders that achieve a specified VMAF score. We use low-complexity features, including bitstream characteristics, video complexity measures, and CLIP-based semantic embeddings. Our results demonstrate that LiteVPNet achieves mean VMAF errors below 1.2 points across a wide range of quality targets. Notably, LiteVPNet achieves VMAF errors within 2 points for over 87% of our test corpus, c.f. approx 61% with state-of-the-art methods. LiteVPNet's performance across various quality regions highlights its applicability for enhancing high-value content transport and streaming for more energy-efficient, high-quality media experiences.

---

### Deep Attention-guided Adaptive Subsampling

Authors: Sharath M Shankaranarayana, Soumava Kumar Roy, Prasad Sudhakar, Chandan Aladahalli
Published: 2025-10-14
Categories: cs.CV, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2510.12376v1

Although deep neural networks have provided impressive gains in performance, these improvements often come at the cost of increased computational complexity and expense. In many cases, such as 3D volume or video classification tasks, not all slices or frames are necessary due to inherent redundancies. To address this issue, we propose a novel learnable subsampling framework that can be integrated into any neural network architecture. Subsampling, being a nondifferentiable operation, poses significant challenges for direct adaptation into deep learning models. While some works, have proposed solutions using the Gumbel-max trick to overcome the problem of non-differentiability, they fall short in a crucial aspect: they are only task-adaptive and not inputadaptive. Once the sampling mechanism is learned, it remains static and does not adjust to different inputs, making it unsuitable for real-world applications. To this end, we propose an attention-guided sampling module that adapts to inputs even during inference. This dynamic adaptation results in performance gains and reduces complexity in deep neural network models. We demonstrate the effectiveness of our method on 3D medical imaging datasets from MedMNIST3D as well as two ultrasound video datasets for classification tasks, one of them being a challenging in-house dataset collected under real-world clinical conditions.

---

### Improved Central Limit Theorem and Bootstrap Approximations for Linear Stochastic Approximation

Authors: Bogdan Butyrin, Eric Moulines, Alexey Naumov, Sergey Samsonov, Qi-Man Shao, Zhuo-Song Zhang
Published: 2025-10-14
Categories: stat.ML, cs.LG, math.OC, math.PR, math.ST
arXiv: https://arxiv.org/abs/2510.12375v1

In this paper, we refine the Berry-Esseen bounds for the multivariate normal approximation of Polyak-Ruppert averaged iterates arising from the linear stochastic approximation (LSA) algorithm with decreasing step size. We consider the normal approximation by the Gaussian distribution with covariance matrix predicted by the Polyak-Juditsky central limit theorem and establish the rate up to order $n^{-1/3}$ in convex distance, where $n$ is the number of samples used in the algorithm. We also prove a non-asymptotic validity of the multiplier bootstrap procedure for approximating the distribution of the rescaled error of the averaged LSA estimator. We establish approximation rates of order up to $1/\sqrt{n}$ for the latter distribution, which significantly improves upon the previous results obtained by Samsonov et al. (2024).

---

### Constrained Sensing and Reliable State Estimation with Shallow Recurrent Decoders on a TRIGA Mark II Reactor

Authors: Stefano Riva, Carolina Introini, Josè Nathan Kutz, Antonio Cammi
Published: 2025-10-14
Categories: cs.CE, cs.LG
arXiv: https://arxiv.org/abs/2510.12368v2

Shallow Recurrent Decoder networks are a novel data-driven methodology able to provide accurate state estimation in engineering systems, such as nuclear reactors. This deep learning architecture is a robust technique designed to map the temporal trajectories of a few sparse measures to the full state space, including unobservable fields, which is agnostic to sensor positions and able to handle noisy data through an ensemble strategy, leveraging the short training times and without the need for hyperparameter tuning. The architecture was successfully applied to the Molten Salt Fast Reactor concept; now, this work considers the performance of Shallow Recurrent Decoders on a deployed reactor concept. The underlying model is represented by a fluid dynamics model of the TRIGA Mark II research reactor; the architecture will use both synthetic temperature data coming from the numerical model and leveraging experimental temperature data recorded during a previous campaign. The objectives of this work are therefore: 1) presenting the first application of SHRED to a deployed nuclear reactor (TRIGA Mark II); 2) the integration of hybrid synthetic/experimental data within the SHRED framework; 3) a systematic assessment of SHRED robustness under physically constrained, low-dynamics sensor locations; and 4) a quantitative evaluation of SHRED self-correction capability when model-to-data discrepancies are present. Indeed, this approach is capable of accurately reconstruct every field of interest in real-time, using both synthetic (with an average relative error lower than 4\% in euclidean norm) and experimental data (with a RMSE of 1.52 K on the temperature, compared of 1.85 K of the CFD), making it suitable for interpretable monitoring and control purposes in the context of building digital twins for nuclear reactors.

---

### LLM-REVal: Can We Trust LLM Reviewers Yet?

Authors: Rui Li, Jia-Chen Gu, Po-Nien Kung, Heming Xia, Junfeng liu, Xiangwen Kong, Zhifang Sui, Nanyun Peng
Published: 2025-10-14
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.12367v1

The rapid advancement of large language models (LLMs) has inspired researchers to integrate them extensively into the academic workflow, potentially reshaping how research is practiced and reviewed. While previous studies highlight the potential of LLMs in supporting research and peer review, their dual roles in the academic workflow and the complex interplay between research and review bring new risks that remain largely underexplored. In this study, we focus on how the deep integration of LLMs into both peer-review and research processes may influence scholarly fairness, examining the potential risks of using LLMs as reviewers by simulation. This simulation incorporates a research agent, which generates papers and revises, alongside a review agent, which assesses the submissions. Based on the simulation results, we conduct human annotations and identify pronounced misalignment between LLM-based reviews and human judgments: (1) LLM reviewers systematically inflate scores for LLM-authored papers, assigning them markedly higher scores than human-authored ones; (2) LLM reviewers persistently underrate human-authored papers with critical statements (e.g., risk, fairness), even after multiple revisions. Our analysis reveals that these stem from two primary biases in LLM reviewers: a linguistic feature bias favoring LLM-generated writing styles, and an aversion toward critical statements. These results highlight the risks and equity concerns posed to human authors and academic research if LLMs are deployed in the peer review cycle without adequate caution. On the other hand, revisions guided by LLM reviews yield quality gains in both LLM-based and human evaluations, illustrating the potential of the LLMs-as-reviewers for early-stage researchers and enhancing low-quality papers.

---

### (R)evolution of Programming: Vibe Coding as a Post-Coding Paradigm

Authors: Kevin Krings, Nino S. Bohn, Thomas Ludwig
Published: 2025-10-14
Categories: cs.SE, cs.AI, cs.HC
arXiv: https://arxiv.org/abs/2510.12364v2

Recent advancements in generative artificial intelligence (GenAI), particularly large language models, have introduced new possibilities for software development practices. In our paper we investigate the emerging Vibe Coding (VC) paradigm that emphasizes intuitive, affect-driven, and improvisational interactions between developers and AI systems. Building upon the discourse of End-User Development (EUD), we explore how VC diverges from conventional programming approaches such as those supported by tools like GitHub Copilot. Through five semi-structured interview sessions with ten experienced software practitioners, we identify five thematic dimensions: creativity, sustainability, the future of programming, collaboration, and criticism. Our analysis conceptualizes VC within the metaphor of co-drifting, contrasting it with the prevalent co-piloting perspective of AI-assisted development. We argue that VC reconfigures the developers role, blurring boundaries between professional and non-developers. While VC enables novel forms of expression and rapid prototyping, it also introduces challenges regarding reproducibility, scalability, and inclusivity. We propose that VC represents a meaningful shift in programming culture, warranting further investigation within human-computer interaction (HCI) and software engineering research.

