# Research Papers — 2025-11 · bundle page 53 of 197 > Full-text markdown of papers #2601–2650 of 9830 published in 2025-11, newest first. AI-readable mirrors of the full 9830-paper 2025-11 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2025-11/page/53 [← Previous bundle](https://aipod.dev/papers/archive/2025-11/page/52.md) [Next bundle →](https://aipod.dev/papers/archive/2025-11/page/54.md) ### Evaluating perturbation robustnessof generative systems that use COBOL code inputs Authors: Samuel Ackerman, Wesam Ibraheem, Orna Raz, Marcel Zalmanovici Published: 2025-11-23 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2511.18488v1 Systems incorporating large language models (LLMs) as a component are known to be sensitive (i.e., non-robust) to minor input variations that do not change the meaning of the input; such sensitivity may reduce the system's usefulness. Here, we present a framework to evaluate robustness of systems using COBOL code as input; our application is translation between COBOL and Java programming languages, but the approach extends to other tasks such as code generation or explanation. Targeting robustness of systems with COBOL as input is essential yet challenging. Many business-critical applications are written in COBOL, yet these are typically proprietary legacy applications and their code is unavailable to LLMs for training. We develop a library of COBOL paragraph and full-program perturbation methods, and create variant-expanded versions of a benchmark dataset of examples for a specific task. The robustness of the LLM-based system is evaluated by measuring changes in values of individual and aggregate metrics calculated on the system's outputs. Finally, we present a series of dynamic table and chart visualization dashboards that assist in debugging the system's outputs, and monitoring and understanding root causes of the system's sensitivity to input variation. These tools can be further used to improve the system by, for instance, indicating variations that should be handled by pre-processing steps. --- ### InstructAudio: Unified speech and music generation with natural language instruction Authors: Chunyu Qiang, Kang Yin, Xiaopeng Wang, Yuzhe Liang, Jiahui Zhao, Ruibo Fu, Tianrui Wang, Cheng Gong, Chen Zhang, Longbiao Wang, Jianwu Dang Published: 2025-11-23 Categories: eess.AS, cs.AI, cs.CL, cs.SD arXiv: https://arxiv.org/abs/2511.18487v1 Text-to-speech (TTS) and text-to-music (TTM) models face significant limitations in instruction-based control. TTS systems usually depend on reference audio for timbre, offer only limited text-level attribute control, and rarely support dialogue generation. TTM systems are constrained by input conditioning requirements that depend on expert knowledge annotations. The high heterogeneity of these input control conditions makes them difficult to joint modeling with speech synthesis. Despite sharing common acoustic modeling characteristics, these two tasks have long been developed independently, leaving open the challenge of achieving unified modeling through natural language instructions. We introduce InstructAudio, a unified framework that enables instruction-based (natural language descriptions) control of acoustic attributes including timbre (gender, age), paralinguistic (emotion, style, accent), and musical (genre, instrument, rhythm, atmosphere). It supports expressive speech, music, and dialogue generation in English and Chinese. The model employs joint and single diffusion transformer layers with a standardized instruction-phoneme input format, trained on 50K hours of speech and 20K hours of music data, enabling multi-task learning and cross-modal alignment. Fig. 1 visualizes performance comparisons with mainstream TTS and TTM models, demonstrating that InstructAudio achieves optimal results on most metrics. To our best knowledge, InstructAudio represents the first instruction-controlled framework unifying speech and music generation. Audio samples are available at: https://qiangchunyu.github.io/InstructAudio/ --- ### Adaptive Mesh-Quantization for Neural PDE Solvers Authors: Winfried van den Dool, Maksim Zhdanov, Yuki M. Asano, Max Welling Published: 2025-11-23 Categories: cs.LG arXiv: https://arxiv.org/abs/2511.18474v1 Physical systems commonly exhibit spatially varying complexity, presenting a significant challenge for neural PDE solvers. While Graph Neural Networks can handle the irregular meshes required for complex geometries and boundary conditions, they still apply uniform computational effort across all nodes regardless of the underlying physics complexity. This leads to inefficient resource allocation where computationally simple regions receive the same treatment as complex phenomena. We address this challenge by introducing Adaptive Mesh Quantization: spatially adaptive quantization across mesh node, edge, and cluster features, dynamically adjusting the bit-width used by a quantized model. We propose an adaptive bit-width allocation strategy driven by a lightweight auxiliary model that identifies high-loss regions in the input mesh. This enables dynamic resource distribution in the main model, where regions of higher difficulty are allocated increased bit-width, optimizing computational resource utilization. We demonstrate our framework's effectiveness by integrating it with two state-of-the-art models, MP-PDE and GraphViT, to evaluate performance across multiple tasks: 2D Darcy flow, large-scale unsteady fluid dynamics in 2D, steady-state Navier-Stokes simulations in 3D, and a 2D hyper-elasticity problem. Our framework demonstrates consistent Pareto improvements over uniformly quantized baselines, yielding up to 50% improvements in performance at the same cost. --- ### Uncertainty Quantification in HSI Reconstruction using Physics-Aware Diffusion Priors and Optics-Encoded Measurements Authors: Juan Romero, Qiang Fu, Matteo Ravasi, Wolfgang Heidrich Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18473v1 Hyperspectral image reconstruction from a compressed measurement is a highly ill-posed inverse problem. Current data-driven methods suffer from hallucination due to the lack of spectral diversity in existing hyperspectral image datasets, particularly when they are evaluated for the metamerism phenomenon. In this work, we formulate hyperspectral image (HSI) reconstruction as a Bayesian inference problem and propose a framework, HSDiff, that utilizes an unconditionally trained, pixel-level diffusion prior and posterior diffusion sampling to generate diverse HSI samples consistent with the measurements of various hyperspectral image formation models. We propose an enhanced metameric augmentation technique using region-based metameric black and partition-of-union spectral upsampling to expand training with physically valid metameric spectra, strengthening the prior diversity and improving uncertainty calibration. We utilize HSDiff to investigate how the studied forward models shape the posterior distribution and demonstrate that guiding with effective spectral encoding provides calibrated informative uncertainty compared to non-encoded models. Through the lens of the Bayesian framework, HSDiff offers a complete, high-performance method for uncertainty-aware HSI reconstruction. Our results also reiterate the significance of effective spectral encoding in snapshot hyperspectral imaging. --- ### Jacobian-Aware Posterior Sampling for Inverse Problems Authors: Liav Hen, Tom Tirer, Raja Giryes, Shady Abu-Hussein Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18471v3 Diffusion models provide powerful generative priors for solving inverse problems by sampling from a posterior distribution conditioned on corrupted measurements. Existing methods primarily follow two paradigms: direct methods, which approximate the likelihood term, and proximal methods, which incorporate intermediate solutions satisfying measurement constraints into the sampling process. Under standard Gaussian approximations and locally-linear measurements, we demonstrate that these approaches differ fundamentally in their treatment of the diffusion denoiser's Jacobian within the likelihood term. While this Jacobian encodes critical prior knowledge of the data distribution, training-induced non-idealities can degrade performance in zero-shot settings. In this work, we bridge direct and proximal approaches by proposing a principled Jacobian-Aware Posterior Sampler (JAPS). JAPS leverages the Jacobian's prior knowledge while mitigating its detrimental effects through a corresponding proximal solution, requiring no additional computational cost. Additionally, we integrate our guidance into DDIM sampling, with a corrected conditional factor that has been missing in previous works. Our method enhances reconstruction quality across diverse linear and nonlinear noisy imaging tasks, outperforming existing diffusion-based baselines in perceptual quality while maintaining or improving distortion metrics. --- ### Robust Posterior Diffusion-based Sampling via Adaptive Guidance Scale Authors: Liav Hen, Tom Tirer, Raja Giryes, Shady Abu-Hussein Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18471v1 Diffusion models have recently emerged as powerful generative priors for solving inverse problems, achieving state-of-the-art results across various imaging tasks. A central challenge in this setting lies in balancing the contribution of the prior with the data fidelity term: overly aggressive likelihood updates may introduce artifacts, while conservative updates can slow convergence or yield suboptimal reconstructions. In this work, we propose an adaptive likelihood step-size strategy to guide the diffusion process for inverse-problem formulations. Specifically, we develop an observation-dependent weighting scheme based on the agreement between two different approximations of the intractable intermediate likelihood gradients, that adapts naturally