# Research Papers — 2025-10 · bundle page 42 of 217

> Full-text markdown of papers #2051–2100 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/42
[← Previous bundle](https://aipod.dev/papers/archive/2025-10/page/41.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-10/page/43.md)

### Mind Your Entropy: From Maximum Entropy to Trajectory Entropy-Constrained RL

Authors: Guojian Zhan, Likun Wang, Pengcheng Wang, Feihong Zhang, Jingliang Duan, Masayoshi Tomizuka, Shengbo Eben Li
Published: 2025-10-25
Categories: cs.LG, cs.AI, stat.ML
arXiv: https://arxiv.org/abs/2511.11592v1

Maximum entropy has become a mainstream off-policy reinforcement learning (RL) framework for balancing exploitation and exploration. However, two bottlenecks still limit further performance improvement: (1) non-stationary Q-value estimation caused by jointly injecting entropy and updating its weighting parameter, i.e., temperature; and (2) short-sighted local entropy tuning that adjusts temperature only according to the current single-step entropy, without considering the effect of cumulative entropy over time. In this paper, we extends maximum entropy framework by proposing a trajectory entropy-constrained reinforcement learning (TECRL) framework to address these two challenges. Within this framework, we first separately learn two Q-functions, one associated with reward and the other with entropy, ensuring clean and stable value targets unaffected by temperature updates. Then, the dedicated entropy Q-function, explicitly quantifying the expected cumulative entropy, enables us to enforce a trajectory entropy constraint and consequently control the policy long-term stochasticity. Building on this TECRL framework, we develop a practical off-policy algorithm, DSAC-E, by extending the state-of-the-art distributional soft actor-critic with three refinements (DSAC-T). Empirical results on the OpenAI Gym benchmark demonstrate that our DSAC-E can achieve higher returns and better stability.

---

### Evaluating LLMs' Reasoning Over Ordered Procedural Steps

Authors: Adrita Anika, Md Messal Monem Miah
Published: 2025-10-25
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2511.04688v2

Reasoning over procedural sequences, where the order of steps directly impacts outcomes, is a critical capability for large language models (LLMs). In this work, we study the task of reconstructing globally ordered sequences from shuffled procedural steps, using a curated dataset of food recipes, a domain where correct sequencing is essential for task success. We evaluate several LLMs under zero-shot and few-shot settings and present a comprehensive evaluation framework that adapts established metrics from ranking and sequence alignment. These include Kendall's Tau, Normalized Longest Common Subsequence (NLCS), and Normalized Edit Distance (NED), which capture complementary aspects of ordering quality. Our analysis shows that model performance declines with increasing sequence length, reflecting the added complexity of longer procedures. We also find that greater step displacement in the input, corresponding to more severe shuffling, leads to further degradation. These findings highlight the limitations of current LLMs in procedural reasoning, especially with longer and more disordered inputs.

---

### FPGA-based Lane Detection System incorporating Temperature and Light Control Units

Authors: Ibrahim Qamar, Saber Mahmoud, Seif Megahed, Mohamed Khaled, Saleh Hesham, Ahmed Matar, Saif Gebril, Mervat Mahmoud
Published: 2025-10-25
Categories: cs.CV, eess.IV
arXiv: https://arxiv.org/abs/2510.24778v1

Intelligent vehicles are one of the most important outcomes gained from the world tendency toward automation. Applications of IVs, whether in urban roads or robot tracks, do prioritize lane path detection. This paper proposes an FPGA-based Lane Detector Vehicle LDV architecture that relies on the Sobel algorithm for edge detection. Operating on 416 x 416 images and 150 MHz, the system can generate a valid output every 1.17 ms. The valid output consists of the number of present lanes, the current lane index, as well as its right and left boundaries. Additionally, the automated light and temperature control units in the proposed system enhance its adaptability to the surrounding environmental conditions.

---

### Cross-Enhanced Multimodal Fusion of Eye-Tracking and Facial Features for Alzheimer's Disease Diagnosis

Authors: Yujie Nie, Jianzhang Ni, Yonglong Ye, Yuan-Ting Zhang, Yun Kwok Wing, Xiangqing Xu, Xin Ma, Lizhou Fan
Published: 2025-10-25
Categories: cs.CV, cs.AI, eess.IV
arXiv: https://arxiv.org/abs/2510.24777v1

Accurate diagnosis of Alzheimer's disease (AD) is essential for enabling timely intervention and slowing disease progression. Multimodal diagnostic approaches offer considerable promise by integrating complementary information across behavioral and perceptual domains. Eye-tracking and facial features, in particular, are important indicators of cognitive function, reflecting attentional distribution and neurocognitive state. However, few studies have explored their joint integration for auxiliary AD diagnosis. In this study, we propose a multimodal cross-enhanced fusion framework that synergistically leverages eye-tracking and facial features for AD detection. The framework incorporates two key modules: (a) a Cross-Enhanced Fusion Attention Module (CEFAM), which models inter-modal interactions through cross-attention and global enhancement, and (b) a Direction-Aware Convolution Module (DACM), which captures fine-grained directional facial features via horizontal-vertical receptive fields. Together, these modules enable adaptive and discriminative multimodal representation learning. To support this work, we constructed a synchronized multimodal dataset, including 25 patients with AD and 25 healthy controls (HC), by recording aligned facial video and eye-tracking sequences during a visual memory-search paradigm, providing an ecologically valid resource for evaluating integration strategies. Extensive experiments on this dataset demonstrate that our framework outperforms traditional late fusion and feature concatenation methods, achieving a classification accuracy of 95.11% in distinguishing AD from HC, highlighting superior robustness and diagnostic performance by explicitly modeling inter-modal dependencies and modality-specific contributions.

---

### CFL-SparseMed: Communication-Efficient Federated Learning for Medical Imaging with Top-k Sparse Updates

Authors: Gousia Habib, Aniket Bhardwaj, Ritvik Sharma, Shoeib Amin Banday, Ishfaq Ahmad Malik
Published: 2025-10-25
Categories: eess.IV, cs.CV, cs.DC, cs.LG
arXiv: https://arxiv.org/abs/2510.24776v1

Secure and reliable medical image classification is crucial for effective patient treatment, but centralized models face challenges due to data and privacy concerns. Federated Learning (FL) enables privacy-preserving collaborations but struggles with heterogeneous, non-IID data and high communication costs, especially in large networks. We propose \textbf{CFL-SparseMed}, an FL approach that uses Top-k Sparsification to reduce communication overhead by transmitting only the top k gradients. This unified solution effectively addresses data heterogeneity while maintaining model accuracy. It enhances FL efficiency, preserves privacy, and improves diagnostic accuracy and patient care in non-IID medical imaging settings. The reproducibility source code is available on \href{https://github.com/Aniket2241/APK_contruct}{Github}.

---

### PANORAMA: A Dataset and Benchmarks Capturing Decision Trails and Rationales in Patent Examination

Authors: Hyunseung Lim, Sooyohn Nam, Sungmin Na, Ji Yong Cho, June Yong Yang, Hyungyu Shin, Yoonjoo Lee, Juho Kim, Moontae Lee, Hwajung Hong
Published: 2025-10-25
Categories: cs.CY, cs.CL
arXiv: https://arxiv.org/abs/2510.24774v1

Patent examination remains an ongoing challenge in the NLP literature even after the advent of large language models (LLMs), as it requires an extensive yet nuanced human judgment on whether a submitted claim meets the statutory standards of novelty and non-obviousness against previously granted claims -- prior art -- in expert domains. Previous NLP studies have approached this challenge as a prediction task (e.g., forecasting grant outcomes) with high-level proxies such as similarity metrics or classifiers trained on historical labels. However, this approach often overlooks the step-by-step evaluations that examiners must make with profound information, including rationales for the decisions provided in office actions documents, which also makes it harder to measure the current state of techniques in patent review processes. To fill this gap, we construct PANORAMA, a dataset of 8,143 U.S. patent examination records that preserves the full decision trails, including original applications, all cited references, Non-Final Rejections, and Notices of Allowance. Also, PANORAMA decomposes the trails into sequential benchmarks that emulate patent professionals' patent review processes and allow researchers to examine large language models' capabilities at each step of them. Our findings indicate that, although LLMs are relatively effective at retrieving relevant prior art and pinpointing the pertinent paragraphs, they struggle to assess the novelty and non-obviousness of patent claims. We discuss these results and argue that advancing NLP, including LLMs, in the patent domain requires a deeper understanding of real-world patent examination. Our dataset is openly available at https://huggingface.co/datasets/LG-AI-Research/PANORAMA.