---

### Pretraining in Actor-Critic Reinforcement Learning for Locomotion

Authors: Jiale Fan, Andrei Cramariuc, Tifanny Portela, Marco Hutter
Published: 2025-10-14
Categories: cs.RO, cs.LG
arXiv: https://arxiv.org/abs/2510.12363v4

The pretraining-finetuning paradigm has facilitated numerous transformative advancements in artificial intelligence research in recent years. However, in the domain of reinforcement learning (RL) for robot locomotion, individual skills are often learned from scratch despite the high likelihood that some generalizable knowledge is shared across all task-specific policies belonging to the same robot embodiment. This work aims to define a paradigm for pretraining neural network models that encapsulate such knowledge and can subsequently serve as a basis for warm-starting the RL process in classic actor-critic algorithms, such as Proximal Policy Optimization (PPO). We begin with a task-agnostic exploration-based data collection algorithm to gather diverse, dynamic transition data, which is then used to train a Proprioceptive Inverse Dynamics Model (PIDM) through supervised learning. The pretrained weights are then loaded into both the actor and critic networks to warm-start the policy optimization of actual tasks. We systematically validated our proposed method with 9 distinct robot locomotion RL environments comprising 3 different robot embodiments, showing significant benefits of this initialization strategy. Our proposed approach on average improves sample efficiency by 36.2% and task performance by 4.3% compared to random initialization. We further present key ablation studies and empirical analyses that shed light on the mechanisms behind the effectiveness of this method.

---

### CurriFlow: Curriculum-Guided Depth Fusion with Optical Flow-Based Temporal Alignment for 3D Semantic Scene Completion

Authors: Jinzhou Lin, Jie Zhou, Wenhao Xu, Rongtao Xu, Changwei Wang, Shunpeng Chen, Kexue Fu, Yihua Shao, Li Guo, Shibiao Xu
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12362v1

Semantic Scene Completion (SSC) aims to infer complete 3D geometry and semantics from monocular images, serving as a crucial capability for camera-based perception in autonomous driving. However, existing SSC methods relying on temporal stacking or depth projection often lack explicit motion reasoning and struggle with occlusions and noisy depth supervision. We propose CurriFlow, a novel semantic occupancy prediction framework that integrates optical flow-based temporal alignment with curriculum-guided depth fusion. CurriFlow employs a multi-level fusion strategy to align segmentation, visual, and depth features across frames using pre-trained optical flow, thereby improving temporal consistency and dynamic object understanding. To enhance geometric robustness, a curriculum learning mechanism progressively transitions from sparse yet accurate LiDAR depth to dense but noisy stereo depth during training, ensuring stable optimization and seamless adaptation to real-world deployment. Furthermore, semantic priors from the Segment Anything Model (SAM) provide category-agnostic supervision, strengthening voxel-level semantic learning and spatial consistency. Experiments on the SemanticKITTI benchmark demonstrate that CurriFlow achieves state-of-the-art performance with a mean IoU of 16.9, validating the effectiveness of our motion-guided and curriculum-aware design for camera-based 3D semantic scene completion.

---

### MoBiLE: Efficient Mixture-of-Experts Inference on Consumer GPU with Mixture of Big Little Experts

Authors: Yushu Zhao, Yubin Qin, Yang Wang, Xiaolong Yang, Huiming Han, Shaojun Wei, Yang Hu, Shouyi Yin
Published: 2025-10-14
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.12357v1

Mixture-of-Experts (MoE) models have recently demonstrated exceptional performance across a diverse range of applications. The principle of sparse activation in MoE models facilitates an offloading strategy, wherein active experts are maintained in GPU HBM, while inactive experts are stored in CPU DRAM. The efficacy of this approach, however, is fundamentally constrained by the limited bandwidth of the CPU-GPU interconnect. To mitigate this bottleneck, existing approaches have employed prefetching to accelerate MoE inference. These methods attempt to predict and prefetch the required experts using specially trained modules. Nevertheless, such techniques are often encumbered by significant training overhead and have shown diminished effectiveness on recent MoE models with fine-grained expert segmentation. In this paper, we propose MoBiLE, a plug-and-play offloading-based MoE inference framework with \textit{mixture of big-little experts}. It reduces the number of experts for unimportant tokens to half for acceleration while maintaining full experts for important tokens to guarantee model quality. Further, a dedicated fallback and prefetching mechanism is designed for switching between little and big experts to improve memory efficiency. We evaluate MoBiLE on four typical modern MoE architectures and challenging generative tasks. Our results show that MoBiLE achieves a speedup of 1.60x to 1.72x compared to the baseline on a consumer GPU system, with negligible degradation in accuracy.

---

### Fine-grained Analysis of Brain-LLM Alignment through Input Attribution

Authors: Michela Proietti, Roberto Capobianco, Mariya Toneva
Published: 2025-10-14
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.12355v2

Understanding the alignment between large language models (LLMs) and human brain activity can reveal computational principles underlying language processing. We introduce a fine-grained input attribution method to identify the specific words most important for brain-LLM alignment, and leverage it to study a contentious research question about brain-LLM alignment: the relationship between brain alignment (BA) and next-word prediction (NWP). Our findings reveal that BA and NWP rely on largely distinct word subsets: NWP exhibits recency and primacy biases with a focus on syntax, while BA prioritizes semantic and discourse-level information with a more targeted recency effect. This work advances our understanding of how LLMs relate to human language processing and highlights differences in feature reliance between BA and NWP. Beyond this study, our attribution method can be broadly applied to explore the cognitive relevance of model predictions in diverse language processing tasks.

---

### O-Forge: An LLM + Computer Algebra Framework for Asymptotic Analysis

Authors: Ayush Khaitan, Vijay Ganesh
Published: 2025-10-14
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.12350v2

Large language models have recently demonstrated advanced capabilities in solving IMO and Putnam problems; yet their role in research mathematics has remained fairly limited. The key difficulty is verification: suggested proofs may look plausible, but cannot be trusted without rigorous checking. We present a framework, called LLM+CAS, and an associated tool, O-Forge, that couples frontier LLMs with a computer algebra systems (CAS) in an In-Context Symbolic Feedback loop to produce proofs that are both creative and symbolically verified. Our focus is on asymptotic inequalities, a topic that often involves difficult proofs and appropriate decomposition of the domain into the "right" subdomains. Many mathematicians, including Terry Tao, have suggested that using AI tools to find the right decompositions can be very useful for research-level asymptotic analysis. In this paper, we show that our framework LLM+CAS turns out to be remarkably effective at proposing such decompositions via a combination of a frontier LLM and a CAS. More precisely, we use an LLM to suggest domain decomposition, and a CAS (such as Mathematica) that provides a verification of each piece axiomatically. Using this loop, we answer a question posed by Terence Tao: whether LLMs coupled with a verifier can be used to help prove intricate asymptotic inequalities. More broadly, we show how AI can move beyond contest math towards research-level tools for professional mathematicians.

---

### Traveling Salesman-Based Token Ordering Improves Stability in Homomorphically Encrypted Language Models

Authors: Donghwan Rho, Sieun Seo, Hyewon Sung, Chohong Min, Ernest K. Ryu
Published: 2025-10-14
Categories: cs.LG, cs.CR
arXiv: https://arxiv.org/abs/2510.12343v1

As users increasingly interact with large language models (LLMs) using private information, secure and encrypted communication becomes essential. Homomorphic encryption (HE) provides a principled solution by enabling computation directly on encrypted data. Although prior work has explored aspects of running LLMs under HE, the challenge of text generation, particularly next-token prediction, has received limited attention and remains a key obstacle to practical encrypted interaction. In this work, we propose a TSP-based token reordering strategy to address the difficulties of encrypted text generation, together with a post-processing step that further reduces approximation error. Theoretical analysis and experimental results demonstrate that our method prevents collapse, improves coherence in generated text, and preserves data privacy throughout. Overall, our contributions advance the feasibility of practical and privacy-preserving LLM inference.

---

### Sliding-Window Signatures for Time Series: Application to Electricity Demand Forecasting

Authors: Nina Drobac, Margaux Brégère, Joseph de Vilmarest, Olivier Wintenberger
Published: 2025-10-14
Categories: stat.ME, stat.ML
arXiv: https://arxiv.org/abs/2510.12337v1

Nonlinear and delayed effects of covariates often render time series forecasting challenging. To this end, we propose a novel forecasting framework based on ridge regression with signature features calculated on sliding windows. These features capture complex temporal dynamics without relying on learned or hand-crafted representations. Focusing on the discrete-time setting, we establish theoretical guarantees, namely universality of approximation and stationarity of signatures. We introduce an efficient sequential algorithm for computing signatures on sliding windows. The method is evaluated on both synthetic and real electricity demand data. Results show that signature features effectively encode temporal and nonlinear dependencies, yielding accurate forecasts competitive with those based on expert knowledge.