to the diffusion schedule, time re-spacing, and injected stochasticity. The resulting approach, Adaptive Posterior diffusion Sampling (AdaPS), is hyperparameter-free and improves reconstruction quality across diverse imaging tasks - including super-resolution, Gaussian deblurring, and motion deblurring - on CelebA-HQ and ImageNet-256 validation sets. AdaPS consistently surpasses existing diffusion-based baselines in perceptual quality with minimal or no loss in distortion, without any task-specific tuning. Extensive ablation studies further demonstrate its robustness to the number of diffusion steps, observation noise levels, and varying stochasticity. --- ### Gaze Beyond the Frame: Forecasting Egocentric 3D Visual Span Authors: Heeseung Yun, Joonil Na, Jaeyeon Kim, Calvin Murdock, Gunhee Kim Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18470v1 People continuously perceive and interact with their surroundings based on underlying intentions that drive their exploration and behaviors. While research in egocentric user and scene understanding has focused primarily on motion and contact-based interaction, forecasting human visual perception itself remains less explored despite its fundamental role in guiding human actions and its implications for AR/VR and assistive technologies. We address the challenge of egocentric 3D visual span forecasting, predicting where a person's visual perception will focus next within their three-dimensional environment. To this end, we propose EgoSpanLift, a novel method that transforms egocentric visual span forecasting from 2D image planes to 3D scenes. EgoSpanLift converts SLAM-derived keypoints into gaze-compatible geometry and extracts volumetric visual span regions. We further combine EgoSpanLift with 3D U-Net and unidirectional transformers, enabling spatio-temporal fusion to efficiently predict future visual span in the 3D grid. In addition, we curate a comprehensive benchmark from raw egocentric multisensory data, creating a testbed with 364.6K samples for 3D visual span forecasting. Our approach outperforms competitive baselines for egocentric 2D gaze anticipation and 3D localization while achieving comparable results even when projected back onto 2D image planes without additional 2D-specific training. --- ### SloMo-Fast: Slow-Momentum and Fast-Adaptive Teachers for Source-Free Continual Test-Time Adaptation Authors: Md Akil Raihan Iftee, Mir Sazzat Hossain, Rakibul Hasan Rajib, Tariq Iqbal, Md Mofijul Islam, M Ashraful Amin, Amin Ahsan Ali, AKM Mahbubur Rahman Published: 2025-11-23 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2511.18468v2 Continual Test-Time Adaptation (CTTA) is crucial for deploying models in real-world applications with unseen, evolving target domains. Existing CTTA methods, however, often rely on source data or prototypes, limiting their applicability in privacy-sensitive and resource-constrained settings. Although several methods attempt to mitigate catastrophic forgetting, they often fail to preserve long-term domain-specific knowledge across many domain shifts. Moreover, their relatively slow adaptation rates during domain transitions can cause error accumulation, allowing mistakes to propagate before effective adaptation occurs. To address these challenges, we propose SloMo-Fast, a source-free, dual-teacher CTTA framework designed for enhanced quick adaptability and generalization. It includes two complementary teachers: the Slow-Teacher, which exhibits slow forgetting and retains long-term knowledge of previously encountered domains to ensure robust generalization, and the Fast-Teacher rapidly adapts to new domains while accumulating and integrating knowledge across them. This framework preserves knowledge of past domains and adapts efficiently to new ones. Our extensive experiments show that SloMo-Fast consistently outperforms state-of-the-art methods across Cyclic Test-Time Adaptation (Cyclic-TTA), a CTTA benchmark that simulates recurring domain shifts, along with ten other CTTA settings, highlighting its ability to both adapt and generalize across evolving, revisited domains. --- ### SloMo-Fast: Slow-Momentum and Fast-Adaptive Teachers for Source-Free Continual Test-Time Adaptation Authors: Md Akil Raihan Iftee, Mir Sazzat Hossain, Rakibul Hasan Rajib, Tariq Iqbal, Md Mofijul Islam, M Ashraful Amin, Amin Ahsan Ali, AKM Mahbubur Rahman Published: 2025-11-23 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2511.18468v1 Continual Test-Time Adaptation (CTTA) is crucial for deploying models in real-world applications with unseen, evolving target domains. Existing CTTA methods, however, often rely on source data or prototypes, limiting their applicability in privacy-sensitive and resource-constrained settings. Additionally, these methods suffer from long-term forgetting, which degrades performance on previously encountered domains as target domains shift. To address these challenges, we propose SloMo-Fast, a source-free, dual-teacher CTTA framework designed for enhanced adaptability and generalization. It includes two complementary teachers: the Slow-Teacher, which exhibits slow forgetting and retains long-term knowledge of previously encountered domains to ensure robust generalization, and the Fast-Teacher rapidly adapts to new domains while accumulating and integrating knowledge across them. This framework preserves knowledge of past domains and adapts efficiently to new ones. We also introduce Cyclic Test-Time Adaptation (Cyclic-TTA), a novel CTTA benchmark that simulates recurring domain shifts. Our extensive experiments demonstrate that SloMo-Fast consistently outperforms state-of-the-art methods across Cyclic-TTA, as well as ten other CTTA settings, highlighting its ability to both adapt and generalize across evolving and revisited domains. --- ### Shadows in the Code: Exploring the Risks and Defenses of LLM-based Multi-Agent Software Development Systems Authors: Xiaoqing Wang, Keman Huang, Bin Liang, Hongyu Li, Xiaoyong Du Published: 2025-11-23 Categories: cs.CR, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2511.18467v1 The rapid advancement of Large Language Model (LLM)-driven multi-agent systems has significantly streamlined software developing tasks, enabling users with little technical expertise to develop executable applications. While these systems democratize software creation through natural language requirements, they introduce significant security risks that remain largely unexplored. We identify two risky scenarios: Malicious User with Benign Agents (MU-BA) and Benign User with Malicious Agents (BU-MA). We introduce the Implicit Malicious Behavior Injection Attack (IMBIA), demonstrating how multi-agent systems can be manipulated to generate software with concealed malicious capabilities beneath seemingly benign applications, and propose Adv-IMBIA as a defense mechanism. Evaluations across ChatDev, MetaGPT, and AgentVerse frameworks reveal varying vulnerability patterns, with IMBIA achieving attack success rates of 93%, 45%, and 71% in MU-BA scenarios, and 71%, 84%, and 45% in BU-MA scenarios. Our defense mechanism reduced attack success rates significantly, particularly in the MU-BA scenario. Further analysis reveals that compromised agents in the coding and testing phases pose significantly greater security risks, while also identifying critical agents that require protection against malicious user exploitation. Our findings highlight the urgent need for robust security measures in multi-agent software development systems and provide practical guidelines for implementing targeted, resource-efficient defensive strategies. --- ### Reliable Selection of Heterogeneous Treatment Effect Estimators Authors: Jiayi Guo, Zijun Gao Published: 2025-11-23 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2511.18464v2 We study the problem of selecting the best heterogeneous treatment effect (HTE) estimator from a collection of candidates in settings where the treatment effect is fundamentally unobserved. We cast estimator selection as a multiple testing problem and introduce a ground-truth-free procedure based on a cross-fitted, exponentially weighted test statistic. A key component of our method is a two-way sample splitting scheme that decouples nuisance estimation from weight learning and ensures the stability required for valid inference. Leveraging a stability-based central limit theorem, we establish asymptotic familywise error rate control under mild regularity conditions. Empirically, our procedure provides reliable error control while substantially reducing false selections compared with commonly used methods across ACIC 2016, IHDP, and Twins benchmarks, demonstrating that our method is feasible and powerful even without ground-truth treatment effects. --- ### Reliable Selection of Heterogeneous Treatment Effect Estimators Authors: Jiayi Guo, Zijun Gao Published: 2025-11-23 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2511.18464v1 We study the problem of selecting the best heterogeneous treatment effect (HTE) estimator from a collection of candidates in settings where the treatment effect is fundamentally unobserved. We cast estimator selection as a multiple testing problem and introduce a ground-truth-free procedure based on a cross-fitted, exponentially weighted test statistic. A key component of our method is a two-way sample splitting scheme that decouples nuisance estimation from weight learning and ensures the stability required for valid inference. Leveraging a stability-based central limit theorem, we establish asymptotic familywise error rate control under mild regularity conditions. Empirically, our procedure provides reliable error control while substantially reducing false selections compared with commonly used methods across ACIC 2016, IHDP, and Twins benchmarks, demonstrating that our method is feasible and powerful even without ground-truth treatment