---

### Error Adjustment Based on Spatiotemporal Correlation Fusion for Traffic Forecasting

Authors: Fuqiang Liu, Weiping Ding, Luis Miranda-Moreno, Lijun Sun
Published: 2025-10-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.23656v1

Deep neural networks (DNNs) play a significant role in an increasing body of research on traffic forecasting due to their effectively capturing spatiotemporal patterns embedded in traffic data. A general assumption of training the said forecasting models via mean squared error estimation is that the errors across time steps and spatial positions are uncorrelated. However, this assumption does not really hold because of the autocorrelation caused by both the temporality and spatiality of traffic data. This gap limits the performance of DNN-based forecasting models and is overlooked by current studies. To fill up this gap, this paper proposes Spatiotemporally Autocorrelated Error Adjustment (SAEA), a novel and general framework designed to systematically adjust autocorrelated prediction errors in traffic forecasting. Unlike existing approaches that assume prediction errors follow a random Gaussian noise distribution, SAEA models these errors as a spatiotemporal vector autoregressive (VAR) process to capture their intrinsic dependencies. First, it explicitly captures both spatial and temporal error correlations by a coefficient matrix, which is then embedded into a newly formulated cost function. Second, a structurally sparse regularization is introduced to incorporate prior spatial information, ensuring that the learned coefficient matrix aligns with the inherent road network structure. Finally, an inference process with test-time error adjustment is designed to dynamically refine predictions, mitigating the impact of autocorrelated errors in real-time forecasting. The effectiveness of the proposed approach is verified on different traffic datasets. Results across a wide range of traffic forecasting models show that our method enhances performance in almost all cases.

---

### The Structural Scalpel: Automated Contiguous Layer Pruning for Large Language Models

Authors: Yao Lu, Yuqi Li, Wenbin Xie, Shanqing Yu, Qi Xuan, Zhaowei Zhu, Shiping Wen
Published: 2025-10-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.23652v1

Although large language models (LLMs) have achieved revolutionary breakthroughs in many fields, their large model size and high computational cost pose significant challenges for practical deployment on resource-constrained edge devices. To this end, layer pruning has been proposed to reduce the computational overhead by directly removing redundant layers. However, existing layer pruning methods typically rely on hand-crafted metrics to evaluate and remove individual layers, while ignoring the dependencies between layers. This can disrupt the model's information flow and severely degrade performance. To address these issues, we propose CLP, a novel continuous layer pruning framework that introduces two key innovations: a differentiable concave gate algorithm that automatically identifies the best continuous layer segments for pruning via gradient-based optimization; and a cutoff endpoint tuning strategy that effectively restores model performance by fine-tuning only the layers adjacent to the pruned segments. Extensive experiments across multiple model architectures (including LLaMA2, LLaMA3 and Qwen) and sizes (from $7$B to $70$B parameters) show that CLP significantly outperforms existing state-of-the-art baselines. For example, at a pruning rate of $20\%$, CLP achieves an average performance retention of $95.34\%$ on LLaMA3-70B, outperforming baselines by $4.29\%$-$30.52\%$. Furthermore, CLP can be seamlessly combined with quantization to further compress the model with only a slight performance loss.

---

### Beyond Hidden-Layer Manipulation: Semantically-Aware Logit Interventions for Debiasing LLMs

Authors: Wei Xia
Published: 2025-10-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.23650v1

We proposed Static and Dynamic -- two zero-shot logits-layer debiasing methods. Dynamic reduces bias by up to 70% with minimal fluency loss. Logits intervention outperforms hidden-layer approaches. We show semantic-aware logits intervention is stable and effective for debiasing aligned LLMs.

---

### Efficient Low Rank Attention for Long-Context Inference in Large Language Models

Authors: Tenghui Li, Guoxu Zhou, Xuyang Zhao, Yuning Qiu, Qibin Zhao
Published: 2025-10-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.23649v3

As the length of input text increases, the key-value (KV) cache in LLMs imposes prohibitive GPU memory costs and limits long-context inference on resource constrained devices. Existing approaches, such as KV quantization and pruning, reduce memory usage but suffer from numerical precision loss or suboptimal retention of key-value pairs. In this work, Low Rank Query and Key attention (LRQK) is introduced, a two-stage framework that jointly decomposes full-precision query and key matrices into compact rank-\(r\) factors during the prefill stage, and then employs these low-dimensional projections to compute proxy attention scores in \(\mathcal{O}(lr)\) time at each decode step. By selecting only the top-\(k\) tokens and a small fixed set of recent tokens, LRQK employs a mixed GPU-CPU cache with a hit-and-miss mechanism where only missing full-precision KV pairs are transferred, thereby preserving exact attention outputs while reducing CPU-GPU data movement. Extensive experiments on the RULER and LongBench benchmarks with LLaMA-3-8B and Qwen2.5-7B demonstrate that LRQK matches or surpasses leading sparse-attention methods in long context settings, while delivering significant memory savings with minimal accuracy loss. Our code is available at https://github.com/tenghuilee/LRQK.

---

### RoGBot: Relationship-Oblivious Graph-based Neural Network with Contextual Knowledge for Bot Detection

Authors: Ashutosh Anshul, Mohammad Zia Ur Rehman, Sri Akash Kadali, Nagendra Kumar
Published: 2025-10-25
Categories: cs.SI, cs.AI
arXiv: https://arxiv.org/abs/2510.23648v1

Detecting automated accounts (bots) among genuine users on platforms like Twitter remains a challenging task due to the evolving behaviors and adaptive strategies of such accounts. While recent methods have achieved strong detection performance by combining text, metadata, and user relationship information within graph-based frameworks, many of these models heavily depend on explicit user-user relationship data. This reliance limits their applicability in scenarios where such information is unavailable. To address this limitation, we propose a novel multimodal framework that integrates detailed textual features with enriched user metadata while employing graph-based reasoning without requiring follower-following data. Our method uses transformer-based models (e.g., BERT) to extract deep semantic embeddings from tweets, which are aggregated using max pooling to form comprehensive user-level representations. These are further combined with auxiliary behavioral features and passed through a GraphSAGE model to capture both local and global patterns in user behavior. Experimental results on the Cresci-15, Cresci-17, and PAN 2019 datasets demonstrate the robustness of our approach, achieving accuracies of 99.8%, 99.1%, and 96.8%, respectively, and highlighting its effectiveness against increasingly sophisticated bot strategies.

---

### Derivative-Free Sequential Quadratic Programming for Equality-Constrained Stochastic Optimization

Authors: Sen Na
Published: 2025-10-25
Categories: math.OC, math.NA, math.ST, stat.CO, stat.ML
arXiv: https://arxiv.org/abs/2510.22458v1

We consider solving nonlinear optimization problems with a stochastic objective and deterministic equality constraints, assuming that only zero-order information is available for both the objective and constraints, and that the objective is also subject to random sampling noise. Under this setting, we propose a Derivative-Free Stochastic Sequential Quadratic Programming (DF-SSQP) method. Due to the lack of derivative information, we adopt a simultaneous perturbation stochastic approximation (SPSA) technique to randomly estimate the gradients and Hessians of both the objective and constraints. This approach requires only a dimension-independent number of zero-order evaluations -- as few as eight -- at each iteration step. A key distinction between our derivative-free and existing derivative-based SSQP methods lies in the intricate random bias introduced into the gradient and Hessian estimates of the objective and constraints, brought by stochastic zero-order approximations. To address this issue, we introduce an online debiasing technique based on momentum-style estimators that properly aggregate past gradient and Hessian estimates to reduce stochastic noise, while avoiding excessive memory costs via a moving averaging scheme. Under standard assumptions, we establish the global almost-sure convergence of the proposed DF-SSQP method. Notably, we further complement the global analysis with local convergence guarantees by demonstrating that the rescaled iterates exhibit asymptotic normality, with a limiting covariance matrix resembling the minimax optimal covariance achieved by derivative-based methods, albeit larger due to the absence of derivative information. Our local analysis enables online statistical inference of model parameters leveraging DF-SSQP. Numerical experiments on benchmark nonlinear problems demonstrate both the global and local behavior of DF-SSQP.