---

### Finite-time Convergence Analysis of Actor-Critic with Evolving Reward

Authors: Rui Hu, Yu Chen, Longbo Huang
Published: 2025-10-14
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.12334v1

Many popular practical reinforcement learning (RL) algorithms employ evolving reward functions-through techniques such as reward shaping, entropy regularization, or curriculum learning-yet their theoretical foundations remain underdeveloped. This paper provides the first finite-time convergence analysis of a single-timescale actor-critic algorithm in the presence of an evolving reward function under Markovian sampling. We consider a setting where the reward parameters may change at each time step, affecting both policy optimization and value estimation. Under standard assumptions, we derive non-asymptotic bounds for both actor and critic errors. Our result shows that an $O(1/\sqrt{T})$ convergence rate is achievable, matching the best-known rate for static rewards, provided the reward parameters evolve slowly enough. This rate is preserved when the reward is updated via a gradient-based rule with bounded gradient and on the same timescale as the actor and critic, offering a theoretical foundation for many popular RL techniques. As a secondary contribution, we introduce a novel analysis of distribution mismatch under Markovian sampling, improving the best-known rate by a factor of $\log^2T$ in the static-reward case.

---

### Leveraging Teleconnections with Physics-Informed Graph Attention Networks for Long-Range Extreme Rainfall Forecasting in Thailand

Authors: Kiattikun Chobtham, Kanoksri Sarinnapakorn, Kritanai Torsri, Prattana Deeprasertkul, Jirawan Kamma
Published: 2025-10-14
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.12328v7

Accurate rainfall forecasting, particularly for extreme events, remains a significant challenge in climatology and the Earth system. This paper presents novel physics-informed Graph Neural Networks (GNNs) combined with extreme-value analysis techniques to improve gauge-station rainfall predictions across Thailand. The model leverages a graph-structured representation of gauge stations to capture complex spatiotemporal patterns, and it offers explainability through teleconnections. We preprocess relevant climate indices that potentially influence regional rainfall. The proposed Graph Attention Network with Long Short-Term Memory (Attention-LSTM) applies the attention mechanism using initial edge features derived from simple orographic-precipitation physics formulation. The embeddings are subsequently processed by LSTM layers. To address extremes, we perform Peak-Over-Threshold (POT) mapping using the novel Spatial Season-aware Generalized Pareto Distribution (GPD) method, which overcomes limitations of traditional machine-learning models. Experiments demonstrate that our method outperforms well-established baselines across most regions, including areas prone to extremes, and remains strongly competitive with the state of the art. Compared with the operational forecasting system SEAS5, our real-world application improves extreme-event prediction and offers a practical enhancement to produce high-resolution maps that support decision-making in long-term water management.

---

### Simple Projection Variants Improve ColBERT Performance

Authors: Benjamin Clavié, Sean Lee, Rikiya Takehi, Aamir Shakir, Makoto P. Kato
Published: 2025-10-14
Categories: cs.IR, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2510.12327v1

Multi-vector dense retrieval methods like ColBERT systematically use a single-layer linear projection to reduce the dimensionality of individual vectors. In this study, we explore the implications of the MaxSim operator on the gradient flows of the training of multi-vector models and show that such a simple linear projection has inherent, if non-critical, limitations in this setting. We then discuss the theoretical improvements that could result from replacing this single-layer projection with well-studied alternative feedforward linear networks (FFN), such as deeper, non-linear FFN blocks, GLU blocks, and skip-connections, could alleviate these limitations. Through the design and systematic evaluation of alternate projection blocks, we show that better-designed final projections positively impact the downstream performance of ColBERT models. We highlight that many projection variants outperform the original linear projections, with the best-performing variants increasing average performance on a range of retrieval benchmarks across domains by over 2 NDCG@10 points. We then conduct further exploration on the individual parameters of these projections block in order to understand what drives this empirical performance, highlighting the particular importance of upscaled intermediate projections and residual connections. As part of these ablation studies, we show that numerous suboptimal projection variants still outperform the traditional single-layer projection across multiple benchmarks, confirming our hypothesis. Finally, we observe that this effect is consistent across random seeds, further confirming that replacing the linear layer of ColBERT models is a robust, drop-in upgrade.

---

### Causal Inspired Multi Modal Recommendation

Authors: Jie Yang, Chenyang Gu, Zixuan Liu
Published: 2025-10-14
Categories: cs.IR, cs.AI
arXiv: https://arxiv.org/abs/2510.12325v1

Multimodal recommender systems enhance personalized recommendations in e-commerce and online advertising by integrating visual, textual, and user-item interaction data. However, existing methods often overlook two critical biases: (i) modal confounding, where latent factors (e.g., brand style or product category) simultaneously drive multiple modalities and influence user preference, leading to spurious feature-preference associations; (ii) interaction bias, where genuine user preferences are mixed with noise from exposure effects and accidental clicks. To address these challenges, we propose a Causal-inspired multimodal Recommendation framework. Specifically, we introduce a dual-channel cross-modal diffusion module to identify hidden modal confounders, utilize back-door adjustment with hierarchical matching and vector-quantized codebooks to block confounding paths, and apply front-door adjustment combined with causal topology reconstruction to build a deconfounded causal subgraph. Extensive experiments on three real-world e-commerce datasets demonstrate that our method significantly outperforms state-of-the-art baselines while maintaining strong interpretability.

---

### RAG-Anything: All-in-One RAG Framework

Authors: Zirui Guo, Xubin Ren, Lingrui Xu, Jiahao Zhang, Chao Huang
Published: 2025-10-14
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.12323v1

Retrieval-Augmented Generation (RAG) has emerged as a fundamental paradigm for expanding Large Language Models beyond their static training limitations. However, a critical misalignment exists between current RAG capabilities and real-world information environments. Modern knowledge repositories are inherently multimodal, containing rich combinations of textual content, visual elements, structured tables, and mathematical expressions. Yet existing RAG frameworks are limited to textual content, creating fundamental gaps when processing multimodal documents. We present RAG-Anything, a unified framework that enables comprehensive knowledge retrieval across all modalities. Our approach reconceptualizes multimodal content as interconnected knowledge entities rather than isolated data types. The framework introduces dual-graph construction to capture both cross-modal relationships and textual semantics within a unified representation. We develop cross-modal hybrid retrieval that combines structural knowledge navigation with semantic matching. This enables effective reasoning over heterogeneous content where relevant evidence spans multiple modalities. RAG-Anything demonstrates superior performance on challenging multimodal benchmarks, achieving significant improvements over state-of-the-art methods. Performance gains become particularly pronounced on long documents where traditional approaches fail. Our framework establishes a new paradigm for multimodal knowledge access, eliminating the architectural fragmentation that constrains current systems. Our framework is open-sourced at: https://github.com/HKUDS/RAG-Anything.

---

### Beating Harmful Stereotypes Through Facts: RAG-based Counter-speech Generation

Authors: Greta Damo, Elena Cabrio, Serena Villata
Published: 2025-10-14
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.12316v1

Counter-speech generation is at the core of many expert activities, such as fact-checking and hate speech, to counter harmful content. Yet, existing work treats counter-speech generation as pure text generation task, mainly based on Large Language Models or NGO experts. These approaches show severe drawbacks due to the limited reliability and coherence in the generated countering text, and in scalability, respectively. To close this gap, we introduce a novel framework to model counter-speech generation as knowledge-wise text generation process. Our framework integrates advanced Retrieval-Augmented Generation (RAG) pipelines to ensure the generation of trustworthy counter-speech for 8 main target groups identified in the hate speech literature, including women, people of colour, persons with disabilities, migrants, Muslims, Jews, LGBT persons, and other. We built a knowledge base over the United Nations Digital Library, EUR-Lex and the EU Agency for Fundamental Rights, comprising a total of 32,792 texts. We use the MultiTarget-CONAN dataset to empirically assess the quality of the generated counter-speech, both through standard metrics (i.e., JudgeLM) and a human evaluation. Results show that our framework outperforms standard LLM baselines and competitive approach, on both assessments. The resulting framework and the knowledge base pave the way for studying trustworthy and sound counter-speech generation, in hate speech and beyond.