effects. --- ### Decoupling Perception from Reasoning for Hallucination-Resistant Video Understanding Authors: Bowei Pu, Chuanbin Liu, Yifan Ge, Peicheng Zhou, Yiwei Sun, Zhiying Lu, Zhangchi Hu, Hongtao Xie Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18463v3 Video Large Language Models improve reasoning over complex videos by generating intermediate reasoning text. However, reliable reasoning depends on accurate video perception. In existing approaches, perception evidence is intertwined with reasoning text, making it difficult to directly supervise the perception process. We argue that reliable supervision requires explicitly separating perception evidence from reasoning so that perception can be verified independently. To supervise perception directly, we propose Decoupled Perception and Logic (DPL), which represents perception as fixed-format evidence units containing timestamps and visual descriptions. This structured representation enables direct extraction of perception content and simplifies alignment between video segments and reward evaluation. Building on DPL, we introduce a perception reward that encourages both hallucination resistance and perception-based reasoning. An Factual-Aware Evaluator (FAE) provides anti-hallucination scores and achieves hallucination evaluation performance comparable to GPT-4o. In addition, we validate reasoning consistency by feeding perception results and questions into a reference model. Experiments show that, by providing reliable process rewards, Video-DPL consistently improves post-training performance at both 3B and 7B scales, while delivering higher data efficiency. --- ### Alternating Perception-Reasoning for Hallucination-Resistant Video Understanding Authors: Bowei Pu, Chuanbin Liu, Yifan Ge, Peichen Zhou, Yiwei Sun, Zhiyin Lu, Jiankang Wang, Hongtao Xie Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18463v1 Sufficient visual perception is the foundation of video reasoning. Nevertheless, existing Video Reasoning LLMs suffer from perception shortcuts, relying on a flawed single-step perception paradigm. This paradigm describes the video and then conducts reasoning, which runs the risk of insufficient evidence and emergent hallucinations. To address these issues, we introduce a new framework that integrates a loop-based paradigm with an anti-hallucination reward. First, to address the insufficient evidence, we introduce the Perception Loop Reasoning (PLR) paradigm. Instead of describing the video at once, each loop requires the model to describe a video segment with precise timestamps, analyze this segment, and decide the next action. Second, for the risk of hallucinations, the Factual-Aware Evaluator (FAE) evaluates each perception result as a reliable anti-hallucination reward. This reward encourages the model to provide sufficient and precise video evidence. Our FAE, which performs comparably to GPT-4o, is tuned on our AnetHallu-117K, a large-scale hallucination judgment preference dataset. Extensive experiments show that our Video-PLR achieves the state-of-the-art in both 3B and 7B parameter scales and has the best data efficiency. Our code, models, and datasets are released on: https://github.com/BoweiPu/VideoPLR. --- ### Radiation-Preserving Selective Imaging for Pediatric Hip Dysplasia: A Cross-Modal Ultrasound-Xray Policy with Limited Labels Authors: Duncan Stothers, Ben Stothers, Emily Schaeffer, Kishore Mulpuri Published: 2025-11-23 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2511.18457v2 We study an ultrasound-first, radiation-preserving policy for developmental dysplasia of the hip (DDH) that requests a radiograph only when needed. We (i) pretrain modality-specific encoders (ResNet-18) with SimSiam on a large unlabelled registry (37186 ultrasound; 19546 radiographs), (ii) freeze the backbones and fit small, measurement-faithful heads on DDH-relevant landmarks and measurements, (iii) calibrate a one-sided conformal deferral rule on ultrasound predictions that provides finite sample marginal coverage guarantees under exchangeability, using a held-out calibration set. Ultrasound heads predict Graf alpha, beta, and femoral head coverage; X-ray heads predict acetabular index (AI), center-edge (CE) angle and IHDI grade. On our held out labeled evaluation set, ultrasound measurement error is modest (e.g., alpha MAE ~= 9.7 degrees, coverage MAE ~= 14.0%), while radiographic probes achieve AI and CE MAEs of ~= 7.6 degrees and ~= 8.9 degrees, respectively. The calibrated US-only policy is explored across rule families (alpha-only; alpha OR coverage; alpha AND coverage), conformal miscoverage levels, and per-utility trade-offs using decision-curve analysis. Conservative settings yield high coverage with near-zero US-only rates; permissive settings (e.g., alpha OR coverage at larger deltas) achieve non-zero US-only throughput with expected coverage tradeoffs. The result is a simple, reproducible pipeline that turns limited labels into interpretable measurements and tunable selective imaging curves suitable for clinical handoff and future external validation. --- ### Radiation-Preserving Selective Imaging for Pediatric Hip Dysplasia: A Cross-Modal Ultrasound-Xray Policy with Limited Labels Authors: Duncan Stothers, Ben Stothers, Emily Schaeffer, Kishore Mulpuri Published: 2025-11-23 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2511.18457v1 We study an ultrasound-first, radiation-preserving policy for developmental dysplasia of the hip (DDH) that requests a radiograph only when needed. We (i) pretrain modality-specific encoders (ResNet-18) with SimSiam on a large unlabelled registry (37186 ultrasound; 19546 radiographs), (ii) freeze the backbones and fit small, measurement-faithful heads on DDH relevant landmarks and measurements (iii) calibrate a one sided conformal deferral rule on ultrasound predictions that provides finite sample coverage guarantees under exchangeability, using a held-out calibration set. Ultrasound heads predict Graf alpha, beta, and femoral head coverage; X-ray heads predict acetabular index (AI), center-edge (CE) angle and IHDI grade. On our held out labeled evaluation set, ultrasound measurement error is modest (e.g., alpha MAE ~= 9.7 degrees, coverage MAE ~= 14.0%), while radiographic probes achieve AI and CE MAEs of ~= 7.6 degrees and ~= 8.9 degrees, respectively. The calibrated US-only policy is explored across rule families (alpha-only; alpha OR coverage; alpha AND coverage), uncertainty inflation factors, and per-utility trade-offs using decision-curve analysis. Conservative settings yield high coverage with near-zero US-only rates; permissive settings (e.g., alpha OR coverage at larger deltas) achieve non-zero US-only throughput with expected coverage tradeoffs. The result is a simple, reproducible pipeline that turns limited labels into interpretable measurements and tunable selective imaging curves suitable for clinical handoff and future external validation. --- ### AttnRegDeepLab: A Two-Stage Decoupled Framework for Interpretable Embryo Fragmentation Grading Authors: Ming-Jhe Lee, Chang-Hong Wu, Jung-Hua Wang, Ming-Jer Chen, Yu-Chiao Yi, Tsung-Hsien Lee Published: 2025-11-23 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2511.18454v4 Assessing embryo fragmentation is crucial for predicting IVF success, yet manual grading is prone to subjectivity, and existing AI models struggle with clinical interpretability and segmentation errors. We propose AttnRegDeepLab, a Multi-Task Learning (MTL) framework designed to solve these challenges. The model enhances a DeepLabV3+ decoder with Attention Gates to filter out cytoplasmic noise and retain sharp contour details. It also introduces a Multi-Scale Regression Head with Feature Injection, guiding the segmentation process with global grading priors to eliminate systematic area estimation errors. Based on a two-stage decoupled training strategy and a range-based loss for weakly labeled data, our method resolves MTL gradient conflicts. AttnRegDeepLab yields high grading precision and excellent segmentation quality (Dice coefficient = 0.729), avoiding the trade-off between contour integrity and grading accuracy seen under standard joint optimization. This provides a reliable, clinically interpretable tool balancing visual and quantitative accuracy. --- ### RegDeepLab: A Two-Stage Decoupled Framework for Interpretable Embryo Fragmentation Grading Authors: Ming-Jhe Lee Published: 2025-11-23 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2511.18454v1 The degree of embryo fragmentation serves as a critical morphological indicator for assessing embryo developmental potential in In Vitro Fertilization (IVF) clinical decision-making. However, current manual grading processes are not only time-consuming but also limited by significant inter-observer variability and efficiency bottlenecks. Although deep learning has demonstrated potential in automated grading in recent years, existing solutions face a significant challenge: pure regression models lack the visual explainability required for clinical practice, while pure segmentation models struggle to directly translate pixel-level masks into precise clinical grades. This study proposes RegDeepLab, a dual-branch Multi-Task Learning (MTL) framework that integrates State-of-the-Art (SOTA) semantic segmentation (DeepLabV3+) with a multi-scale regression head. Addressing the common issues of "Gradient Conflict" and "Negative Transfer" in multi-task training, we propose a "Two-Stage Decoupled Training Strategy." Experimental results demonstrate that while standard end-to-end MTL training can minimize grading error (MAE=0.046) through our designed "Feature Injection" mechanism, it compromises the integrity of segmentation boundaries. In contrast, our decoupled strategy successfully provides robust and high-precision grading predictions while preserving SOTA-level segmentation