---

### Evaluating Multimodal Large Language Models on Core Music Perception Tasks

Authors: Brandon James Carone, Iran R. Roman, Pablo Ripollés
Published: 2025-10-25
Categories: cs.SD, cs.AI, eess.AS
arXiv: https://arxiv.org/abs/2510.22455v1

Multimodal Large Language Models (LLMs) claim "musical understanding" via evaluations that conflate listening with score reading. We benchmark three SOTA LLMs (Gemini 2.5 Pro, Gemini 2.5 Flash, and Qwen2.5-Omni) across three core music skills: Syncopation Scoring, Transposition Detection, and Chord Quality Identification. Moreover, we separate three sources of variability: (i) perceptual limitations (audio vs. MIDI inputs), (ii) exposure to examples (zero- vs. few-shot manipulations), and (iii) reasoning strategies (Standalone, CoT, LogicLM). For the latter we adapt LogicLM, a framework combining LLMs with symbolic solvers to perform structured reasoning, to music. Results reveal a clear perceptual gap: models perform near ceiling on MIDI but show accuracy drops on audio. Reasoning and few-shot prompting offer minimal gains. This is expected for MIDI, where performance reaches saturation, but more surprising for audio, where LogicLM, despite near-perfect MIDI accuracy, remains notably brittle. Among models, Gemini Pro achieves the highest performance across most conditions. Overall, current systems reason well over symbols (MIDI) but do not yet "listen" reliably from audio. Our method and dataset make the perception-reasoning boundary explicit and offer actionable guidance for building robust, audio-first music systems.

---

### SemiETPicker: Fast and Label-Efficient Particle Picking for CryoET Tomography Using Semi-Supervised Learning

Authors: Linhan Wang, Jianwen Dou, Wang Li, Shengkun Wang, Zhiwu Xie, Chang-Tien Lu, Yinlin Chen
Published: 2025-10-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.22454v2

Cryogenic Electron Tomography (CryoET) combined with sub-volume averaging (SVA) is the only imaging modality capable of resolving protein structures inside cells at molecular resolution. Particle picking, the task of localizing and classifying target proteins in 3D CryoET volumes, remains the main bottleneck. Due to the reliance on time-consuming manual labels, the vast reserve of unlabeled tomograms remains underutilized. In this work, we present a fast, label-efficient semi-supervised framework that exploits this untapped data. Our framework consists of two components: (i) an end-to-end heatmap-supervised detection model inspired by keypoint detection, and (ii) a teacher-student co-training mechanism that enhances performance under sparse labeling conditions. Furthermore, we introduce multi-view pseudo-labeling and a CryoET-specific DropBlock augmentation strategy to further boost performance. Extensive evaluations on the large-scale CZII dataset show that our approach improves F1 by 10% over supervised baselines, underscoring the promise of semi-supervised learning for leveraging unlabeled CryoET data.

---

### Confidence Sets for Multidimensional Scaling

Authors: Siddharth Vishwanath, Ery Arias-Castro
Published: 2025-10-25
Categories: math.ST, cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2510.22452v1

We develop a formal statistical framework for classical multidimensional scaling (CMDS) applied to noisy dissimilarity data. We establish distributional convergence results for the embeddings produced by CMDS for various noise models, which enable the construction of \emph{bona~fide} uniform confidence sets for the latent configuration, up to rigid transformations. We further propose bootstrap procedures for constructing these confidence sets and provide theoretical guarantees for their validity. We find that the multiplier bootstrap adapts automatically to heteroscedastic noise such as multiplicative noise, while the empirical bootstrap seems to require homoscedasticity. Either form of bootstrap, when valid, is shown to substantially improve finite-sample accuracy. The empirical performance of the proposed methods is demonstrated through numerical experiments.

---

### GraphTOP: Graph Topology-Oriented Prompting for Graph Neural Networks

Authors: Xingbo Fu, Zhenyu Lei, Zihan Chen, Binchi Zhang, Chuxu Zhang, Jundong Li
Published: 2025-10-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.22451v1

Graph Neural Networks (GNNs) have revolutionized the field of graph learning by learning expressive graph representations from massive graph data. As a common pattern to train powerful GNNs, the "pre-training, adaptation" scheme first pre-trains GNNs over unlabeled graph data and subsequently adapts them to specific downstream tasks. In the adaptation phase, graph prompting is an effective strategy that modifies input graph data with learnable prompts while keeping pre-trained GNN models frozen. Typically, existing graph prompting studies mainly focus on *feature-oriented* methods that apply graph prompts to node features or hidden representations. However, these studies often achieve suboptimal performance, as they consistently overlook the potential of *topology-oriented* prompting, which adapts pre-trained GNNs by modifying the graph topology. In this study, we conduct a pioneering investigation of graph prompting in terms of graph topology. We propose the first **Graph** **T**opology-**O**riented **P**rompting (GraphTOP) framework to effectively adapt pre-trained GNN models for downstream tasks. More specifically, we reformulate topology-oriented prompting as an edge rewiring problem within multi-hop local subgraphs and relax it into the continuous probability space through reparameterization while ensuring tight relaxation and preserving graph sparsity. Extensive experiments on five graph datasets under four pre-training strategies demonstrate that our proposed GraphTOP outshines six baselines on multiple node classification datasets. Our code is available at https://github.com/xbfu/GraphTOP.

---

### SmartMixed: A Two-Phase Training Strategy for Adaptive Activation Function Learning in Neural Networks

Authors: Amin Omidvar
Published: 2025-10-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.22450v4

The choice of activation function plays a critical role in neural networks, yet most architectures still rely on fixed, uniform activation functions across all neurons. We introduce SmartMixed, a novel two-phase training strategy that allows networks to learn optimal per-neuron activation functions while preserving computational efficiency at inference. In the first phase, neurons adaptively select from a pool of candidate activation functions (ReLU, Sigmoid, Tanh, Leaky_ReLU, ELU, SELU) using a differentiable hard mixture mechanism. In the second phase, each neuron's activation function is fixed according to the learned selection, resulting in a computationally efficient network that supports continued training with optimized vectorized operations. We evaluate SmartMixed on the MNIST dataset using feedforward neural networks of different architectures. Our analysis reveals that neurons in different layers exhibit distinct preferences for activation functions, providing insights into the functional diversity within neural architectures. We also demonstrated that SmartMixed effectively trains the network by allowing neurons to select their preferred activation functions, competing against models using a single fixed state-of-the-art activation function.

---

### Benchmarking Egocentric Multimodal Goal Inference for Assistive Wearable Agents

Authors: Vijay Veerabadran, Fanyi Xiao, Nitin Kamra, Pedro Matias, Joy Chen, Caley Drooff, Brett D Roads, Riley Williams, Ethan Henderson, Xuanyi Zhao, Kevin Carlberg, Joseph Tighe et al.
Published: 2025-10-25
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.22443v1

There has been a surge of interest in assistive wearable agents: agents embodied in wearable form factors (e.g., smart glasses) who take assistive actions toward a user's goal/query (e.g. "Where did I leave my keys?"). In this work, we consider the important complementary problem of inferring that goal from multi-modal contextual observations. Solving this "goal inference" problem holds the promise of eliminating the effort needed to interact with such an agent. This work focuses on creating WAGIBench, a strong benchmark to measure progress in solving this problem using vision-language models (VLMs). Given the limited prior work in this area, we collected a novel dataset comprising 29 hours of multimodal data from 348 participants across 3,477 recordings, featuring ground-truth goals alongside accompanying visual, audio, digital, and longitudinal contextual observations. We validate that human performance exceeds model performance, achieving 93% multiple-choice accuracy compared with 84% for the best-performing VLM. Generative benchmark results that evaluate several families of modern vision-language models show that larger models perform significantly better on the task, yet remain far from practical usefulness, as they produce relevant goals only 55% of the time. Through a modality ablation, we show that models benefit from extra information in relevant modalities with minimal performance degradation from irrelevant modalities.

---

### Low-Precision Streaming PCA

Authors: Sanjoy Dasgupta, Syamantak Kumar, Shourya Pandey, Purnamrita Sarkar
Published: 2025-10-25
Categories: cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2510.22440v1

Low-precision streaming PCA estimates the top principal component in a streaming setting under limited precision. We establish an information-theoretic lower bound on the quantization resolution required to achieve a target accuracy for the leading eigenvector. We study Oja's algorithm for streaming PCA under linear and nonlinear stochastic quantization. The quantized variants use unbiased stochastic quantization of the weight vector and the updates. Under mild moment and spectral-gap assumptions on the data distribution, we show that a batched version achieves the lower bound up to logarithmic factors under both schemes. This leads to a nearly dimension-free quantization error in the nonlinear quantization setting. Empirical evaluations on synthetic streams validate our theoretical findings and demonstrate that our low-precision methods closely track the performance of standard Oja's algorithm.