---

### Deep SPI: Safe Policy Improvement via World Models

Authors: Florent Delgrange, Raphael Avalos, Willem Röpke
Published: 2025-10-14
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.12312v2

Safe policy improvement (SPI) offers theoretical control over policy updates, yet existing guarantees largely concern offline, tabular reinforcement learning (RL). We study SPI in general online settings, when combined with world model and representation learning. We develop a theoretical framework showing that restricting policy updates to a well-defined neighborhood of the current policy ensures monotonic improvement and convergence. This analysis links transition and reward prediction losses to representation quality, yielding online, "deep" analogues of classical SPI theorems from the offline RL literature. Building on these results, we introduce DeepSPI, a principled on-policy algorithm that couples local transition and reward losses with regularised policy updates. On the ALE-57 benchmark, DeepSPI matches or exceeds strong baselines, including PPO and DeepMDPs, while retaining theoretical guarantees.

---

### Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics

Authors: Joanna Marks, Tim Y. J. Wang, O. Deniz Akyildiz
Published: 2025-10-14
Categories: stat.ML, cs.LG, stat.CO
arXiv: https://arxiv.org/abs/2510.12311v2

We develop interacting particle algorithms for learning latent variable models with energy-based priors. To do so, we leverage recent developments in particle-based methods for solving maximum marginal likelihood estimation (MMLE) problems. Specifically, we provide a continuous-time framework for learning latent energy-based models, by defining stochastic differential equations (SDEs) that provably solve the MMLE problem. We obtain a practical algorithm as a discretisation of these SDEs and provide theoretical guarantees for the convergence of the proposed algorithm. Finally, we empirically validate the effectiveness of our method on synthetic and image datasets and demonstrate that using a particle based approach offers significant improvement in computational efficiency.

---

### DeepTrust: Multi-Step Classification through Dissimilar Adversarial Representations for Robust Android Malware Detection

Authors: Daniel Pulido-Cortázar, Daniel Gibert, Felip Manyà
Published: 2025-10-14
Categories: cs.CR, cs.LG
arXiv: https://arxiv.org/abs/2510.12310v1

Over the last decade, machine learning has been extensively applied to identify malicious Android applications. However, such approaches remain vulnerable against adversarial examples, i.e., examples that are subtly manipulated to fool a machine learning model into making incorrect predictions. This research presents DeepTrust, a novel metaheuristic that arranges flexible classifiers, like deep neural networks, into an ordered sequence where the final decision is made by a single internal model based on conditions activated in cascade. In the Robust Android Malware Detection competition at the 2025 IEEE Conference SaTML, DeepTrust secured the first place and achieved state-of-the-art results, outperforming the next-best competitor by up to 266% under feature-space evasion attacks. This is accomplished while maintaining the highest detection rate on non-adversarial malware and a false positive rate below 1%. The method's efficacy stems from maximizing the divergence of the learned representations among the internal models. By using classifiers inducing fundamentally dissimilar embeddings of the data, the decision space becomes unpredictable for an attacker. This frustrates the iterative perturbation process inherent to evasion attacks, enhancing system robustness without compromising accuracy on clean examples.

---

### Hybrid Gaussian Splatting for Novel Urban View Synthesis

Authors: Mohamed Omran, Farhad Zanjani, Davide Abati, Jens Petersen, Amirhossein Habibian
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12308v1

This paper describes the Qualcomm AI Research solution to the RealADSim-NVS challenge, hosted at the RealADSim Workshop at ICCV 2025. The challenge concerns novel view synthesis in street scenes, and participants are required to generate, starting from car-centric frames captured during some training traversals, renders of the same urban environment as viewed from a different traversal (e.g. different street lane or car direction). Our solution is inspired by hybrid methods in scene generation and generative simulators merging gaussian splatting and diffusion models, and it is composed of two stages: First, we fit a 3D reconstruction of the scene and render novel views as seen from the target cameras. Then, we enhance the resulting frames with a dedicated single-step diffusion model. We discuss specific choices made in the initialization of gaussian primitives as well as the finetuning of the enhancer model and its training data curation. We report the performance of our model design and we ablate its components in terms of novel view quality as measured by PSNR, SSIM and LPIPS. On the public leaderboard reporting test results, our proposal reaches an aggregated score of 0.432, achieving the second place overall.

---

### A large-scale pipeline for LLM-assisted corpus annotation: variation and change in the English consider construction

Authors: Cameron Morin, Matti Marttinen Larsson
Published: 2025-10-14
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.12306v5

As natural language corpora expand at an unprecedented rate, manual annotation remains a significant methodological bottleneck in corpus linguistic work. We address this challenge by presenting a scalable pipeline for automating grammatical annotation in voluminous corpora using large language models (LLMs). Unlike previous supervised and iterative approaches, our method employs a four-phase workflow: prompt engineering, pre-hoc evaluation, automated batch processing, and post-hoc validation. We demonstrate the pipeline's accessibility and effectiveness through a diachronic case study of variation in the English evaluative consider construction (consider X as/to be/Ø Y). We annotate 143,933 'consider' concordance lines from the Corpus of Historical American English (COHA) via the OpenAI API in under 60 hours, achieving 98%+ accuracy on two sophisticated annotation procedures. A Bayesian multinomial GAM fitted to 44,527 true positives of the evaluative construction reveals previously undocumented genre-specific trajectories of change, enabling us to advance new hypotheses about the relationship between register formality and competing pressures of morphosyntactic reduction and enhancement. Our results suggest that LLMs can perform a range of data preparation tasks at scale with minimal human intervention, unlocking substantive research questions previously beyond practical reach, though implementation requires attention to costs, licensing, and other ethical considerations.

---

### General Fourier Feature Physics-Informed Extreme Learning Machine (GFF-PIELM) for High-Frequency PDEs

Authors: Fei Ren, Sifan Wang, Pei-Zhi Zhuang, Hai-Sui Yu, He Yang
Published: 2025-10-14
Categories: cs.LG, cs.NE, physics.comp-ph
arXiv: https://arxiv.org/abs/2510.12293v1

Conventional physics-informed extreme learning machine (PIELM) often faces challenges in solving partial differential equations (PDEs) involving high-frequency and variable-frequency behaviors. To address these challenges, we propose a general Fourier feature physics-informed extreme learning machine (GFF-PIELM). We demonstrate that directly concatenating multiple Fourier feature mappings (FFMs) and an extreme learning machine (ELM) network makes it difficult to determine frequency-related hyperparameters. Fortunately, we find an alternative to establish the GFF-PIELM in three main steps. First, we integrate a variation of FFM into ELM as the Fourier-based activation function, so there is still one hidden layer in the GFF-PIELM framework. Second, we assign a set of frequency coefficients to the hidden neurons, which enables ELM network to capture diverse frequency components of target solutions. Finally, we develop an innovative, straightforward initialization method for these hyperparameters by monitoring the distribution of ELM output weights. GFF-PIELM not only retains the high accuracy, efficiency, and simplicity of the PIELM framework but also inherits the ability of FFMs to effectively handle high-frequency problems. We carry out five case studies with a total of ten numerical examples to highlight the feasibility and validity of the proposed GFF-PIELM, involving high frequency, variable frequency, multi-scale behaviour, irregular boundary and inverse problems. Compared to conventional PIELM, the GFF-PIELM approach significantly improves predictive accuracy without additional cost in training time and architecture complexity. Our results confirm that that PIELM can be extended to solve high-frequency and variable-frequency PDEs with high accuracy, and our initialization strategy may further inspire advances in other physics-informed machine learning (PIML) frameworks.

---

### Vision Language Models Map Logos to Text via Semantic Entanglement in the Visual Projector

Authors: Sifan Li, Hongkai Chen, Yujun Cai, Qingwen Ye, Liyang Chen, Junsong Yuan, Yiwei Wang
Published: 2025-10-14
Categories: cs.CV, cs.CL
arXiv: https://arxiv.org/abs/2510.12287v1

Vision Language Models (VLMs) have achieved impressive progress in multimodal reasoning; yet, they remain vulnerable to hallucinations, where outputs are not grounded in visual evidence. In this paper, we investigate a previously overlooked setting: logo hallucination, where models generate brand names or textual content despite logos containing no visible words. Using curated splits of pure symbols, hybrids, and text-bearing logos, as well as the challenging Hard-60 subset, we systematically measure hallucination across leading VLMs. We further probe robustness through nine structured perturbations and show that hallucinations persist even under strong distortions, with occlusion exposing the sharpest weaknesses. Embedding-level analysis with open-weight LLaVA demonstrates that hallucination is tied to a small subset of projector dimensions, and targeted ablation substantially reduces errors while preserving OCR accuracy. Together, these findings reveal that VLMs often rely on symbolic priors rather than genuine glyph perception, particularly for iconic circular logos, and that projector subspaces play a decisive role in this failure mode. Our work contributes both a novel diagnostic lens and actionable mitigation insights, highlighting projector disentanglement and OCR-guided decoding as promising directions for building more trustworthy multimodal systems.