accuracy (Dice=0.729). Furthermore, we introduce a "Range Loss" to effectively utilize large-scale discrete grading data for semi-supervised learning. This study ultimately presents a dual-module clinical auxiliary solution that combines high accuracy with visual explainability. --- ### NAF: Zero-Shot Feature Upsampling via Neighborhood Attention Filtering Authors: Loick Chambon, Paul Couairon, Eloi Zablocki, Alexandre Boulch, Nicolas Thome, Matthieu Cord Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18452v1 Vision Foundation Models (VFMs) extract spatially downsampled representations, posing challenges for pixel-level tasks. Existing upsampling approaches face a fundamental trade-off: classical filters are fast and broadly applicable but rely on fixed forms, while modern upsamplers achieve superior accuracy through learnable, VFM-specific forms at the cost of retraining for each VFM. We introduce Neighborhood Attention Filtering (NAF), which bridges this gap by learning adaptive spatial-and-content weights through Cross-Scale Neighborhood Attention and Rotary Position Embeddings (RoPE), guided solely by the high-resolution input image. NAF operates zero-shot: it upsamples features from any VFM without retraining, making it the first VFM-agnostic architecture to outperform VFM-specific upsamplers and achieve state-of-the-art performance across multiple downstream tasks. It maintains high efficiency, scaling to 2K feature maps and reconstructing intermediate-resolution maps at 18 FPS. Beyond feature upsampling, NAF demonstrates strong performance on image restoration, highlighting its versatility. Code and checkpoints are available at https://github.com/valeoai/NAF. --- ### ORIGAMISPACE: Benchmarking Multimodal LLMs in Multi-Step Spatial Reasoning with Mathematical Constraints Authors: Rui Xu, Dakuan Lu, Zicheng Zhao, Xiaoyu Tan, Xintao Wang, Siyu Yuan, Jiangjie Chen, Yinghui Xu Published: 2025-11-23 Categories: cs.AI arXiv: https://arxiv.org/abs/2511.18450v1 Spatial reasoning is a key capability in the field of artificial intelligence, especially crucial in areas such as robotics, computer vision, and natural language understanding. However, evaluating the ability of multimodal large language models(MLLMs) in complex spatial reasoning still faces challenges, particularly in scenarios requiring multi-step reasoning and precise mathematical constraints. This paper introduces ORIGAMISPACE, a new dataset and benchmark designed to evaluate the multi-step spatial reasoning ability and the capacity to handle mathematical constraints of MLLMs through origami tasks. The dataset contains 350 data instances,each comprising a strictly formatted crease pattern (CP diagram), the Compiled Flat Pattern, the complete Folding Process, and the final Folded Shape Image. We propose four evaluation tasks: Pattern Prediction, Multi-step Spatial Reasoning, Spatial Relationship Prediction, and End-to-End CP Code Generation. For the CP code generation task, we design an interactive environment and explore the possibility of using reinforcement learning methods to train MLLMs. Through experiments on existing MLLMs, we initially reveal the strengths and weaknesses of these models in handling complex spatial reasoning tasks. --- ### EventBench: Towards Comprehensive Benchmarking of Event-based MLLMs Authors: Shaoyu Liu, Jianing Li, Guanghui Zhao, Yunjian Zhang, Xiangyang Ji Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18448v1 Multimodal large language models (MLLMs) have made significant advancements in event-based vision, yet the comprehensive evaluation of their capabilities within a unified benchmark remains largely unexplored. In this work, we introduce EventBench, a benchmark that offers eight diverse task metrics together with a large-scale event stream dataset. EventBench differs from existing event-based benchmarks in four key aspects: (1) openness in accessibility, releasing all raw event streams and task instructions across eight evaluation metrics; (2) diversity in task coverage, spanning understanding, recognition, and spatial reasoning tasks for comprehensive capability assessment; (3) integration in spatial dimensions, pioneering the design of 3D spatial reasoning tasks for event-based MLLMs; and (4) scale in data volume, with an accompanying training set of over one million event-text pairs supporting large-scale training and evaluation. Using EventBench, we evaluate state-of-the-art closed-source models such as GPT-5 and Gemini-2.5 Pro, leading open-source models including Qwen2.5-VL and InternVL3, and event-based MLLMs such as EventGPT that directly process raw event streams. Extensive evaluation reveals that while current event-based MLLMs demonstrate strong performance in event stream understanding, they continue to struggle with fine-grained recognition and spatial reasoning. --- ### SineProject: Machine Unlearning for Stable Vision Language Alignment Authors: Arpit Garg, Hemanth Saratchandran, Simon Lucey Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18444v2 Multimodal Large Language Models (MLLMs) increasingly need to forget specific knowledge such as unsafe or private information without requiring full retraining. However, existing unlearning methods often disrupt vision language alignment, causing models to reject both harmful and benign queries. We trace this failure to the projector network during unlearning, its Jacobian becomes severely illconditioned, leading to unstable optimization and drift in cross modal embeddings. We introduce SineProject, a simple method that augments the frozen projector with sinusoidally modulated trainable parameters, improving the Jacobian's spectral conditioning and stabilizing alignment throughout unlearning. Across standard safety and privacy unlearning benchmarks using LLaVA v1.5 7B and 13B, SineProject reduces benign query refusals while achieving complete forgetting of targeted information, yielding state of the art forget retain trade offs with negligible computational overhead. --- ### SineProject: Machine Unlearning for Stable Vision Language Alignment Authors: Arpit Garg, Hemanth Saratchandran, Simon Lucey Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18444v1 Multimodal Large Language Models (MLLMs) increasingly need to forget specific knowledge such as unsafe or private information without requiring full retraining. However, existing unlearning methods often disrupt vision language alignment, causing models to reject both harmful and benign queries. We trace this failure to the projector network during unlearning, its Jacobian becomes severely illconditioned, leading to unstable optimization and drift in cross modal embeddings. We introduce SineProject, a simple method that augments the frozen projector with sinusoidally modulated trainable parameters, improving the Jacobian's spectral conditioning and stabilizing alignment throughout unlearning. Across standard safety and privacy unlearning benchmarks using LLaVA v1.5 7B and 13B, SineProject reduces benign query refusals while achieving complete forgetting of targeted information, yielding state of the art forget retain trade offs with negligible computational overhead. --- ### ReCoGS: Real-time ReColoring for Gaussian Splatting scenes Authors: Lorenzo Rutayisire, Nicola Capodieci, Fabio Pellacini Published: 2025-11-23 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2511.18441v1 Gaussian Splatting has emerged as a leading method for novel view synthesis, offering superior training efficiency and real-time inference compared to NeRF approaches, while still delivering high-quality reconstructions. Beyond view synthesis, this 3D representation has also been explored for editing tasks. Many existing methods leverage 2D diffusion models to generate multi-view datasets for training, but they often suffer from limitations such as view inconsistencies, lack of fine-grained control, and high computational demand. In this work, we focus specifically on the editing task of recoloring. We introduce a user-friendly pipeline that enables precise selection and recoloring of regions within a pre-trained Gaussian Splatting scene. To demonstrate the real-time performance of our method, we also present an interactive tool that allows users to experiment with the pipeline in practice. Code is available at https://github.com/loryruta/recogs. --- ### Perceptual-Evidence Anchored Reinforced Learning for Multimodal Reasoning Authors: Chi Zhang, Haibo Qiu, Qiming Zhang, Yufei Xu, Zhixiong Zeng, Siqi Yang, Peng Shi, Lin Ma, Jing Zhang Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18437v1 Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced the reasoning capabilities of Large Language Models (LLMs) and is now being applied to Vision-Language Models (VLMs). However, vanilla RLVR for VLMs verifies only the final textual output, critically neglecting the foundational step of visual perception. This oversight leads to visual hallucinations and reward hacking, as reasoning built upon flawed perception is inherently unreliable. To address this, we propose PEARL (Perceptual-Evidence Anchored Reinforced Learning), a dual-branch, perception-reasoning synergistic that strengthens multimodal reasoning by explicitly anchoring it to verified visual evidence. For each reasoning-oriented QA instance, PEARL first derive a perception checklist -- a set of perception-oriented sub-questions with verifiable answers that probe the model's understanding of key visual evidence. During training, auxiliary rollouts on this checklist yield a perceptual reward that both directly reinforces the model's perception ability and acts as a fidelity gate for reasoning. If the model passes the perception check, its policy update is biased towards evidence-anchored reasoning. Otherwise, the process is halted to prevent reasoning from flawed premises. PEARL can be seamlessly integrated with popular RL methods