---

### PromptReverb: Multimodal Room Impulse Response Generation Through Latent Rectified Flow Matching

Authors: Ali Vosoughi, Yongyi Zang, Qihui Yang, Nathan Paek, Randal Leistikow, Chenliang Xu
Published: 2025-10-25
Categories: cs.SD, cs.AI
arXiv: https://arxiv.org/abs/2510.22439v2

Room impulse response (RIR) generation remains a critical challenge for creating immersive virtual acoustic environments. Current methods suffer from two fundamental limitations: the scarcity of full-band RIR datasets and the inability of existing models to generate acoustically accurate responses from diverse input modalities. We present PromptReverb, a two-stage generative framework that addresses these challenges. Our approach combines a variational autoencoder that upsamples band-limited RIRs to full-band quality (48 kHz), and a conditional diffusion transformer model based on rectified flow matching that generates RIRs from descriptions in natural language. Empirical evaluation demonstrates that PromptReverb produces RIRs with superior perceptual quality and acoustic accuracy compared to existing methods, achieving 8.8% mean RT60 error compared to -37% for widely used baselines and yielding more realistic room-acoustic parameters. Our method enables practical applications in virtual reality, architectural acoustics, and audio production where flexible, high-quality RIR synthesis is essential.

---

### Modeling Hierarchical Thinking in Large Reasoning Models

Authors: G M Shahariar, Erfan Shayegani, Ali Nazari, Nael Abu-Ghazaleh
Published: 2025-10-25
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2510.22437v2

Large Reasoning Models (LRMs) solve complex tasks by generating long Chain-of-Thought (CoT) sequences; however, the emergent dynamics governing reasoning trajectories are not well understood and can lead to inconsistencies and reasoning pathologies. In this work, we propose to approximate LRM's emerging hierarchical reasoning dynamics as a trajectory within a Finite State Machine (FSM) transitioning among six abstract cognitive states. We demonstrate that these states and transitions can be captured in the latent state of the model. We believe that this representation can have different applications in the interpretability and optimization of LRM models. For example, by analyzing the topology of these transitions, we identify statistical shifts in reasoning strategies that help identify effective reasoning chains from those that fail. To illustrate these potential advantages, we propose Q-Value guided steering, a training-free inference-time control method that treats reasoning as a planning problem. We estimate the long-horizon utility of state transitions and apply sparse, orthogonal activation steering at sentence boundaries to align the CoT generation with optimal reasoning policies. Experiments across four benchmarks (AIME25, MATH-500, GSM8k, and GPQA Diamond) using three state-of-the-art open reasoning models demonstrate that Q-Value steering policy achieves significant performance gains with "surgical" efficiency, often requiring 25 times fewer interventions than greedy and weighted baselines, which suggests that reasoning can be effectively controlled by guiding high-level cognitive dynamics rather than micro-managing token generation. Code is available at: https://github.com/shahariar-shibli/CoT-FSM.

---

### 3D Roadway Scene Object Detection with LIDARs in Snowfall Conditions

Authors: Ghazal Farhani, Taufiq Rahman, Syed Mostaquim Ali, Andrew Liu, Mohamed Zaki, Dominique Charlebois, Benoit Anctil
Published: 2025-10-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.22436v1

Because 3D structure of a roadway environment can be characterized directly by a Light Detection and Ranging (LiDAR) sensors, they can be used to obtain exceptional situational awareness for assitive and autonomous driving systems. Although LiDARs demonstrate good performance in clean and clear weather conditions, their performance significantly deteriorates in adverse weather conditions such as those involving atmospheric precipitation. This may render perception capabilities of autonomous systems that use LiDAR data in learning based models to perform object detection and ranging ineffective. While efforts have been made to enhance the accuracy of these models, the extent of signal degradation under various weather conditions remains largely not quantified. In this study, we focus on the performance of an automotive grade LiDAR in snowy conditions in order to develop a physics-based model that examines failure modes of a LiDAR sensor. Specifically, we investigated how the LiDAR signal attenuates with different snowfall rates and how snow particles near the source serve as small but efficient reflectors. Utilizing our model, we transform data from clear conditions to simulate snowy scenarios, enabling a comparison of our synthetic data with actual snowy conditions. Furthermore, we employ this synthetic data, representative of different snowfall rates, to explore the impact on a pre-trained object detection model, assessing its performance under varying levels of snowfall

---

### Hollywood Town: Long-Video Generation via Cross-Modal Multi-Agent Orchestration

Authors: Zheng Wei, Mingchen Li, Zeqian Zhang, Ruibin Yuan, Pan Hui, Huamin Qu, James Evans, Maneesh Agrawala, Anyi Rao
Published: 2025-10-25
Categories: cs.MA, cs.CV
arXiv: https://arxiv.org/abs/2510.22431v1

Recent advancements in multi-agent systems have demonstrated significant potential for enhancing creative task performance, such as long video generation. This study introduces three innovations to improve multi-agent collaboration. First, we propose OmniAgent, a hierarchical, graph-based multi-agent framework for long video generation that leverages a film-production-inspired architecture to enable modular specialization and scalable inter-agent collaboration. Second, inspired by context engineering, we propose hypergraph nodes that enable temporary group discussions among agents lacking sufficient context, reducing individual memory requirements while ensuring adequate contextual information. Third, we transition from directed acyclic graphs (DAGs) to directed cyclic graphs with limited retries, allowing agents to reflect and refine outputs iteratively, thereby improving earlier stages through feedback from subsequent nodes. These contributions lay the groundwork for developing more robust multi-agent systems in creative tasks.

---

### Reinforcement learning-guided optimization of critical current in high-temperature superconductors

Authors: Mouyang Cheng, Qiwei Wan, Bowen Yu, Eunbi Rha, Michael J Landry, Mingda Li
Published: 2025-10-25
Categories: cond-mat.mtrl-sci, cond-mat.supr-con, cs.LG
arXiv: https://arxiv.org/abs/2510.22424v1

High-temperature superconductors are essential for next-generation energy and quantum technologies, yet their performance is often limited by the critical current density ($J_c$), which is strongly influenced by microstructural defects. Optimizing $J_c$ through defect engineering is challenging due to the complex interplay of defect type, density, and spatial correlation. Here we present an integrated workflow that combines reinforcement learning (RL) with time-dependent Ginzburg-Landau (TDGL) simulations to autonomously identify optimal defect configurations that maximize $J_c$. In our framework, TDGL simulations generate current-voltage characteristics to evaluate $J_c$, which serves as the reward signal that guides the RL agent to iteratively refine defect configurations. We find that the agent discovers optimal defect densities and correlations in two-dimensional thin-film geometries, enhancing vortex pinning and $J_c$ relative to the pristine thin-film, approaching 60\% of theoretical depairing limit with up to 15-fold enhancement compared to random initialization. This RL-driven approach provides a scalable strategy for defect engineering, with broad implications for advancing HTS applications in fusion magnets, particle accelerators, and other high-field technologies.

---

### Group size effects and collective misalignment in LLM multi-agent systems

Authors: Ariel Flint, Luca Maria Aiello, Romualdo Pastor-Satorras, Andrea Baronchelli
Published: 2025-10-25
Categories: cs.MA, cs.AI, cs.CY, physics.soc-ph
arXiv: https://arxiv.org/abs/2510.22422v1

Multi-agent systems of large language models (LLMs) are rapidly expanding across domains, introducing dynamics not captured by single-agent evaluations. Yet, existing work has mostly contrasted the behavior of a single agent with that of a collective of fixed size, leaving open a central question: how does group size shape dynamics? Here, we move beyond this dichotomy and systematically explore outcomes across the full range of group sizes. We focus on multi-agent misalignment, building on recent evidence that interacting LLMs playing a simple coordination game can generate collective biases absent in individual models. First, we show that collective bias is a deeper phenomenon than previously assessed: interaction can amplify individual biases, introduce new ones, or override model-level preferences. Second, we demonstrate that group size affects the dynamics in a non-linear way, revealing model-dependent dynamical regimes. Finally, we develop a mean-field analytical approach and show that, above a critical population size, simulations converge to deterministic predictions that expose the basins of attraction of competing equilibria. These findings establish group size as a key driver of multi-agent dynamics and highlight the need to consider population-level effects when deploying LLM-based systems at scale.