---

### Chinese ModernBERT with Whole-Word Masking

Authors: Zeyu Zhao, Ningtao Wang, Xing Fu, Yu Cheng
Published: 2025-10-14
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.12285v1

Encoder-only Transformers have advanced along three axes -- architecture, data, and systems -- yielding Pareto gains in accuracy, speed, and memory efficiency. Yet these improvements have not fully transferred to Chinese, where tokenization and morphology differ markedly from English. We introduce Chinese ModernBERT, a from-scratch Chinese encoder that couples: (i) a hardware-aware 32k BPE vocabulary tailored to frequent Chinese affixes/compounds, lowering the embedding budget; (ii) whole-word masking (WWM) with a dynamic masking curriculum (30% -> 15%) to align task difficulty with training progress; (iii) a two-stage pre-training pipeline that extends the native context from 1,024 to 8,192 tokens using RoPE and alternating local/global attention; and (iv) a damped-cosine learning-rate schedule for stable long-horizon optimization. We pre-train on ~1.2T Chinese tokens from CCI3-HQ, CCI4 (Chinese), and Cosmopedia-Chinese. On CLUE, Chinese ModernBERT is competitive with strong Chinese encoders under a unified fine-tuning protocol. Under bf16 it achieves high long-sequence throughput while maintaining strong short-sequence speed, reflecting benefits from budget allocation and attention design. To probe retrieval-oriented quality, we add a small amount of open contrastive data: fine-tuning on SimCLUE (~3M pairs) improves further when adding T2Ranking (~2M), reaching 0.505 (Pearson) / 0.537 (Spearman) on the SimCLUE test set. Under this open-data setting, Chinese ModernBERT surpasses Qwen-0.6B-embedding on SimCLUE, suggesting a clear scaling path for STS with additional curated pairs. We will release tokenizer and weights to facilitate reproducible research.

---

### Dual Learning with Dynamic Knowledge Distillation and Soft Alignment for Partially Relevant Video Retrieval

Authors: Jianfeng Dong, Lei Huang, Daizong Liu, Xianke Chen, Xun Yang, Changting Lin, Xun Wang, Meng Wang
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12283v1

Almost all previous text-to-video retrieval works ideally assume that videos are pre-trimmed with short durations containing solely text-related content. However, in practice, videos are typically untrimmed in long durations with much more complicated background content. Therefore, in this paper, we focus on the more practical yet challenging task of Partially Relevant Video Retrieval (PRVR), which aims to retrieve partially relevant untrimmed videos with the given query. To tackle this task, we propose a novel framework that distills generalization knowledge from a powerful large-scale vision-language pre-trained model and transfers it to a lightweight, task-specific PRVR network. Specifically, we introduce a Dual Learning framework with Dynamic Knowledge Distillation (DL-DKD++), where a large teacher model provides supervision to a compact dual-branch student network. The student model comprises two branches: an inheritance branch that absorbs transferable knowledge from the teacher, and an exploration branch that learns task-specific information from the PRVR dataset to address domain gaps. To further enhance learning, we incorporate a dynamic soft-target construction mechanism. By replacing rigid hard-target supervision with adaptive soft targets that evolve during training, our method enables the model to better capture the fine-grained, partial relevance between videos and queries. Experiment results demonstrate that our proposed model achieves state-of-the-art performance on TVR, ActivityNet, and Charades-STA datasets for PRVR. The code is available at https://github.com/HuiGuanLab/DL-DKD.

---

### PAGS: Priority-Adaptive Gaussian Splatting for Dynamic Driving Scenes

Authors: Ying A, Wenzhang Sun, Chang Zeng, Chunfeng Wang, Hao Li, Jianxun Cui
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12282v1

Reconstructing dynamic 3D urban scenes is crucial for autonomous driving, yet current methods face a stark trade-off between fidelity and computational cost. This inefficiency stems from their semantically agnostic design, which allocates resources uniformly, treating static backgrounds and safety-critical objects with equal importance. To address this, we introduce Priority-Adaptive Gaussian Splatting (PAGS), a framework that injects task-aware semantic priorities directly into the 3D reconstruction and rendering pipeline. PAGS introduces two core contributions: (1) Semantically-Guided Pruning and Regularization strategy, which employs a hybrid importance metric to aggressively simplify non-critical scene elements while preserving fine-grained details on objects vital for navigation. (2) Priority-Driven Rendering pipeline, which employs a priority-based depth pre-pass to aggressively cull occluded primitives and accelerate the final shading computations. Extensive experiments on the Waymo and KITTI datasets demonstrate that PAGS achieves exceptional reconstruction quality, particularly on safety-critical objects, while significantly reducing training time and boosting rendering speeds to over 350 FPS.

---

### Quantum Annealing for Staff Scheduling in Educational Environments

Authors: Alessia Ciacco, Francesca Guerriero, Eneko Osaba
Published: 2025-10-14
Categories: cs.ET, cs.AI
arXiv: https://arxiv.org/abs/2510.12278v2

We address a novel staff allocation problem that arises in the organization of collaborators among multiple school sites and educational levels. The problem emerges from a real case study in a public school in Calabria, Italy, where staff members must be distributed across kindergartens, primary, and secondary schools under constraints of availability, competencies, and fairness. To tackle this problem, we develop an optimization model and investigate a solution approach based on quantum annealing. Our computational experiments on real-world data show that quantum annealing is capable of producing balanced assignments in short runtimes. These results provide evidence of the practical applicability of quantum optimization methods in educational scheduling and, more broadly, in complex resource allocation tasks.

---

### TFGA-Net: Temporal-Frequency Graph Attention Network for Brain-Controlled Speaker Extraction

Authors: Youhao Si, Yuan Liao, Qiushi Han, Yuhang Yang, Rui Dai, Liya Huang
Published: 2025-10-14
Categories: cs.SD, cs.AI
arXiv: https://arxiv.org/abs/2510.12275v1

The rapid development of auditory attention decoding (AAD) based on electroencephalography (EEG) signals offers the possibility EEG-driven target speaker extraction. However, how to effectively utilize the target-speaker common information between EEG and speech remains an unresolved problem. In this paper, we propose a model for brain-controlled speaker extraction, which utilizes the EEG recorded from the listener to extract the target speech. In order to effectively extract information from EEG signals, we derive multi-scale time--frequency features and further incorporate cortical topological structures that are selectively engaged during the task. Moreover, to effectively exploit the non-Euclidean structure of EEG signals and capture their global features, the graph convolutional networks and self-attention mechanism are used in the EEG encoder. In addition, to make full use of the fused EEG and speech feature and preserve global context and capture speech rhythm and prosody, we introduce MossFormer2 which combines MossFormer and RNN-Free Recurrent as separator. Experimental results on both the public Cocktail Party and KUL dataset in this paper show that our TFGA-Net model significantly outper-forms the state-of-the-art method in certain objective evaluation metrics. The source code is available at: https://github.com/LaoDa-X/TFGA-NET.

---

### Multi-Action Self-Improvement for Neural Combinatorial Optimization

Authors: Laurin Luttmann, Lin Xie
Published: 2025-10-14
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.12273v1

Self-improvement has emerged as a state-of-the-art paradigm in Neural Combinatorial Optimization (NCO), where models iteratively refine their policies by generating and imitating high-quality solutions. Despite strong empirical performance, existing methods face key limitations. Training is computationally expensive, as policy updates require sampling numerous candidate solutions per instance to extract a single expert trajectory. More fundamentally, these approaches fail to exploit the structure of combinatorial problems involving the coordination of multiple agents, such as vehicles in min-max routing or machines in scheduling. By supervising on single-action trajectories, they fail to exploit agent-permutation symmetries, where distinct sequences of actions yield identical solutions, hindering generalization and the ability to learn coordinated behavior. We address these challenges by extending self-improvement to operate over joint multi-agent actions. Our model architecture predicts complete agent-task assignments jointly at each decision step. To explicitly leverage symmetries, we employ a set-prediction loss, which supervises the policy on multiple expert assignments for any given state. This approach enhances sample efficiency and the model's ability to learn coordinated behavior. Furthermore, by generating multi-agent actions in parallel, it drastically accelerates the solution generation phase of the self-improvement loop. Empirically, we validate our method on several combinatorial problems, demonstrating consistent improvements in the quality of the final solution and a reduced generation latency compared to standard self-improvement.