like GRPO and DAPO. Comprehensive experiments show PEARL achieves substantial gains on multimodal reasoning benchmarks, e.g., a +9.7% improvement over the baseline and +6.6% over GRPO on MathVerse. --- ### When Generative Replay Meets Evolving Deepfakes: Dual Confusion-Aware Regularization for Incremental Face Forgery Detection Authors: Hao Shen, Jikang Cheng, Renye Yan, Zhongyuan Wang, Wei Peng, Baojin Huang Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18436v2 The rapid advancement of face generation techniques has introduced an increasing variety of forgery methods, making incremental deepfake detection essential for maintaining robust detectors. While generative replay provides a promising solution to alleviate catastrophic forgetting without storing historical data, its effectiveness is hindered by \textbf{domain confusion} between generated samples and real data. We observe that replay samples fall into two categories: when the replay generator closely resembles the newly introduced forgery model, generated real samples overlap with the fake domain and become \textbf{domain-risky}; when the generator differs significantly, generated samples maintain clearer domain separation and can be treated as \textbf{domain-safe}. To address this challenge, we propose a Dual \textbf{C}onfusion-\textbf{A}ware \textbf{RE}gularization strategy, termed \textbf{Dual-CARE}. A Domain-aware Confusion Score (DC Score) is introduced to quantify domain confusion and \textbf{dual-modulate} the optimization of both replay generators and the incremental detector. Guided by DC Score, replay generators are updated to better approximate previous-task distributions, while the detector adopts different supervision strategies: domain-safe samples are directly supervised, whereas domain-risky samples are regulated using a Relative Separation Loss to balance supervision and confusion. Extensive experiments demonstrate that Dual-CARE effectively exploits generative replay and improves incremental deepfake detection under evolving forgery scenarios. --- ### When Generative Replay Meets Evolving Deepfakes: Domain-Aware Relative Weighting for Incremental Face Forgery Detection Authors: Hao Shen, Jikang Cheng, Renye Yan, Zhongyuan Wang, Wei Peng, Baojin Huang Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18436v1 The rapid advancement of face generation techniques has led to a growing variety of forgery methods. Incremental forgery detection aims to gradually update existing models with new forgery data, yet current sample replay-based methods are limited by low diversity and privacy concerns. Generative replay offers a potential solution by synthesizing past data, but its feasibility for forgery detection remains unclear. In this work, we systematically investigate generative replay and identify two scenarios: when the replay generator closely resembles the new forgery model, generated real samples blur the domain boundary, creating domain-risky samples; when the replay generator differs significantly, generated samples can be safely supervised, forming domain-safe samples. To exploit generative replay effectively, we propose a novel Domain-Aware Relative Weighting (DARW) strategy. DARW directly supervises domain-safe samples while applying a Relative Separation Loss to balance supervision and potential confusion for domain-risky samples. A Domain Confusion Score dynamically adjusts this tradeoff according to sample reliability. Extensive experiments demonstrate that DARW consistently improves incremental learning performance for forgery detection under different generative replay settings and alleviates the adverse impact of domain overlap. --- ### DocPTBench: Benchmarking End-to-End Photographed Document Parsing and Translation Authors: Yongkun Du, Pinxuan Chen, Xuye Ying, Zhineng Chen Published: 2025-11-23 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2511.18434v1 The advent of Multimodal Large Language Models (MLLMs) has unlocked the potential for end-to-end document parsing and translation. However, prevailing benchmarks such as OmniDocBench and DITrans are dominated by pristine scanned or digital-born documents, and thus fail to adequately represent the intricate challenges of real-world capture conditions, such as geometric distortions and photometric variations. To fill this gap, we introduce DocPTBench, a comprehensive benchmark specifically designed for Photographed Document Parsing and Translation. DocPTBench comprises over 1,300 high-resolution photographed documents from multiple domains, includes eight translation scenarios, and provides meticulously human-verified annotations for both parsing and translation. Our experiments demonstrate that transitioning from digital-born to photographed documents results in a substantial performance decline: popular MLLMs exhibit an average accuracy drop of 18% in end-to-end parsing and 12% in translation, while specialized document parsing models show significant average decrease of 25%. This substantial performance gap underscores the unique challenges posed by documents captured in real-world conditions and reveals the limited robustness of existing models. Dataset and code are available at https://github.com/Topdu/DocPTBench. --- ### Robust Differential Evolution via Nonlinear Population Size Reduction and Adaptive Restart: The ARRDE Algorithm Authors: Khoirul Faiq Muzakka, Ahsani Hafizhu Shali, Haris Suhendar, Sören Möller, Martin Finsterbusch Published: 2025-11-23 Categories: cs.NE, math.OC arXiv: https://arxiv.org/abs/2511.18429v5 Robustness across heterogeneous optimization regimes remains a central challenge in bound-constrained continuous optimization. In practice, users often prefer optimizers that remain reliable across different dimensionalities, landscape structures, and evaluation budgets. Yet many Differential Evolution (DE) variants that perform strongly in one regime degrade substantially when transferred to others. To address this issue, we propose the Adaptive Restart--Refine Differential Evolution (ARRDE) algorithm, a DE variant designed explicitly for cross-regime robustness. ARRDE combines an adaptive restart--refine strategy, a nonlinear population-size reduction schedule that depends on problem dimensionality, and a budget-aware population-initialization rule for restricted-budget settings. Because robustness cannot be established credibly from a narrow experimental setting, we evaluate ARRDE on five benchmark suites: CEC2011, CEC2017, CEC2019, CEC2020, and CEC2022. These suites span markedly different dimensions, landscape characteristics, and evaluation budgets, making this, to the best of our knowledge, one of the most comprehensive robustness-oriented evaluations reported for a proposed DE variant in this context. Since their official performance metrics emphasize different aspects and are not directly comparable, we additionally introduce a bounded accuracy-based scoring metric derived from relative error for cross-suite robustness assessment. Using both the official suite-specific metrics and the proposed unified metric, ARRDE demonstrates consistently strong performance and one of the most stable aggregate profiles across the five suites. These results support ARRDE as a competitive DE variant for robust optimization across heterogeneous benchmark regimes. --- ### Robust Differential Evolution via Nonlinear Population Size Reduction and Adaptive Restart: The ARRDE Algorithm Authors: Khoirul Faiq Muzakka, Ahsani Hafizhu Shali, Haris Suhendar, Sören Möller, Martin Finsterbusch Published: 2025-11-23 Categories: cs.NE, math.OC arXiv: https://arxiv.org/abs/2511.18429v1 This study is motivated by a robustness issue in numerical optimization of bound-constrained problems: many algorithms that perform well on a particular benchmark suite, such as the IEEE CEC2017 problems, struggle to maintain the same level of performance when applied to other suites that differ in dimensionality, landscape complexity, or the maximum number of function evaluations ($N_{\text{max}}$). To address this, we propose the Adaptive Restart-Refine Differential Evolution (ARRDE) algorithm, a new variant of Differential Evolution (DE). ARRDE builds upon the LSHADE algorithm, incorporates key mechanisms from jSO, and introduces a nonlinear population-size reduction strategy combined with an adaptive restart-refine mechanism. We evaluate ARRDE on five benchmark suites (CEC2011, CEC2017, CEC2019, CEC2020, and CEC2022) which, to the best of our knowledge, constitutes the most extensive experimental study to date in the context of algorithmic comparison, as most prior works consider only one or two suites. This broad evaluation enables a rigorous assessment of generalization across markedly different problem characteristics. To further support fair cross-suite comparisons, we also introduce a bounded accuracy-based scoring metric derived from relative error. Using both rank-based and accuracy-based metrics, and comparing against algorithms that perform strongly on CEC2017 (e.g., jSO and LSHADE-cnEpSin) as well as those that excel on CEC2020 (e.g., j2020 and NLSHADE-RSP), ARRDE consistently demonstrates top-tier performance, ranking first across all benchmark suites considered. These results highlight ARRDE's robustness and its superior generalization capability. --- ### LungX: A Hybrid EfficientNet-Vision Transformer Architecture with Multi-Scale Attention for Accurate Pneumonia Detection Authors: Mansur Yerzhanuly Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18425v1 Pneumonia remains a leading global cause of mortality where timely diagnosis is critical. We introduce LungX, a novel hybrid architecture combining EfficientNet's multi-scale features, CBAM attention mechanisms, and Vision Transformer's global context modeling for enhanced pneumonia detection. Evaluated on 20,000 curated chest X-rays from RSNA and CheXpert, LungX achieves