---

### Extragradient Method for $(L_0, L_1)$-Lipschitz Root-finding Problems

Authors: Sayantan Choudhury, Nicolas Loizou
Published: 2025-10-25
Categories: math.OC, cs.LG
arXiv: https://arxiv.org/abs/2510.22421v1

Introduced by Korpelevich in 1976, the extragradient method (EG) has become a cornerstone technique for solving min-max optimization, root-finding problems, and variational inequalities (VIs). Despite its longstanding presence and significant attention within the optimization community, most works focusing on understanding its convergence guarantees assume the strong L-Lipschitz condition. In this work, building on the proposed assumptions by Zhang et al. (2024b) for minimization and Vankov et al.(2024) for VIs, we focus on the more relaxed $α$-symmetric $(L_0, L_1)$-Lipschitz condition. This condition generalizes the standard Lipschitz assumption by allowing the Lipschitz constant to scale with the operator norm, providing a more refined characterization of problem structures in modern machine learning. Under the $α$-symmetric $(L_0, L_1)$-Lipschitz condition, we propose a novel step size strategy for EG to solve root-finding problems and establish sublinear convergence rates for monotone operators and linear convergence rates for strongly monotone operators. Additionally, we prove local convergence guarantees for weak Minty operators. We supplement our analysis with experiments validating our theory and demonstrating the effectiveness and robustness of the proposed step sizes for EG.

---

### A PyTorch Framework for Scalable Non-Crossing Quantile Regression

Authors: Kaihua Chang
Published: 2025-10-25
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2510.22419v2

Quantile regression is fundamental to distributional modeling, yet independent estimation of multiple quantiles frequently produces crossing -- where estimated quantile functions violate monotonicity, implying impossible negative probability densities. While Constrained Joint Quantile Regression (CJQR) elegantly enforces non-crossing by construction, existing formulations via Linear Programming exhibit $O((qn)^3)$ complexity, rendering them impractical for large-scale applications. We present the first scalable solution using PyTorch automatic differentiation: \textbf{CJQR-ALM}, combining the \textbf{Augmented Lagrangian Method} with \textbf{differentiable pinball loss} and \textbf{L-BFGS} optimization. Our approach reduces computational complexity to $O(n)$, achieving near-zero crossing rates on datasets exceeding 70,000 observations within minutes. The differentiable formulation naturally extends to neural network architectures for non-linear conditional quantile estimation. Application to Student Growth Percentile calculations demonstrates practical utility for educational assessment, while simulation studies show negligible accuracy cost (RMSE increase $\approx 2.4$ points) relative to unconstrained estimation -- a favorable trade-off for applications requiring valid probability statements across finance, healthcare, and engineering.

---

### Genetic Optimization of a Software-Defined GNSS Receiver

Authors: Laura Train, Rodrigo Castellanos, Miguel Gómez-López
Published: 2025-10-25
Categories: eess.SP, cs.NE, math.OC
arXiv: https://arxiv.org/abs/2510.22417v1

Commercial off-the-shelf (COTS) Global Navigation Satellite System (GNSS) receivers face significant limitations under high-dynamic conditions, particularly in high-acceleration environments such as those experienced by launch vehicles. These performance degradations, often observed as discontinuities in the navigation solution, arise from the inability of traditional tracking loop bandwidths to cope with rapid variations in synchronization parameters. Software-Defined Radio (SDR) receivers overcome these constraints by enabling flexible reconfiguration of tracking loops; however, manual tuning involves a complex, multidimensional search and seldom ensures optimal performance. This work introduces a genetic algorithm-based optimization framework that autonomously explores the receiver configuration space to determine optimal loop parameters for phase, frequency, and delay tracking. The approach is validated within an SDR environment using realistically simulated GPS L1 signals for three representative dynamic regimes -guided rocket flight, Low Earth Orbit (LEO) satellite, and static receiver-processed with the open-source GNSS-SDR architecture. Results demonstrate that evolutionary optimization enables SDR receivers to maintain robust and accurate Position, Velocity, and Time (PVT) solutions across diverse dynamic conditions. The optimized configurations yielded maximum position and velocity errors of approximately 6 m and 0.08 m/s for the static case, 12 m and 2.5 m/s for the rocket case, and 5 m and 0.2 m/s for the LEO case.

---

### Knowledge-guided Continual Learning for Behavioral Analytics Systems

Authors: Yasas Senarath, Hemant Purohit
Published: 2025-10-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2510.22405v2

User behavior on online platforms is evolving, reflecting real-world changes in how people post, whether it's helpful messages or hate speech. Models that learn to capture this content can experience a decrease in performance over time due to data drift, which can lead to ineffective behavioral analytics systems. However, fine-tuning such a model over time with new data can be detrimental due to catastrophic forgetting. Replay-based approaches in continual learning offer a simple yet efficient method to update such models, minimizing forgetting by maintaining a buffer of important training instances from past learned tasks. However, the main limitation of this approach is the fixed size of the buffer. External knowledge bases can be utilized to overcome this limitation through data augmentation. We propose a novel augmentation-based approach to incorporate external knowledge in the replay-based continual learning framework. We evaluate several strategies with three datasets from prior studies related to deviant behavior classification to assess the integration of external knowledge in continual learning and demonstrate that augmentation helps outperform baseline replay-based approaches.

---

### NetBurst: Event-Centric Forecasting of Bursty, Intermittent Time Series

Authors: Satyandra Guthula, Jaber Daneshamooz, Charles Fleming, Kesheng Wu, Walter Willinger, Arpit Gupta
Published: 2025-10-25
Categories: cs.NI, cs.LG
arXiv: https://arxiv.org/abs/2510.22397v2

Network operators monitor their infrastructure by collecting telemetry data such as packet counts, byte rates, or flow volumes, yet answering the questions that effective operations demand -- forecasting future load, diagnosing and characterizing anomalies, and searching for and retrieving historical precedents -- requires more than raw measurements. Bridging this gap calls for learned representations: compact per-entity summaries that capture temporal dynamics from each entity's univariate time series. Time-series foundation models are the natural starting point, but they are designed for dense, periodic benchmark datasets -- the \emph{mild} statistical regime. However, network telemetry data inhabits the \emph{wild} regime: operationally relevant events are rare, separated by variable-length stretches of low or no activity (``ebbs''), with intermittent bursts of heavy-tailed extremes (``tides''). We present NetBurst, an event-centric pipeline that collapses ebbs, separates each time series into a stream of burst timings and a stream of burst magnitudes, and learns a single representation serving all three operational tasks. Compared to the strongest competitors among eight baselines -- including Amazon's Chronos-2 and Datadog's Toto -- and across nine production telemetry configurations, NetBurst reduces median forecasting error by $1.3$--$116\times$ on wild-regime data with a $1.0$--$7.5\times$ better match to the true burst distribution, and matches baselines on mild-regime benchmarks. For characterizing anomalies, NetBurst produces balanced, well-spread clusters that are $16\times$ more describable in operator-familiar terms under a novel interpretability score, and cluster-filtered search delivers $7.5\times$ faster end-to-end retrieval.

---

### Confabulations from ACL Publications (CAP): A Dataset for Scientific Hallucination Detection

Authors: Federica Gamba, Aman Sinha, Timothee Mickus, Raul Vazquez, Patanjali Bhamidipati, Claudio Savelli, Ahana Chattopadhyay, Laura A. Zanella, Yash Kankanampati, Binesh Arakkal Remesh, Aryan Ashok Chandramania, Rohit Agarwal et al.
Published: 2025-10-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.22395v1

We introduce the CAP (Confabulations from ACL Publications) dataset, a multilingual resource for studying hallucinations in large language models (LLMs) within scientific text generation. CAP focuses on the scientific domain, where hallucinations can distort factual knowledge, as they frequently do. In this domain, however, the presence of specialized terminology, statistical reasoning, and context-dependent interpretations further exacerbates these distortions, particularly given LLMs' lack of true comprehension, limited contextual understanding, and bias toward surface-level generalization. CAP operates in a cross-lingual setting covering five high-resource languages (English, French, Hindi, Italian, and Spanish) and four low-resource languages (Bengali, Gujarati, Malayalam, and Telugu). The dataset comprises 900 curated scientific questions and over 7000 LLM-generated answers from 16 publicly available models, provided as question-answer pairs along with token sequences and corresponding logits. Each instance is annotated with a binary label indicating the presence of a scientific hallucination, denoted as a factuality error, and a fluency label, capturing issues in the linguistic quality or naturalness of the text. CAP is publicly released to facilitate advanced research on hallucination detection, multilingual evaluation of LLMs, and the development of more reliable scientific NLP systems.