---

### Heterogeneous RBCs via Deep Multi-Agent Reinforcement Learning

Authors: Federico Gabriele, Aldo Glielmo, Marco Taboga
Published: 2025-10-14
Categories: cs.MA, cs.LG, econ.TH
arXiv: https://arxiv.org/abs/2510.12272v2

Current macroeconomic models with agent heterogeneity can be broadly divided into two main groups. Heterogeneous-agent general equilibrium (GE) models, such as those based on Heterogeneous Agent New Keynesian (HANK) or Krusell-Smith (KS) approaches, rely on GE and 'rational expectations', somewhat unrealistic assumptions that make the models very computationally cumbersome, which in turn limits the amount of heterogeneity that can be modelled. In contrast, agent-based models (ABMs) can flexibly encompass a large number of arbitrarily heterogeneous agents, but typically require the specification of explicit behavioural rules, which can lead to a lengthy trial-and-error model-development process. To address these limitations, we introduce MARL-BC, a framework that integrates deep multi-agent reinforcement learning (MARL) with real business cycle (RBC) models. We demonstrate that MARL-BC can: (1) recover textbook RBC results when using a single agent; (2) recover the results of the mean-field KS model using a large number of identical agents; and (3) effectively simulate rich heterogeneity among agents, a hard task for traditional GE approaches. Our framework can be thought of as an ABM if used with a variety of heterogeneous interacting agents, and can reproduce GE results in limit cases. As such, it is a step towards a synthesis of these often opposed modelling paradigms.

---

### The Living Forecast: Evolving Day-Ahead Predictions into Intraday Reality

Authors: Kutay Bölat, Peter Palensky, Simon Tindemans
Published: 2025-10-14
Categories: stat.AP, cs.LG
arXiv: https://arxiv.org/abs/2510.12271v2

Accurate intraday forecasts are essential for power system operations, complementing day-ahead forecasts that gradually lose relevance as new information becomes available. This paper introduces a Bayesian updating mechanism that converts fully probabilistic day-ahead forecasts into intraday forecasts without retraining or re-inference. The approach conditions the Gaussian mixture output of a conditional variational autoencoder-based forecaster on observed measurements, yielding an updated distribution for the remaining horizon that preserves its probabilistic structure. This enables consistent point, quantile, and ensemble forecasts while remaining computationally efficient and suitable for real-time applications. Experiments on household electricity consumption and photovoltaic generation datasets demonstrate that the proposed method improves forecast accuracy up to 25% across likelihood-, sample-, quantile-, and point-based metrics. The largest gains occur in time steps with strong temporal correlation to observed data, and the use of pattern dictionary-based covariance structures further enhances performance. The results highlight a theoretically grounded framework for intraday forecasting in modern power systems.

---

### Tensor Logic: The Language of AI

Authors: Pedro Domingos
Published: 2025-10-14
Categories: cs.AI, cs.LG, cs.NE, cs.PL, stat.ML
arXiv: https://arxiv.org/abs/2510.12269v3

Progress in AI is hindered by the lack of a programming language with all the requisite features. Libraries like PyTorch and TensorFlow provide automatic differentiation and efficient GPU implementation, but are additions to Python, which was never intended for AI. Their lack of support for automated reasoning and knowledge acquisition has led to a long and costly series of hacky attempts to tack them on. On the other hand, AI languages like LISP and Prolog lack scalability and support for learning. This paper proposes tensor logic, a language that solves these problems by unifying neural and symbolic AI at a fundamental level. The sole construct in tensor logic is the tensor equation, based on the observation that logical rules and Einstein summation are essentially the same operation, and all else can be reduced to them. I show how to elegantly implement key forms of neural, symbolic and statistical AI in tensor logic, including transformers, formal reasoning, kernel machines and graphical models. Most importantly, tensor logic makes new directions possible, such as sound reasoning in embedding space. This combines the scalability and learnability of neural networks with the reliability and transparency of symbolic reasoning, and is potentially a basis for the wider adoption of AI.

---

### SpineBench: Benchmarking Multimodal LLMs for Spinal Pathology Analysis

Authors: Chenghanyu Zhang, Zekun Li, Peipei Li, Xing Cui, Shuhan Xia, Weixiang Yan, Yiqiao Zhang, Qianyu Zhuang
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12267v1

With the increasing integration of Multimodal Large Language Models (MLLMs) into the medical field, comprehensive evaluation of their performance in various medical domains becomes critical. However, existing benchmarks primarily assess general medical tasks, inadequately capturing performance in nuanced areas like the spine, which relies heavily on visual input. To address this, we introduce SpineBench, a comprehensive Visual Question Answering (VQA) benchmark designed for fine-grained analysis and evaluation of MLLMs in the spinal domain. SpineBench comprises 64,878 QA pairs from 40,263 spine images, covering 11 spinal diseases through two critical clinical tasks: spinal disease diagnosis and spinal lesion localization, both in multiple-choice format. SpineBench is built by integrating and standardizing image-label pairs from open-source spinal disease datasets, and samples challenging hard negative options for each VQA pair based on visual similarity (similar but not the same disease), simulating real-world challenging scenarios. We evaluate 12 leading MLLMs on SpineBench. The results reveal that these models exhibit poor performance in spinal tasks, highlighting limitations of current MLLM in the spine domain and guiding future improvements in spinal medicine applications. SpineBench is publicly available at https://zhangchenghanyu.github.io/SpineBench.github.io/.

---

### HiLoRA: Adaptive Hierarchical LoRA Routing for Training-Free Domain Generalization

Authors: Ziyi Han, Huanyu Wang, Zeyu Zhang, Xiangxiang Dai, Xutong Liu, John C. S. Lui
Published: 2025-10-14
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.12266v1

Low-Rank Adaptation (LoRA) has emerged as a widely used technique for adapting large language models (LLMs) to new domains, due to its modular design and broad availability on platforms such as HuggingFace. This availability has motivated efforts to reuse existing LoRAs for domain generalization. However, existing methods often rely on explicit task labels or additional training, which are impractical for deployment. Moreover, they typically activate a fixed number of entire LoRA modules, leading to parameter redundancy or insufficiency that degrade performance. In this paper, we propose \texttt{HiLoRA}, a training-free framework that performs adaptive hierarchical routing over LoRA pools. Drawing on structural properties of LoRA, we define rank-one components (ROCs), in which each rank parameter is regarded as an independent unit. For a given input sequence, \texttt{HiLoRA} first adaptively selects a subset of LoRAs and determines their ROC allocation based on Gaussian likelihoods at the sequence level. At the token level, it further refines routing by activating only the most informative ROCs. We further provide theoretical guarantees that \texttt{HiLoRA} selects the most relevant LoRAs with high probability. Extensive experiments show that \texttt{HiLoRA} achieves substantial improvements in domain generalization, with accuracy gains of up to {\small $55\%$} over state-of-the-art baselines, while maintaining comparable inference throughput.

---

### Human-in-the-Loop Bandwidth Estimation for Quality of Experience Optimization in Real-Time Video Communication

Authors: Sami Khairy, Gabriel Mittag, Vishak Gopal, Ross Cutler
Published: 2025-10-14
Categories: cs.MM, cs.AI, cs.NI, eess.SY
arXiv: https://arxiv.org/abs/2510.12265v1

The quality of experience (QoE) delivered by video conferencing systems is significantly influenced by accurately estimating the time-varying available bandwidth between the sender and receiver. Bandwidth estimation for real-time communications remains an open challenge due to rapidly evolving network architectures, increasingly complex protocol stacks, and the difficulty of defining QoE metrics that reliably improve user experience. In this work, we propose a deployed, human-in-the-loop, data-driven framework for bandwidth estimation to address these challenges. Our approach begins with training objective QoE reward models derived from subjective user evaluations to measure audio and video quality in real-time video conferencing systems. Subsequently, we collect roughly $1$M network traces with objective QoE rewards from real-world Microsoft Teams calls to curate a bandwidth estimation training dataset. We then introduce a novel distributional offline reinforcement learning (RL) algorithm to train a neural-network-based bandwidth estimator aimed at improving QoE for users. Our real-world A/B test demonstrates that the proposed approach reduces the subjective poor call ratio by $11.41\%$ compared to the baseline bandwidth estimator. Furthermore, the proposed offline RL algorithm is benchmarked on D4RL tasks to demonstrate its generalization beyond bandwidth estimation.