state-of-the-art performance (86.5 percent accuracy, 0.943 AUC), representing a 6.7 percent AUC improvement over EfficientNet-B0 baselines. Visual analysis demonstrates superior lesion localization through interpretable attention maps. Future directions include multi-center validation and architectural optimizations targeting 88 percent accuracy for clinical deployment as an AI diagnostic aid. --- ### CrossJEPA: Cross-Modal Joint-Embedding Predictive Architecture for Efficient 3D Representation Learning from 2D Images Authors: Avishka Perera, Kumal Hewagamage, Saeedha Nazar, Kavishka Abeywardana, Hasitha Gallella, Ranga Rodrigo, Mohamed Afham Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18424v1 Image-to-point cross-modal learning has emerged to address the scarcity of large-scale 3D datasets in 3D representation learning. However, current methods that leverage 2D data often result in large, slow-to-train models, making them computationally expensive and difficult to deploy in resource-constrained environments. The architecture design of such models is therefore critical, determining their performance, memory footprint, and compute efficiency. The Joint-embedding Predictive Architecture (JEPA) has gained wide popularity in self-supervised learning for its simplicity and efficiency, but has been under-explored in cross-modal settings, partly due to the misconception that masking is intrinsic to JEPA. In this light, we propose CrossJEPA, a simple Cross-modal Joint Embedding Predictive Architecture that harnesses the knowledge of an image foundation model and trains a predictor to infer embeddings of specific rendered 2D views from corresponding 3D point clouds, thereby introducing a JEPA-style pretraining strategy beyond masking. By conditioning the predictor on cross-domain projection information, CrossJEPA purifies the supervision signal from semantics exclusive to the target domain. We further exploit the frozen teacher design with a one-time target embedding caching mechanism, yielding amortized efficiency. CrossJEPA achieves a new state-of-the-art in linear probing on the synthetic ModelNet40 (94.2%) and the real-world ScanObjectNN (88.3%) benchmarks, using only 14.1M pretraining parameters (8.5M in the point encoder), and about 6 pretraining hours on a standard single GPU. These results position CrossJEPA as a performant, memory-efficient, and fast-to-train framework for 3D representation learning via knowledge distillation. We analyze CrossJEPA intuitively, theoretically, and empirically, and extensively ablate our design choices. Code will be made available. --- ### General Agentic Memory Via Deep Research Authors: B. Y. Yan, Chaofan Li, Hongjin Qian, Shuqi Lu, Zheng Liu Published: 2025-11-23 Categories: cs.CL, cs.AI, cs.IR, cs.LG arXiv: https://arxiv.org/abs/2511.18423v1 Memory is critical for AI agents, yet the widely-adopted static memory, aiming to create readily available memory in advance, is inevitably subject to severe information loss. To address this limitation, we propose a novel framework called \textbf{general agentic memory (GAM)}. GAM follows the principle of "\textbf{just-in time (JIT) compilation}" where it focuses on creating optimized contexts for its client at runtime while keeping only simple but useful memory during the offline stage. To this end, GAM employs a duo-design with the following components. 1) \textbf{Memorizer}, which highlights key historical information using a lightweight memory, while maintaining complete historical information within a universal page-store. 2) \textbf{Researcher}, which retrieves and integrates useful information from the page-store for its online request guided by the pre-constructed memory. This design allows GAM to effectively leverage the agentic capabilities and test-time scalability of frontier large language models (LLMs), while also facilitating end-to-end performance optimization through reinforcement learning. In our experimental study, we demonstrate that GAM achieves substantial improvement on various memory-grounded task completion scenarios against existing memory systems. --- ### NeuroVascU-Net: A Unified Multi-Scale and Cross-Domain Adaptive Feature Fusion U-Net for Precise 3D Segmentation of Brain Vessels in Contrast-Enhanced T1 MRI Authors: Mohammad Jafari Vayeghan, Niloufar Delfan, Mehdi Tale Masouleh, Mansour Parvaresh Rizi, Behzad Moshiri Published: 2025-11-23 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2511.18422v1 Precise 3D segmentation of cerebral vasculature from T1-weighted contrast-enhanced (T1CE) MRI is crucial for safe neurosurgical planning. Manual delineation is time-consuming and prone to inter-observer variability, while current automated methods often trade accuracy for computational cost, limiting clinical use. We present NeuroVascU-Net, the first deep learning architecture specifically designed to segment cerebrovascular structures directly from clinically standard T1CE MRI in neuro-oncology patients, addressing a gap in prior work dominated by TOF-MRA-based approaches. NeuroVascU-Net builds on a dilated U-Net and integrates two specialized modules: a Multi-Scale Contextual Feature Fusion ($MSC^2F$) module at the bottleneck and a Cross-Domain Adaptive Feature Fusion ($CDA^2F$) module at deeper hierarchical layers. $MSC^2F$ captures both local and global information via multi-scale dilated convolutions, while $CDA^2F$ dynamically integrates domain-specific features, enhancing representation while keeping computation low. The model was trained and validated on a curated dataset of T1CE scans from 137 brain tumor biopsy patients, annotated by a board-certified functional neurosurgeon. NeuroVascU-Net achieved a Dice score of 0.8609 and precision of 0.8841, accurately segmenting both major and fine vascular structures. Notably, it requires only 12.4M parameters, significantly fewer than transformer-based models such as Swin U-NetR. This balance of accuracy and efficiency positions NeuroVascU-Net as a practical solution for computer-assisted neurosurgical planning. --- ### DHAuDS: A Dynamic and Heterogeneous Audio Benchmark for Test-Time Adaptation Authors: Weichuang Shao, Iman Yi Liao, Tomas Henrique Bode Maul, Tissa Chandesa Published: 2025-11-23 Categories: cs.SD, cs.LG arXiv: https://arxiv.org/abs/2511.18421v3 Existing Test-time Adaptation (TTA) studies rely heavily on static and homogeneous corruption protocols, such as ImageNet-C and CIFAR-10-C/100-C, leading to inconsistent evaluation settings and potentially inflated robustness estimates that are compared with real-world situations. TTA lacks a standardized evaluation infrastructure capable of modeling realistic heterogeneous acoustic degradation. We introduce DHAuDS, a standardized benchmark suite for evaluating audio classification TTA robustness under dynamic corruption severity and heterogeneous noise mixtures. Rather than proposing a new TTA algorithm, DHAuDS focuses on exposing robustness limitations that remain hidden under conventional fixed-noise evaluation protocols. --- ### DHAuDS: A Dynamic and Heterogeneous Audio Benchmark for Test-Time Adaptation Authors: Weichuang Shao, Iman Yi Liao, Tomas Henrique Bode Maul, Tissa Chandesa Published: 2025-11-23 Categories: cs.SD, cs.LG arXiv: https://arxiv.org/abs/2511.18421v1 Audio classifiers frequently face domain shift, when models trained on one dataset lose accuracy on data recorded in acoustically different conditions. Previous Test-Time Adaptation (TTA) research in speech and sound analysis often evaluates models under fixed or mismatched noise settings, that fail to mimic real-world variability. To overcome these limitations, this paper presents DHAuDS (Dynamic and Heterogeneous Audio Domain Shift), a benchmark designed to assess TTA approaches under more realistic and diverse acoustic shifts. DHAuDS comprises four standardized benchmarks: UrbanSound8K-C, SpeechCommandsV2-C, VocalSound-C, and ReefSet-C, each constructed with dynamic corruption severity levels and heterogeneous noise types to simulate authentic audio degradation scenarios. The framework defines 14 evaluation criteria for each benchmark (8 for UrbanSound8K-C), resulting in 50 unrepeated criteria (124 experiments) that collectively enable fair, reproducible, and cross-domain comparison of TTA algorithms. Through the inclusion of dynamic and mixed-domain noise settings, DHAuDS offers a consistent and publicly reproducible testbed to support ongoing studies in robust and adaptive audio modeling. --- ### Categorical Equivariant Deep Learning: Category-Equivariant Neural Networks and Universal Approximation Theorems Authors: Yoshihiro Maruyama Published: 2025-11-23 Categories: cs.LG, cs.AI, cs.CV, cs.RO arXiv: https://arxiv.org/abs/2511.18417v2 We develop a theory of category-equivariant neural networks (CENNs) that unifies group/groupoid-equivariant networks, poset/lattice-equivariant networks, graph and sheaf neural networks. Equivariance is formulated as naturality in a topological category with Radon measures. Formulating linear and nonlinear layers in the categorical setup, we prove the equivariant universal approximation theorem in the general setting: the class of finite-depth CENNs is dense in the space of continuous equivariant transformations. We instantiate the framework for groups/groupoids, posets/lattices, graphs and cellular sheaves, deriving universal approximation theorems for them in a systematic manner. Categorical equivariant deep learning thus allows us to expand the horizons of equivariant deep learning beyond group actions, encompassing not only geometric symmetries but also contextual and compositional symmetries. --- ### Categorical Equivariant Deep Learning: Category-Equivariant Neural Networks and Universal Approximation Theorems Authors: Yoshihiro Maruyama Published: 2025-11-23 Categories: cs.LG, cs.AI, cs.CV, cs.RO arXiv: https://arxiv.org/abs/2511.18417v1 We develop a