---

### Top-Down Semantic Refinement for Image Captioning

Authors: Jusheng Zhang, Kaitong Cai, Jing Yang, Jian Wang, Chengpei Tang, Keze Wang
Published: 2025-10-25
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2510.22391v2

Large Vision-Language Models (VLMs) face an inherent contradiction in image captioning: their powerful single-step generation capabilities often lead to a myopic decision-making process. This makes it difficult to maintain global narrative coherence while capturing rich details, a limitation that is particularly pronounced in tasks that require multi-step and complex scene description. To overcome this fundamental challenge, we redefine image captioning as a goal-oriented hierarchical refinement planning problem, and further propose a novel framework, named Top-Down Semantic Refinement (TDSR), which models the generation process as a Markov Decision Process (MDP). However, planning within the vast state space of a VLM presents a significant computational hurdle. Our core contribution, therefore, is the design of a highly efficient Monte Carlo Tree Search (MCTS) algorithm tailored for VLMs. By incorporating a visual-guided parallel expansion and a lightweight value network, our TDSR reduces the call frequency to the expensive VLM by an order of magnitude without sacrificing planning quality. Furthermore, an adaptive early stopping mechanism dynamically matches computational overhead to the image's complexity. Extensive experiments on multiple benchmarks, including DetailCaps, COMPOSITIONCAP, and POPE, demonstrate that our TDSR, as a plug-and-play module, can significantly enhance the performance of existing VLMs (e.g., LLaVA-1.5, Qwen2.5-VL) by achieving state-of-the-art or highly competitive results in fine-grained description, compositional generalization, and hallucination suppression.

---

### A Fully Interpretable Statistical Approach for Roadside LiDAR Background Subtraction

Authors: Aitor Iglesias, Nerea Aranjuelo, Patricia Javierre, Ainhoa Menendez, Ignacio Arganda-Carreras, Marcos Nieto
Published: 2025-10-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.22390v2

We present a fully interpretable and flexible statistical method for background subtraction in roadside LiDAR data, aimed at enhancing infrastructure-based perception in automated driving. Our approach introduces both a Gaussian distribution grid (GDG), which models the spatial statistics of the background using background-only scans, and a filtering algorithm that uses this representation to classify LiDAR points as foreground or background. The method supports diverse LiDAR types, including multiline 360 degree and micro-electro-mechanical systems (MEMS) sensors, and adapts to various configurations. Evaluated on the publicly available RCooper dataset, it outperforms state-of-the-art techniques in accuracy and flexibility, even with minimal background data. Its efficient implementation ensures reliable performance on low-resource hardware, enabling scalable real-world deployment.

---

### Can Small and Reasoning Large Language Models Score Journal Articles for Research Quality and Do Averaging and Few-shot Help?

Authors: Mike Thelwall, Ehsan Mohammadi
Published: 2025-10-25
Categories: cs.DL, cs.AI
arXiv: https://arxiv.org/abs/2510.22389v2

Previous research has shown that journal article quality ratings from the cloud based Large Language Model (LLM) families ChatGPT and Gemini and the medium sized open weights LLM Gemma3 27b correlate moderately with expert research quality scores. This article assesses whether other medium sized LLMs, smaller LLMs, and reasoning models have similar abilities. This is tested with Gemma3 variants, Llama4 Scout, Qwen3, Magistral Small and DeepSeek R1 on a dataset of 2,780 medical, health and life science papers in 6 fields, with two different gold standards, one novel. Few-shot and score averaging approaches are also evaluated. The results suggest that medium-sized LLMs have similar performance to ChatGPT 4o-mini and Gemini 2.0 Flash, but that 1b parameters may often, and 4b sometimes, be too few. Reasoning models did not have a clear advantage. Moreover, averaging scores from multiple identical queries seems to be a universally successful strategy, and there is weak evidence that few-shot prompts (four examples) tend to help. Overall, the results show, for the first time, that smaller LLMs >4b have a substantial capability to rate journal articles for research quality, especially if score averaging is used, but that reasoning does not give an advantage for this task; it is therefore not recommended because it is slow. The use of LLMs to support research evaluation is now more credible since multiple variants have a similar ability, including many that can be deployed offline in a secure environment without substantial computing resources.

---

### Privacy-Aware Federated nnU-Net for ECG Page Digitization

Authors: Nader Nemati
Published: 2025-10-25
Categories: cs.CR, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.22387v1

Deep neural networks can convert ECG page images into analyzable waveforms, yet centralized training often conflicts with cross-institutional privacy and deployment constraints. A cross-silo federated digitization framework is presented that trains a full-model nnU-Net segmentation backbone without sharing images and aggregates updates across sites under realistic non-IID heterogeneity (layout, grid style, scanner profile, noise). The protocol integrates three standard server-side aggregators--FedAvg, FedProx, and FedAdam--and couples secure aggregation with central, user-level differential privacy to align utility with formal guarantees. Key features include: (i) end-to-end full-model training and synchronization across clients; (ii) secure aggregation so the server only observes a clipped, weighted sum once a participation threshold is met; (iii) central Gaussian DP with Renyi accounting applied post-aggregation for auditable user-level privacy; and (iv) a calibration-aware digitization pipeline comprising page normalization, trace segmentation, grid-leakage suppression, and vectorization to twelve-lead signals. Experiments on ECG pages rendered from PTB-XL show consistently faster convergence and higher late-round plateaus with adaptive server updates (FedAdam) relative to FedAvg and FedProx, while approaching centralized performance. The privacy mechanism maintains competitive accuracy while preventing exposure of raw images or per-client updates, yielding deployable, auditable guarantees suitable for multi-institution settings.

---

### Dynamic Dropout: Leveraging Conway's Game of Life for Neural Networks Regularization

Authors: David Freire-Obregón, José Salas-Cáceres, Modesto Castrillón-Santana
Published: 2025-10-25
Categories: cs.LG, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2510.22383v1

Regularization techniques play a crucial role in preventing overfitting and improving the generalization performance of neural networks. Dropout, a widely used regularization technique, randomly deactivates units during training to introduce redundancy and prevent co-adaptation among neurons. Despite its effectiveness, dropout has limitations, such as its static nature and lack of interpretability. In this paper, we propose a novel approach to regularization by substituting dropout with Conway's Game of Life (GoL), a cellular automata with simple rules that govern the evolution of a grid of cells. We introduce dynamic unit deactivation during training by representing neural network units as cells in a GoL grid and applying the game's rules to deactivate units. This approach allows for the emergence of spatial patterns that adapt to the training data, potentially enhancing the network's ability to generalize. We demonstrate the effectiveness of our approach on the CIFAR-10 dataset, showing that dynamic unit deactivation using GoL achieves comparable performance to traditional dropout techniques while offering insights into the network's behavior through the visualization of evolving patterns. Furthermore, our discussion highlights the applicability of our proposal in deeper architectures, demonstrating how it enhances the performance of different dropout techniques.

---

### Efficient Large-Deformation Medical Image Registration via Recurrent Dynamic Correlation

Authors: Tianran Li, Marius Staring, Yuchuan Qiao
Published: 2025-10-25
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2510.22380v1

Deformable image registration estimates voxel-wise correspondences between images through spatial transformations, and plays a key role in medical imaging. While deep learning methods have significantly reduced runtime, efficiently handling large deformations remains a challenging task. Convolutional networks aggregate local features but lack direct modeling of voxel correspondences, promoting recent works to explore explicit feature matching. Among them, voxel-to-region matching is more efficient for direct correspondence modeling by computing local correlation features whithin neighbourhoods, while region-to-region matching incurs higher redundancy due to excessive correlation pairs across large regions. However, the inherent locality of voxel-to-region matching hinders the capture of long-range correspondences required for large deformations. To address this, we propose a Recurrent Correlation-based framework that dynamically relocates the matching region toward more promising positions. At each step, local matching is performed with low cost, and the estimated offset guides the next search region, supporting efficient convergence toward large deformations. In addition, we uses a lightweight recurrent update module with memory capacity and decouples motion-related and texture features to suppress semantic redundancy. We conduct extensive experiments on brain MRI and abdominal CT datasets under two settings: with and without affine pre-registration. Results show that our method exibits a strong accuracy-computation trade-off, surpassing or matching the state-of-the-art performance. For example, it achieves comparable performance on the non-affine OASIS dataset, while using only 9.5% of the FLOPs and running 96% faster than RDP, a representative high-performing method.

---

### TraceTrans: Translation and Spatial Tracing for Surgical Prediction

Authors: Xiyu Luo, Haodong Li, Xinxing Cheng, He Zhao, Yang Hu, Xuan Song, Tianyang Zhang
Published: 2025-10-25
Categories: eess.IV, cs.AI, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2510.22379v4

Image-to-image translation models have achieved notable success in converting images across visual domains and are increasingly used for medical tasks such as predicting post-operative outcomes and modeling disease progression. However, most existing methods primarily aim to match the target distribution and often neglect spatial correspondences between the source and translated images. This limitation can lead to structural inconsistencies and hallucinations, undermining the reliability and interpretability of the predictions. These challenges are accentuated in clinical applications by the stringent requirement for anatomical accuracy. In this work, we present TraceTrans, a novel deformable image translation model designed for post-operative prediction that generates images aligned with the target distribution while explicitly revealing spatial correspondences with the pre-operative input. The framework employs an encoder for feature extraction and dual decoders for predicting spatial deformations and synthesizing the translated image. The predicted deformation field imposes spatial constraints on the generated output, ensuring anatomical consistency with the source. Extensive experiments on medical cosmetology and brain MRI datasets demonstrate that TraceTrans delivers accurate and interpretable post-operative predictions, highlighting its potential for reliable clinical deployment.