---

### Reducing Belief Deviation in Reinforcement Learning for Active Reasoning

Authors: Deyu Zou, Yongqiang Chen, Jianxiang Wang, Haochen Yang, Mufei Li, James Cheng, Pan Li, Yu Gong
Published: 2025-10-14
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.12264v2

Active reasoning requires large language model (LLM) agents to interact with external sources and strategically gather information to solve problems in multiple turns. Central to this process is belief tracking: maintaining an accurate representation of the underlying state and uncertainty in understanding and solving the problem. However, due to limited reasoning capabilities, LLM-based agents often suffer belief deviation: their internal beliefs drift from the true problem state, leading to loss of state awareness and uninformative or repetitive actions. Once this happens, errors compound in the trajectories used for reinforcement learning (RL), leading to misattributed credits and limited exploration. To address this issue, we propose to track belief deviation and develop $\mathbf{T^3}$, a simple yet principled method that detects excessive deviation and truncates training trajectories to suppress uninformative tail effects. Hence, $\mathbf{T^3}$ preserves credits for informative prefixes and systematically improves policy optimization. Across 5 challenging tasks, $\mathbf{T^3}$ consistently enhances training stability and yields performance gains of up to 30 points while cutting token cost by up to 34%. These results highlight belief control as a key principle for building robust LLM agents capable of active reasoning.

---

### AngularFuse: A Closer Look at Angle-based Perception for Spatial-Sensitive Multi-Modality Image Fusion

Authors: Xiaopeng Liu, Yupei Lin, Sen Zhang, Xiao Wang, Yukai Shi, Liang Lin
Published: 2025-10-14
Categories: cs.CV, cs.LG, eess.IV
arXiv: https://arxiv.org/abs/2510.12260v1

Visible-infrared image fusion is crucial in key applications such as autonomous driving and nighttime surveillance. Its main goal is to integrate multimodal information to produce enhanced images that are better suited for downstream tasks. Although deep learning based fusion methods have made significant progress, mainstream unsupervised approaches still face serious challenges in practical applications. Existing methods mostly rely on manually designed loss functions to guide the fusion process. However, these loss functions have obvious limitations. On one hand, the reference images constructed by existing methods often lack details and have uneven brightness. On the other hand, the widely used gradient losses focus only on gradient magnitude. To address these challenges, this paper proposes an angle-based perception framework for spatial-sensitive image fusion (AngularFuse). At first, we design a cross-modal complementary mask module to force the network to learn complementary information between modalities. Then, a fine-grained reference image synthesis strategy is introduced. By combining Laplacian edge enhancement with adaptive histogram equalization, reference images with richer details and more balanced brightness are generated. Last but not least, we introduce an angle-aware loss, which for the first time constrains both gradient magnitude and direction simultaneously in the gradient domain. AngularFuse ensures that the fused images preserve both texture intensity and correct edge orientation. Comprehensive experiments on the MSRS, RoadScene, and M3FD public datasets show that AngularFuse outperforms existing mainstream methods with clear margin. Visual comparisons further confirm that our method produces sharper and more detailed results in challenging scenes, demonstrating superior fusion capability.

---

### Local Background Features Matter in Out-of-Distribution Detection

Authors: Jinlun Ye, Zhuohao Sun, Yiqiao Qiu, Qiu Li, Zhijun Tan, Ruixuan Wang
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12259v1

Out-of-distribution (OOD) detection is crucial when deploying deep neural networks in the real world to ensure the reliability and safety of their applications. One main challenge in OOD detection is that neural network models often produce overconfident predictions on OOD data. While some methods using auxiliary OOD datasets or generating fake OOD images have shown promising OOD detection performance, they are limited by the high costs of data collection and training. In this study, we propose a novel and effective OOD detection method that utilizes local background features as fake OOD features for model training. Inspired by the observation that OOD images generally share similar background regions with ID images, the background features are extracted from ID images as simulated OOD visual representations during training based on the local invariance of convolution. Through being optimized to reduce the $L_2$-norm of these background features, the neural networks are able to alleviate the overconfidence issue on OOD data. Extensive experiments on multiple standard OOD detection benchmarks confirm the effectiveness of our method and its wide combinatorial compatibility with existing post-hoc methods, with new state-of-the-art performance achieved from our method.

---

### Multiplicative Loss for Enhancing Semantic Segmentation in Medical and Cellular Images

Authors: Yuto Yokoi, Kazuhiro Hotta
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12258v1

We propose two novel loss functions, Multiplicative Loss and Confidence-Adaptive Multiplicative Loss, for semantic segmentation in medical and cellular images. Although Cross Entropy and Dice Loss are widely used, their additive combination is sensitive to hyperparameters and often performs suboptimally, especially with limited data. Medical images suffer from data scarcity due to privacy, ethics, and costly annotations, requiring robust and efficient training objectives. Our Multiplicative Loss combines Cross Entropy and Dice losses multiplicatively, dynamically modulating gradients based on prediction confidence. This reduces penalties for confident correct predictions and amplifies gradients for incorrect overconfident ones, stabilizing optimization. Building on this, Confidence-Adaptive Multiplicative Loss applies a confidence-driven exponential scaling inspired by Focal Loss, integrating predicted probabilities and Dice coefficients to emphasize difficult samples. This enhances learning under extreme data scarcity by strengthening gradients when confidence is low. Experiments on cellular and medical segmentation benchmarks show our framework consistently outperforms tuned additive and existing loss functions, offering a simple, effective, and hyperparameter-free mechanism for robust segmentation under challenging data limitations.

---

### Vectorized Video Representation with Easy Editing via Hierarchical Spatio-Temporally Consistent Proxy Embedding

Authors: Ye Chen, Liming Tan, Yupeng Zhu, Yuanbin Wang, Bingbing Ni
Published: 2025-10-14
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.12256v1

Current video representations heavily rely on unstable and over-grained priors for motion and appearance modelling, \emph{i.e.}, pixel-level matching and tracking. A tracking error of just a few pixels would lead to the collapse of the visual object representation, not to mention occlusions and large motion frequently occurring in videos. To overcome the above mentioned vulnerability, this work proposes spatio-temporally consistent proxy nodes to represent dynamically changing objects/scenes in the video. On the one hand, the hierarchical proxy nodes have the ability to stably express the multi-scale structure of visual objects, so they are not affected by accumulated tracking error, long-term motion, occlusion, and viewpoint variation. On the other hand, the dynamic representation update mechanism of the proxy nodes adequately leverages spatio-temporal priors of the video to mitigate the impact of inaccurate trackers, thereby effectively handling drastic changes in scenes and objects. Additionally, the decoupled encoding manner of the shape and texture representations across different visual objects in the video facilitates controllable and fine-grained appearance editing capability. Extensive experiments demonstrate that the proposed representation achieves high video reconstruction accuracy with fewer parameters and supports complex video processing tasks, including video in-painting and keyframe-based temporally consistent video editing.

---

### Shallow Robustness, Deep Vulnerabilities: Multi-Turn Evaluation of Medical LLMs

Authors: Blazej Manczak, Eric Lin, Francisco Eiras, James O' Neill, Vaikkunth Mugunthan
Published: 2025-10-14
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2510.12255v1

Large language models (LLMs) are rapidly transitioning into medical clinical use, yet their reliability under realistic, multi-turn interactions remains poorly understood. Existing evaluation frameworks typically assess single-turn question answering under idealized conditions, overlooking the complexities of medical consultations where conflicting input, misleading context, and authority influence are common. We introduce MedQA-Followup, a framework for systematically evaluating multi-turn robustness in medical question answering. Our approach distinguishes between shallow robustness (resisting misleading initial context) and deep robustness (maintaining accuracy when answers are challenged across turns), while also introducing an indirect-direct axis that separates contextual framing (indirect) from explicit suggestion (direct). Using controlled interventions on the MedQA dataset, we evaluate five state-of-the-art LLMs and find that while models perform reasonably well under shallow perturbations, they exhibit severe vulnerabilities in multi-turn settings, with accuracy dropping from 91.2% to as low as 13.5% for Claude Sonnet 4. Counterintuitively, indirect, context-based interventions are often more harmful than direct suggestions, yielding larger accuracy drops across models and exposing a significant vulnerability for clinical deployment. Further compounding analyses reveal model differences, with some showing additional performance drops under repeated interventions while others partially recovering or even improving. These findings highlight multi-turn robustness as a critical but underexplored dimension for safe and reliable deployment of medical LLMs.