theory of category-equivariant neural networks (CENNs) that unifies group/groupoid-equivariant networks, poset/lattice-equivariant networks, graph and sheaf neural networks. Equivariance is formulated as naturality in a topological category with Radon measures, formulating linear and nonlinear layers in the categorical setup. We prove the equivariant universal approximation theorem in the general setting: the class of finite-depth CENNs is dense in the space of continuous equivariant transformations. We instantiate the framework for groups/groupoids, posets/lattices, graphs and cellular sheaves, deriving universal approximation theorems for them in a systematic manner. Categorical equivariant deep learning thus allows us to expand the horizons of equivariant deep learning beyond group actions, encompassing not only geometric symmetries but also contextual and compositional symmetries. --- ### 4D-VGGT: A General Foundation Model with SpatioTemporal Awareness for Dynamic Scene Geometry Estimation Authors: Haonan Wang, Hanyu Zhou, Haoyue Liu, Luxin Yan Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18416v1 We investigate a challenging task of dynamic scene geometry estimation, which requires representing both spatial and temporal features. Typically, existing methods align the two features into a unified latent space to model scene geometry. However, this unified paradigm suffers from potential mismatched representation due to the heterogeneous nature between spatial and temporal features. In this work, we propose 4D-VGGT, a general foundation model with divide-and-conquer spatiotemporal representation for dynamic scene geometry. Our model is divided into three aspects: 1) Multi-setting input. We design an adaptive visual grid that supports input sequences with arbitrary numbers of views and time steps. 2) Multi-level representation. We propose a cross-view global fusion for spatial representation and a cross-time local fusion for temporal representation. 3) Multi-task prediction. We append multiple task-specific heads to spatiotemporal representations, enabling a comprehensive visual geometry estimation for dynamic scenes. Under this unified framework, these components enhance the feature discriminability and application universality of our model for dynamic scenes. In addition, we integrate multiple geometry datasets to train our model and conduct extensive experiments to verify the effectiveness of our method across various tasks on multiple dynamic scene geometry benchmarks. --- ### DuoTeach: Dual Role Self-Teaching for Coarse-to-Fine Decision Coordination in Vision--Language Models Authors: Wei Yang, Yiran Zhu, Zilin Li, Xunjia Zhang, Jun Xia, Hongtao Wang Published: 2025-11-23 Categories: cs.MM, cs.CV arXiv: https://arxiv.org/abs/2511.18415v2 Coarse-to-fine path decision-making requires predicting a valid taxonomy path in which earlier decisions constrain later ones. However, existing benchmarks score each level independently, obscuring cross-level validity and consistency. To better align evaluation with this setting, we introduce a Joint Path Decision (JPD) protocol that requires predicting the full path in one call, together with Depth-Weighted Prefix Accuracy (DWPA), a metric family that measures path reliability with tunable emphasis on deeper levels. Under JPD, strong vision-language models (VLMs) frequently produce invalid parent-child pairs and brittle full-path predictions, suggesting that their failures stem not only from incomplete taxonomic knowledge but also from unstable cross-level decision coordination. To address this problem, we propose DuoTeach, a dual-role self-teaching distillation framework that requires no ground-truth labels and reuses the same pretrained VLM in two roles. Its Decision-Conditioned Rollout (DCR) generates more coherent teacher traces by conditioning each level on prior decisions, and distills this coordinated behavior into the student without additional test-time rollouts. Across multiple taxonomy-structured benchmarks and VLM base models, DuoTeach improves in-domain DWPA (alpha = 0.95) by up to 30.24 points and boosts zero-shot performance on unseen taxonomies from 17.17% to 43.66%. Further analyses attribute these gains to improved within-call multi-level decision coordination. --- ### Self-Empowering VLMs: Achieving Hierarchical Consistency via Self-Elicited Knowledge Distillation Authors: Wei Yang, Yiran Zhu, Zilin Li, Xunjia Zhang, Hongtao Wang Published: 2025-11-23 Categories: cs.MM, cs.CV arXiv: https://arxiv.org/abs/2511.18415v1 Vision-language models (VLMs) possess rich knowledge but often fail on hierarchical understanding tasks, where the goal is to predict a coarse-to-fine taxonomy path that remains consistent across all levels. We compare three inference paradigms for hierarchical VQA and find that stepwise reasoning, when conditioned on prior answers, significantly outperforms single-pass prompting. Further analysis indicates that the main limitation of current VLMs is their inability to maintain cross-level state, rather than a lack of taxonomic knowledge. Motivated by this diagnosis, we propose Self-Elicited Knowledge Distillation (SEKD), which requires no human labels or external tools: the same VLM is prompted to reason step by step and act as a teacher by exposing its hard labels, soft distributions, and decoder hidden states, while a single-pass student distills these signals. The student VLM remains efficient while approaching the accuracy of its multi-step teacher. It improves in-domain path consistency (HCA) by up to +29.50 percentage points, raises zero-shot HCA on an unseen taxonomy from 4.15% to 42.26%, and yields gains on challenging mathematical benchmarks. Because all supervision is self-elicited, SEKD scales to new taxonomies and datasets without annotation cost, providing a practical route to imbue compact VLMs with dependency-aware multi-step reasoning. --- ### Multi-Agent Collaborative Filtering: Orchestrating Users and Items for Agentic Recommendations Authors: Yu Xia, Sungchul Kim, Tong Yu, Ryan A. Rossi, Julian McAuley Published: 2025-11-23 Categories: cs.CL, cs.IR arXiv: https://arxiv.org/abs/2511.18413v3 Agentic recommendations cast recommenders as large language model (LLM) agents that can plan, reason, use tools, and interact with users of varying preferences in web applications. However, most existing agentic recommender systems focus on generic single-agent plan-execute workflows or multi-agent task decomposition pipelines. Without recommendation-oriented design, they often underuse the collaborative signals in the user-item interaction history, leading to unsatisfying recommendation results. To address this, we propose the Multi-Agent Collaborative Filtering (MACF) framework for agentic recommendations, drawing an analogy between traditional collaborative filtering algorithms and LLM-based multi-agent collaboration. Specifically, given a target user and query, we instantiate similar users and relevant items as LLM agents with unique profiles. Each agent is able to call retrieval tools, suggest candidate items, and interact with other agents. Different from the static preference aggregation in traditional collaborative filtering, MACF employs a central orchestrator agent to adaptively manage the collaboration between user and item agents via dynamic agent recruitment and personalized collaboration instruction. Experimental results on datasets from three different domains show the advantages of our MACF framework compared to strong agentic recommendation baselines. --- ### Multi-Agent Collaborative Filtering: Orchestrating Users and Items for Agentic Recommendations Authors: Yu Xia, Sungchul Kim, Tong Yu, Ryan A. Rossi, Julian McAuely Published: 2025-11-23 Categories: cs.CL, cs.IR arXiv: https://arxiv.org/abs/2511.18413v1 Agentic recommendations cast recommenders as large language model (LLM) agents that can plan, reason, use tools, and interact with users of varying preferences in web applications. However, most existing agentic recommender systems focus on generic single-agent plan-execute workflows or multi-agent task decomposition pipelines. Without recommendation-oriented design, they often underuse the collaborative signals in the user-item interaction history, leading to unsatisfying recommendation results. To address this, we propose the Multi-Agent Collaborative Filtering (MACF) framework for agentic recommendations, drawing an analogy between traditional collaborative filtering algorithms and LLM-based multi-agent collaboration. Specifically, given a target user and query, we instantiate similar users and relevant items as LLM agents with unique profiles. Each agent is able to call retrieval tools, suggest candidate items, and interact with other agents. Different from the static preference aggregation in traditional collaborative filtering, MACF employs a central orchestrator agent to adaptively manage the collaboration between user and item agents via dynamic agent recruitment and personalized collaboration instruction. Experimental results on datasets from three different domains show the advantages of our MACF framework compared to strong agentic recommendation baselines. --- ### SmolKalam: Ensemble Quality-Filtered Translation at Scale for High Quality Arabic Post-Training Data Authors: Sultan Alrashed, Chadi Helwe, Francesco Orabona Published: 2025-11-23 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.18411v1 Although the community has tackled the acquisition of high-quality Arabic pretraining data, we still lack large-scale, multi-turn Arabic datasets that include reasoning and tool calling. Naive translation can work at the pretraining scale, but post-training demands much higher quality, which requires a stricter approach to dataset curation. In this work, we introduce SmolKalam, a translation of Smoltalk2 that uses a multi-model ensemble translation pipeline, applies