---

### Label Smoothing Improves Gradient Ascent in LLM Unlearning

Authors: Zirui Pang, Hao Zheng, Zhijie Deng, Ling Li, Zixin Zhong, Jiaheng Wei
Published: 2025-10-25
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2510.22376v1

LLM unlearning has emerged as a promising approach, aiming to enable models to forget hazardous/undesired knowledge at low cost while preserving as much model utility as possible. Among existing techniques, the most straightforward method is performing Gradient Ascent (GA) w.r.t. the forget data, thereby forcing the model to unlearn the forget dataset. However, GA suffers from severe instability, as it drives updates in a divergent direction, often resulting in drastically degraded model utility. To address this issue, we propose Smoothed Gradient Ascent (SGA). SGA combines the forget data with multiple constructed normal data through a tunable smoothing rate. Intuitively, this extends GA from learning solely on the forget data to jointly learning across both forget and normal data, enabling more stable unlearning while better preserving model utility. Theoretically, we provide the theoretical guidance on the selection of the optimal smoothing rate. Empirically, we evaluate SGA on three benchmarks: TOFU, Harry Potter, and MUSE-NEWS. Experimental results demonstrate that SGA consistently outperforms the original Gradient Ascent (GA) method across all metrics and achieves top-2 performance among all baseline methods on several key metrics.

---

### Conformalized Polynomial Chaos Expansion for Uncertainty-aware Surrogate Modeling

Authors: Dimitrios Loukrezis, Dimitris G. Giovanis
Published: 2025-10-25
Categories: stat.ME, stat.CO, stat.ML
arXiv: https://arxiv.org/abs/2510.22375v3

This work introduces a method to equip data-driven polynomial chaos expansion surrogate models with intervals that quantify the predictive uncertainty of the surrogate. To that end, jackknife-based conformal prediction is integrated into regression-based polynomial chaos expansions. The jackknife algorithm uses leave-one-out residuals to generate predictive intervals around the predictions of the polynomial chaos surrogate. The jackknife+ extension additionally requires leave-one-out model predictions. Both methods allow to use the entire dataset for model training and do not require a hold-out dataset for prediction interval calibration. The key to efficient implementation is to leverage the linearity of the polynomial chaos regression model, so that leave-one-out residuals and, if necessary, leave-one-out model predictions can be computed with analytical, closed-form expressions. This eliminates the need for repeated model re-training. The conformalized polynomial chaos expansion method is first validated on four benchmark models and then applied to two electrical engineering design use-cases. The method produces predictive intervals that provide the target coverages, even for low-accuracy models trained with small datasets. At the same time, training data availability plays a crucial role in improving the empirical coverage and narrowing the predictive interval, as well as in reducing their variability over different training datasets.

---

### VisJudge-Bench: Aesthetics and Quality Assessment of Visualizations

Authors: Yupeng Xie, Zhiyang Zhang, Yifan Wu, Sirong Lu, Jiayi Zhang, Zhaoyang Yu, Jinlin Wang, Sirui Hong, Bang Liu, Chenglin Wu, Yuyu Luo
Published: 2025-10-25
Categories: cs.CL, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2510.22373v3

Visualization, a domain-specific yet widely used form of imagery, is an effective way to turn complex datasets into intuitive insights, and its value depends on whether data are faithfully represented, clearly communicated, and aesthetically designed. However, evaluating visualization quality is challenging: unlike natural images, it requires simultaneous judgment across data encoding accuracy, information expressiveness, and visual aesthetics. Although multimodal large language models (MLLMs) have shown promising performance in aesthetic assessment of natural images, no systematic benchmark exists for measuring their capabilities in evaluating visualizations. To address this, we propose VisJudge-Bench, the first comprehensive benchmark for evaluating MLLMs' performance in assessing visualization aesthetics and quality. It contains 3,090 expert-annotated samples from real-world scenarios, covering single visualizations, multiple visualizations, and dashboards across 32 chart types. Systematic testing on this benchmark reveals that even the most advanced MLLMs (such as GPT-5) still exhibit significant gaps compared to human experts in judgment, with a Mean Absolute Error (MAE) of 0.553 and a correlation with human ratings of only 0.428. To address this issue, we propose VisJudge, a model specifically designed for visualization aesthetics and quality assessment. Experimental results demonstrate that VisJudge significantly narrows the gap with human judgment, reducing the MAE to 0.421 (a 23.9% reduction) and increasing the consistency with human experts to 0.687 (a 60.5% improvement) compared to GPT-5. The benchmark is available at https://github.com/HKUSTDial/VisJudgeBench.

---

### Reasoning Models Reason Well, Until They Don't

Authors: Revanth Rameshkumar, Jimson Huang, Yunxin Sun, Fei Xia, Abulhair Saparov
Published: 2025-10-25
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2510.22371v1

Large language models (LLMs) have shown significant progress in reasoning tasks. However, recent studies show that transformers and LLMs fail catastrophically once reasoning problems exceed modest complexity. We revisit these findings through the lens of large reasoning models (LRMs) -- LLMs fine-tuned with incentives for step-by-step argumentation and self-verification. LRM performance on graph and reasoning benchmarks such as NLGraph seem extraordinary, with some even claiming they are capable of generalized reasoning and innovation in reasoning-intensive fields such as mathematics, physics, medicine, and law. However, by more carefully scaling the complexity of reasoning problems, we show existing benchmarks actually have limited complexity. We develop a new dataset, the Deep Reasoning Dataset (DeepRD), along with a generative process for producing unlimited examples of scalable complexity. We use this dataset to evaluate model performance on graph connectivity and natural language proof planning. We find that the performance of LRMs drop abruptly at sufficient complexity and do not generalize. We also relate our LRM results to the distributions of the complexities of large, real-world knowledge graphs, interaction graphs, and proof datasets. We find the majority of real-world examples fall inside the LRMs' success regime, yet the long tails expose substantial failure potential. Our analysis highlights the near-term utility of LRMs while underscoring the need for new methods that generalize beyond the complexity of examples in the training distribution.

---

### BLIP-FusePPO: A Vision-Language Deep Reinforcement Learning Framework for Lane Keeping in Autonomous Vehicles

Authors: Seyed Ahmad Hosseini Miangoleh, Amin Jalal Aghdasian, Farzaneh Abdollahi
Published: 2025-10-25
Categories: cs.RO, cs.AI, cs.CV, cs.LG, cs.SE
arXiv: https://arxiv.org/abs/2510.22370v1

In this paper, we propose Bootstrapped Language-Image Pretraining-driven Fused State Representation in Proximal Policy Optimization (BLIP-FusePPO), a novel multimodal reinforcement learning (RL) framework for autonomous lane-keeping (LK), in which semantic embeddings generated by a vision-language model (VLM) are directly fused with geometric states, LiDAR observations, and Proportional-Integral-Derivative-based (PID) control feedback within the agent observation space. The proposed method lets the agent learn driving rules that are aware of their surroundings and easy to understand by combining high-level scene understanding from the VLM with low-level control and spatial signals. Our architecture brings together semantic, geometric, and control-aware representations to make policy learning more robust. A hybrid reward function that includes semantic alignment, LK accuracy, obstacle avoidance, and speed regulation helps learning to be more efficient and generalizable. Our method is different from the approaches that only use semantic models to shape rewards. Instead, it directly embeds semantic features into the state representation. This cuts down on expensive runtime inference and makes sure that semantic guidance is always available. The simulation results show that the proposed model is better at LK stability and adaptability than the best vision-based and multimodal RL baselines in a wide range of difficult driving situations. We make our code publicly available.