---

### FedMMKT:Co-Enhancing a Server Text-to-Image Model and Client Task Models in Multi-Modal Federated Learning

Authors: Ningxin He, Yang Liu, Wei Sun, Xiaozhou Ye, Ye Ouyang, Tiegang Gao, Zehui Zhang
Published: 2025-10-14
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.12254v1

Text-to-Image (T2I) models have demonstrated their versatility in a wide range of applications. However, adaptation of T2I models to specialized tasks is often limited by the availability of task-specific data due to privacy concerns. On the other hand, harnessing the power of rich multimodal data from modern mobile systems and IoT infrastructures presents a great opportunity. This paper introduces Federated Multi-modal Knowledge Transfer (FedMMKT), a novel framework that enables co-enhancement of a server T2I model and client task-specific models using decentralized multimodal data without compromising data privacy.

---

### Diffusion Models for Reinforcement Learning: Foundations, Taxonomy, and Development

Authors: Changfu Xu, Jianxiong Guo, Yuzhu Liang, Haiyang Huang, Haodong Zou, Xi Zheng, Shui Yu, Xiaowen Chu, Jiannong Cao, Tian Wang
Published: 2025-10-14
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.12253v1

Diffusion Models (DMs), as a leading class of generative models, offer key advantages for reinforcement learning (RL), including multi-modal expressiveness, stable training, and trajectory-level planning. This survey delivers a comprehensive and up-to-date synthesis of diffusion-based RL. We first provide an overview of RL, highlighting its challenges, and then introduce the fundamental concepts of DMs, investigating how they are integrated into RL frameworks to address key challenges in this research field. We establish a dual-axis taxonomy that organizes the field along two orthogonal dimensions: a function-oriented taxonomy that clarifies the roles DMs play within the RL pipeline, and a technique-oriented taxonomy that situates implementations across online versus offline learning regimes. We also provide a comprehensive examination of this progression from single-agent to multi-agent domains, thereby forming several frameworks for DM-RL integration and highlighting their practical utility. Furthermore, we outline several categories of successful applications of diffusion-based RL across diverse domains, discuss open research issues of current methodologies, and highlight key directions for future research to advance the field. Finally, we summarize the survey to identify promising future development directions. We are actively maintaining a GitHub repository (https://github.com/ChangfuXu/D4RL-FTD) for papers and other related resources to apply DMs for RL.

---

### PromptLocate: Localizing Prompt Injection Attacks

Authors: Yuqi Jia, Yupei Liu, Zedian Shao, Jinyuan Jia, Neil Gong
Published: 2025-10-14
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2510.12252v2

Prompt injection attacks deceive a large language model into completing an attacker-specified task instead of its intended task by contaminating its input data with an injected prompt, which consists of injected instruction(s) and data. Localizing the injected prompt within contaminated data is crucial for post-attack forensic analysis and data recovery. Despite its growing importance, prompt injection localization remains largely unexplored. In this work, we bridge this gap by proposing PromptLocate, the first method for localizing injected prompts. PromptLocate comprises three steps: (1) splitting the contaminated data into semantically coherent segments, (2) identifying segments contaminated by injected instructions, and (3) pinpointing segments contaminated by injected data. We show PromptLocate accurately localizes injected prompts across eight existing and eight adaptive attacks.

---

### DSAS: A Universal Plug-and-Play Framework for Attention Optimization in Multi-Document Question Answering

Authors: Jiakai Li, Rongzheng Wang, Yizhuo Ma, Shuang Liang, Guangchun Luo, Ke Qin
Published: 2025-10-14
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.12251v1

While large language models (LLMs) show considerable promise across various fields, they have notable limitations in handling multi-document question answering (Multi-doc QA) tasks. The first challenge is long-range dependency modeling, where LLMs struggle to focus on key information in long texts, which weakens important semantic connections. Second, most LLMs suffer from the ''lost-in-the-middle'' issue, where they have difficulty processing information in the middle of long inputs. Current solutions either truncate global dependencies or demand costly finetuning, ultimately lacking a universal and simple solution for these challenges. To resolve these limitations, we propose Dual-Stage Adaptive Sharpening (DSAS) containing two modules. (i) The Contextual Gate Weighting (CGW) module alleviates ''lost-in-the-middle'' by assessing paragraph relevance through layer-wise attention tracking and position-aware weighting. (ii) The Reciprocal Attention Suppression (RAS) module enhances focus on critical paragraphs by suppressing information exchange between key and irrelevant texts, thus mitigating the limitations in long-range dependency modeling. Notably, DSAS functions as a plug-and-play solution requiring no architectural modifications or extra training parameters. Extensive experiments on four benchmarks demonstrate DSAS's efficacy across mainstream LLMs (Llama, Qwen, Mistral, and Deepseek), with an average F1-score improvement of 4.2% in Multi-doc QA tasks on Llama-3.1-8B-Instruct and Qwen2.5-14B-Instruct. Ablation studies confirm the essential contributions of both the CGW and RAS modules. In addition, detailed discussions in the Appendix further validate the robustness and scalability of DSAS.

---

### Optimal Regularization for Performative Learning

Authors: Edwige Cyffers, Alireza Mirrokni, Marco Mondelli
Published: 2025-10-14
Categories: cs.LG
arXiv: https://arxiv.org/abs/2510.12249v2

In performative learning, the data distribution reacts to the deployed model - for example, because strategic users adapt their features to game it - which creates a more complex dynamic than in classical supervised learning. One should thus not only optimize the model for the current data but also take into account that the model might steer the distribution in a new direction, without knowing the exact nature of the potential shift. We explore how regularization can help cope with performative effects by studying its impact in high-dimensional ridge regression. We show that, while performative effects worsen the test risk in the population setting, they can be beneficial in the over-parameterized regime where the number of features exceeds the number of samples. We show that the optimal regularization scales with the overall strength of the performative effect, making it possible to set the regularization in anticipation of this effect. We illustrate this finding through empirical evaluations of the optimal regularization parameter on both synthetic and real-world datasets.

---

### PromptFlow: Training Prompts Like Neural Networks

Authors: Jingyi Wang, Hongyuan Zhu, Ye Niu, Yunhui Deng
Published: 2025-10-14
Categories: cs.AI
arXiv: https://arxiv.org/abs/2510.12246v1

Large Language Models (LLMs) have demonstrated profound impact on Natural Language Processing (NLP) tasks. However, their effective deployment across diverse domains often require domain-specific adaptation strategies, as generic models may underperform when faced with specialized data distributions. Recent advances in prompt engineering (PE) offer a promising alternative to extensive retraining by refining input instructions to align LLM outputs with task objectives. This paradigm has emerged as a rapid and versatile approach for model fine-tuning. Despite its potential, manual prompt design remains labor-intensive and heavily depends on specialized expertise, often requiring iterative human effort to achieve optimal formulations. To address this limitation, automated prompt engineering methodologies have been developed to systematically generate task-specific prompts. However, current implementations predominantly employ static update rules and lack mechanisms for dynamic strategy selection, resulting in suboptimal adaptation to varying NLP task requirements. Furthermore, most methods treat and update the whole prompts at each step, without considering editing prompt sections at a finer granularity. At last, in particular, the problem of how to recycle experience in LLM is still underexplored. To this end, we propose the PromptFlow, a modular training framework inspired by TensorFlow, which integrates meta-prompts, operators, optimization, and evaluator. Our framework can be equipped with the latest optimization methods and autonomously explores optimal prompt refinement trajectories through gradient-based meta-learning, requiring minimal task-specific training data. Specifically, we devise a reinforcement learning method to recycle experience for LLM in the PE process. Finally, we conduct extensive experiments on various datasets, and demonstrate the effectiveness of PromptFlow.

---