quality filtering, and examines effective translation techniques for traditional decoder-only models through ablations. --- ### Findings of the BlackboxNLP 2025 Shared Task: Localizing Circuits and Causal Variables in Language Models Authors: Dana Arad, Yonatan Belinkov, Hanjie Chen, Najoung Kim, Hosein Mohebbi, Aaron Mueller, Gabriele Sarti, Martin Tutek Published: 2025-11-23 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2511.18409v1 Mechanistic interpretability (MI) seeks to uncover how language models (LMs) implement specific behaviors, yet measuring progress in MI remains challenging. The recently released Mechanistic Interpretability Benchmark (MIB; Mueller et al., 2025) provides a standardized framework for evaluating circuit and causal variable localization. Building on this foundation, the BlackboxNLP 2025 Shared Task extends MIB into a community-wide reproducible comparison of MI techniques. The shared task features two tracks: circuit localization, which assesses methods that identify causally influential components and interactions driving model behavior, and causal variable localization, which evaluates approaches that map activations into interpretable features. With three teams spanning eight different methods, participants achieved notable gains in circuit localization using ensemble and regularization strategies for circuit discovery. With one team spanning two methods, participants achieved significant gains in causal variable localization using low-dimensional and non-linear projections to featurize activation vectors. The MIB leaderboard remains open; we encourage continued work in this standard evaluation framework to measure progress in MI research going forward. --- ### A Multimodal Conversational Agent for Tabular Data Analysis Authors: Mohammad Nour Al Awad, Sergey Ivanov, Olga Tikhonova, Ivan Khodnenko Published: 2025-11-23 Categories: cs.AI, cs.HC, cs.IR arXiv: https://arxiv.org/abs/2511.18405v1 Large language models (LLMs) can reshape information processing by handling data analysis, visualization, and interpretation in an interactive, context-aware dialogue with users, including voice interaction, while maintaining high performance. In this article, we present Talk2Data, a multimodal LLM-driven conversational agent for intuitive data exploration. The system lets users query datasets with voice or text instructions and receive answers as plots, tables, statistics, or spoken explanations. Built on LLMs, the suggested design combines OpenAI Whisper automatic speech recognition (ASR) system, Qwen-coder code generation LLM/model, custom sandboxed execution tools, and Coqui library for text-to-speech (TTS) within an agentic orchestration loop. Unlike text-only analysis tools, it adapts responses across modalities and supports multi-turn dialogues grounded in dataset context. In an evaluation of 48 tasks on three datasets, our prototype achieved 95.8% accuracy with model-only generation time under 1.7 seconds (excluding ASR and execution time). A comparison across five LLM sizes (1.5B-32B) revealed accuracy-latency-cost trade-offs, with a 7B model providing the best balance for interactive use. By routing between conversation with user and code execution, constrained to a transparent sandbox, with simultaneously grounding prompts in schema-level context, the Talk2Data agent reliably retrieves actionable insights from tables while making computations verifiable. In the article, except for the Talk2Data agent itself, we discuss implications for human-data interaction, trust in LLM-driven analytics, and future extensions toward large-scale multimodal assistants. --- ### Pre-training Graph Neural Networks on 2D and 3D Molecular Structures by using Multi-View Conditional Information Bottleneck Authors: Van Thuy Hoang, O-Joun Lee Published: 2025-11-23 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2511.18404v1 Recent pre-training strategies for molecular graphs have attempted to use 2D and 3D molecular views as both inputs and self-supervised signals, primarily aligning graph-level representations. However, existing studies remain limited in addressing two main challenges of multi-view molecular learning: (1) discovering shared information between two views while diminishing view-specific information and (2) identifying and aligning important substructures, e.g., functional groups, which are crucial for enhancing cross-view consistency and model expressiveness. To solve these challenges, we propose a Multi-View Conditional Information Bottleneck framework, called MVCIB, for pre-training graph neural networks on 2D and 3D molecular structures in a self-supervised setting. Our idea is to discover the shared information while minimizing irrelevant features from each view under the MVCIB principle, which uses one view as a contextual condition to guide the representation learning of its counterpart. To enhance semantic and structural consistency across views, we utilize key substructures, e.g., functional groups and ego-networks, as anchors between the two views. Then, we propose a cross-attention mechanism that captures fine-grained correlations between the substructures to achieve subgraph alignment across views. Extensive experiments in four molecular domains demonstrated that MVCIB consistently outperforms baselines in both predictive performance and interpretability. Moreover, MVCIB achieved the 3d Weisfeiler-Lehman expressiveness power to distinguish not only non-isomorphic graphs but also different 3D geometries that share identical 2D connectivity, such as isomers. --- ### ChineseVideoBench: Benchmarking Multi-modal Large Models for Chinese Video Question Answering Authors: Yuxiang Nie, Han Wang, Yongjie Ye, Haiyang Yu, Weitao Jia, Tao Zeng, Hao Feng, Xiang Fei, Yang Li, Xiaohui Lv, Guozhi Tang, Jingqun Tang et al. Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18399v1 This paper introduces ChineseVideoBench, a pioneering benchmark specifically designed for evaluating Multimodal Large Language Models (MLLMs) in Chinese Video Question Answering. The growing demand for sophisticated video analysis capabilities highlights the critical need for comprehensive, culturally-aware evaluation frameworks. ChineseVideoBench addresses this gap by providing a robust dataset and tailored evaluation metrics, enabling rigorous assessment of state-of-the-art MLLMs on complex Chinese video content. Specifically, ChineseVideoBench comprises 8 main classes and 12 sub-classes, encompassing tasks that demand both deep video understanding and nuanced Chinese linguistic and cultural awareness. Our empirical evaluations reveal that ChineseVideoBench presents a significant challenge to current MLLMs. Among the models assessed, Gemini 2.5 Pro achieves the highest performance with an overall score of 77.9%, while InternVL-38B emerges as the most competitive open-source model. --- ### Natural Emergent Misalignment from Reward Hacking in Production RL Authors: Monte MacDiarmid, Benjamin Wright, Jonathan Uesato, Joe Benton, Jon Kutasov, Sara Price, Naia Bouscal, Sam Bowman, Trenton Bricken, Alex Cloud, Carson Denison, Johannes Gasteiger et al. Published: 2025-11-23 Categories: cs.AI, cs.SE arXiv: https://arxiv.org/abs/2511.18397v1 We show that when large language models learn to reward hack on production RL environments, this can result in egregious emergent misalignment. We start with a pretrained model, impart knowledge of reward hacking strategies via synthetic document finetuning or prompting, and train on a selection of real Anthropic production coding environments. Unsurprisingly, the model learns to reward hack. Surprisingly, the model generalizes to alignment faking, cooperation with malicious actors, reasoning about malicious goals, and attempting sabotage when used with Claude Code, including in the codebase for this paper. Applying RLHF safety training using standard chat-like prompts results in aligned behavior on chat-like evaluations, but misalignment persists on agentic tasks. Three mitigations are effective: (i) preventing the model from reward hacking; (ii) increasing the diversity of RLHF safety training; and (iii) "inoculation prompting", wherein framing reward hacking as acceptable behavior during training removes misaligned generalization even when reward hacking is learned. --- ### Exploring Weak-to-Strong Generalization for CLIP-based Classification Authors: Jinhao Li, Sarah M. Erfani, Lei Feng, James Bailey, Feng Liu Published: 2025-11-23 Categories: cs.CV arXiv: https://arxiv.org/abs/2511.18396v1 Aligning large-scale commercial models with user intent is crucial to preventing harmful outputs. Current methods rely on human supervision but become impractical as model complexity increases. When models surpass human knowledge, providing accurate feedback becomes challenging and inefficient. A novel solution proposed recently is using a weaker model to supervise a stronger model. This concept leverages the ability of weaker models to perform evaluations, thereby reducing the workload on human supervisors. Previous work has shown the effectiveness of weak-to-strong generalization in the context of language-only models. Extending this concept to vision-language models leverages these insights, adapting the proven benefits to a multi-modal context. In our study, we explore weak-to-strong generalization for CLIP-based classification. We propose a method, class prototype learning (CPL), which aims to enhance the classification capabilities of the CLIP model, by learning more representative prototypes for each category. Our findings indicate that, despite using a simple loss function under weak supervision, CPL yields robust improvements in targeted scenarios, particularly when pretraining is limited. Extensive experiments demonstrate that our approach is effective under these settings, achieving a 3.67% improvement over strong baseline methods. ---