---

### GigaEmbeddings: Efficient Russian Language Embedding Model

Authors: Egor Kolodin, Daria Khomich, Nikita Savushkin, Anastasia Ianina, Fyodor Minkin
Published: 2025-10-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.22369v1

We introduce GigaEmbeddings, a novel framework for training high-performance Russian-focused text embeddings through hierarchical instruction tuning of the decoder-only LLM designed specifically for Russian language (GigaChat-3B). Our three-stage pipeline, comprising large-scale contrastive pre-training in web-scale corpora, fine-tuning with hard negatives, and multitask generalization across retrieval, classification, and clustering tasks, addresses key limitations of existing methods by unifying diverse objectives and leveraging synthetic data generation. Architectural innovations include bidirectional attention for contextual modeling, latent attention pooling for robust sequence aggregation, and strategic pruning of 25% of transformer layers to enhance efficiency without compromising performance. Evaluated on the ruMTEB benchmark spanning 23 multilingual tasks, GigaEmbeddings achieves state-of-the-art results (69.1 avg. score), outperforming strong baselines with a larger number of parameters.

---

### T2SMark: Balancing Robustness and Diversity in Noise-as-Watermark for Diffusion Models

Authors: Jindong Yang, Han Fang, Weiming Zhang, Nenghai Yu, Kejiang Chen
Published: 2025-10-25
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2510.22366v1

Diffusion models have advanced rapidly in recent years, producing high-fidelity images while raising concerns about intellectual property protection and the misuse of generative AI. Image watermarking for diffusion models, particularly Noise-as-Watermark (NaW) methods, encode watermark as specific standard Gaussian noise vector for image generation, embedding the infomation seamlessly while maintaining image quality. For detection, the generation process is inverted to recover the initial noise vector containing the watermark before extraction. However, existing NaW methods struggle to balance watermark robustness with generation diversity. Some methods achieve strong robustness by heavily constraining initial noise sampling, which degrades user experience, while others preserve diversity but prove too fragile for real-world deployment. To address this issue, we propose T2SMark, a two-stage watermarking scheme based on Tail-Truncated Sampling (TTS). Unlike prior methods that simply map bits to positive or negative values, TTS enhances robustness by embedding bits exclusively in the reliable tail regions while randomly sampling the central zone to preserve the latent distribution. Our two-stage framework then ensures sampling diversity by integrating a randomly generated session key into both encryption pipelines. We evaluate T2SMark on diffusion models with both U-Net and DiT backbones. Extensive experiments show that it achieves an optimal balance between robustness and diversity. Our code is available at \href{https://github.com/0xD009/T2SMark}{https://github.com/0xD009/T2SMark}.

---

### Bias Begins with Data: The FairGround Corpus for Robust and Reproducible Research on Algorithmic Fairness

Authors: Jan Simson, Alessandro Fabris, Cosima Fröhner, Frauke Kreuter, Christoph Kern
Published: 2025-10-25
Categories: cs.LG, cs.CY, stat.ML
arXiv: https://arxiv.org/abs/2510.22363v1

As machine learning (ML) systems are increasingly adopted in high-stakes decision-making domains, ensuring fairness in their outputs has become a central challenge. At the core of fair ML research are the datasets used to investigate bias and develop mitigation strategies. Yet, much of the existing work relies on a narrow selection of datasets--often arbitrarily chosen, inconsistently processed, and lacking in diversity--undermining the generalizability and reproducibility of results. To address these limitations, we present FairGround: a unified framework, data corpus, and Python package aimed at advancing reproducible research and critical data studies in fair ML classification. FairGround currently comprises 44 tabular datasets, each annotated with rich fairness-relevant metadata. Our accompanying Python package standardizes dataset loading, preprocessing, transformation, and splitting, streamlining experimental workflows. By providing a diverse and well-documented dataset corpus along with robust tooling, FairGround enables the development of fairer, more reliable, and more reproducible ML models. All resources are publicly available to support open and collaborative research.

---

### Mapping Faithful Reasoning in Language Models

Authors: Jiazheng Li, Andreas Damianou, J Rosser, José Luis Redondo García, Konstantina Palla
Published: 2025-10-25
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2510.22362v1

Chain-of-thought (CoT) traces promise transparency for reasoning language models, but prior work shows they are not always faithful reflections of internal computation. This raises challenges for oversight: practitioners may misinterpret decorative reasoning as genuine. We introduce Concept Walk, a general framework for tracing how a model's internal stance evolves with respect to a concept direction during reasoning. Unlike surface text, Concept Walk operates in activation space, projecting each reasoning step onto the concept direction learned from contrastive data. This allows us to observe whether reasoning traces shape outcomes or are discarded. As a case study, we apply Concept Walk to the domain of Safety using Qwen 3-4B. We find that in 'easy' cases, perturbed CoTs are quickly ignored, indicating decorative reasoning, whereas in 'hard' cases, perturbations induce sustained shifts in internal activations, consistent with faithful reasoning. The contribution is methodological: Concept Walk provides a lens to re-examine faithfulness through concept-specific internal dynamics, helping identify when reasoning traces can be trusted and when they risk misleading practitioners.

---

### EndoSfM3D: Learning to 3D Reconstruct Any Endoscopic Surgery Scene using Self-supervised Foundation Model

Authors: Changhao Zhang, Matthew J. Clarkson, Mobarak I. Hoque
Published: 2025-10-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2510.22359v1

3D reconstruction of endoscopic surgery scenes plays a vital role in enhancing scene perception, enabling AR visualization, and supporting context-aware decision-making in image-guided surgery. A critical yet challenging step in this process is the accurate estimation of the endoscope's intrinsic parameters. In real surgical settings, intrinsic calibration is hindered by sterility constraints and the use of specialized endoscopes with continuous zoom and telescope rotation. Most existing methods for endoscopic 3D reconstruction do not estimate intrinsic parameters, limiting their effectiveness for accurate and reliable reconstruction. In this paper, we integrate intrinsic parameter estimation into a self-supervised monocular depth estimation framework by adapting the Depth Anything V2 (DA2) model for joint depth, pose, and intrinsics prediction. We introduce an attention-based pose network and a Weight-Decomposed Low-Rank Adaptation (DoRA) strategy for efficient fine-tuning of DA2. Our method is validated on the SCARED and C3VD public datasets, demonstrating superior performance compared to recent state-of-the-art approaches in self-supervised monocular depth estimation and 3D reconstruction. Code and model weights can be found in project repository: https://github.com/MOYF-beta/EndoSfM3D.

---

### Irony Detection in Urdu Text: A Comparative Study Using Machine Learning Models and Large Language Models

Authors: Fiaz Ahmad, Nisar Hussain, Amna Qasim, Momina Hafeez, Muhammad Usman Grigori Sidorov, Alexander Gelbukh
Published: 2025-10-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2510.22356v1

Ironic identification is a challenging task in Natural Language Processing, particularly when dealing with languages that differ in syntax and cultural context. In this work, we aim to detect irony in Urdu by translating an English Ironic Corpus into the Urdu language. We evaluate ten state-of-the-art machine learning algorithms using GloVe and Word2Vec embeddings, and compare their performance with classical methods. Additionally, we fine-tune advanced transformer-based models, including BERT, RoBERTa, LLaMA 2 (7B), LLaMA 3 (8B), and Mistral, to assess the effectiveness of large-scale models in irony detection. Among machine learning models, Gradient Boosting achieved the best performance with an F1-score of 89.18%. Among transformer-based models, LLaMA 3 (8B) achieved the highest performance with an F1-score of 94.61%. These results demonstrate that combining transliteration techniques with modern NLP models enables robust irony detection in Urdu, a historically low-resource language.

---

### Design Stability in Adaptive Experiments: Implications for Treatment Effect Estimation

Authors: Saikat Sengupta, Koulik Khamaru, Suvrojit Ghosh, Tirthankar Dasgupta
Published: 2025-10-25
Categories: math.ST, stat.ME, stat.ML
arXiv: https://arxiv.org/abs/2510.22351v2

We study the problem of estimating the average treatment effect (ATE) under sequentially adaptive treatment assignment mechanisms. In contrast to classical completely randomized designs, we consider a setting in which the probability of assigning treatment to each experimental unit may depend on prior assignments and observed outcomes. Within the potential outcomes framework, we propose and analyze two natural estimators for the ATE: the inverse propensity weighted (IPW) estimator and an augmented IPW (AIPW) estimator. The cornerstone of our analysis is the concept of design stability, which requires that as the number of units grows, either the assignment probabilities converge, or sample averages of the inverse propensity scores and of the inverse complement propensity scores converge in probability to fixed, non-random limits. Our main results establish central limit theorems for both the IPW and AIPW estimators under design stability and provide explicit expressions for their asymptotic variances. We further propose estimators for these variances, enabling the construction of asymptotically valid confidence intervals. Finally, we illustrate our theoretical results in the context of Wei's adaptive coin design and Efron's biased coin design, highlighting the applicability of the proposed methods to sequential experimentation with adaptive randomization.

---