# Research Papers — 2025-11 · bundle page 37 of 197

> Full-text markdown of papers #1801–1850 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/37
[← Previous bundle](https://aipod.dev/papers/archive/2025-11/page/36.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-11/page/38.md)

### Beyond Components: Singular Vector-Based Interpretability of Transformer Circuits

Authors: Areeb Ahmad, Abhinav Joshi, Ashutosh Modi
Published: 2025-11-25
Categories: cs.LG, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2511.20273v1

Transformer-based language models exhibit complex and distributed behavior, yet their internal computations remain poorly understood. Existing mechanistic interpretability methods typically treat attention heads and multilayer perceptron layers (MLPs) (the building blocks of a transformer architecture) as indivisible units, overlooking possibilities of functional substructure learned within them. In this work, we introduce a more fine-grained perspective that decomposes these components into orthogonal singular directions, revealing superposed and independent computations within a single head or MLP. We validate our perspective on widely used standard tasks like Indirect Object Identification (IOI), Gender Pronoun (GP), and Greater Than (GT), showing that previously identified canonical functional heads, such as the name mover, encode multiple overlapping subfunctions aligned with distinct singular directions. Nodes in a computational graph, that are previously identified as circuit elements show strong activation along specific low-rank directions, suggesting that meaningful computations reside in compact subspaces. While some directions remain challenging to interpret fully, our results highlight that transformer computations are more distributed, structured, and compositional than previously assumed. This perspective opens new avenues for fine-grained mechanistic interpretability and a deeper understanding of model internals.

---

### VKnowU: Evaluating Visual Knowledge Understanding in Multimodal LLMs

Authors: Tianxiang Jiang, Sheng Xia, Yicheng Xu, Linquan Wu, Xiangyu Zeng, Limin Wang, Yu Qiao, Yi Wang
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20272v3

While Multimodal Large Language Models (MLLMs) have become adept at recognizing objects, they often lack the intuitive, human-like understanding of the world's underlying physical and social principles. This high-level vision-grounded semantics, which we term visual knowledge, forms a bridge between perception and reasoning, yet remains an underexplored area in current MLLMs. To systematically evaluate this capability, we present VKnowU, a comprehensive benchmark featuring 1,680 questions in 1,249 videos, covering 8 core types of visual knowledge spanning both world-centric (e.g., intuitive physics) and human-centric (e.g., subjective intentions). Evaluation of 28 SOTA MLLMs reveals that leading models still fall short of human performance, with particularly notable gaps in the world-centric. To bridge this gap, we introduce a new dataset, VKnowQA, and VideoKnow+, a baseline model that explicitly incorporates visual knowledge into MLLMs. VideoKnow+ follows a structured See-Think-Answer paradigm and adopts reinforcement learning with visual knowledge reward, achieving a +3.7% improvement on VKnowU and consistent gains on MVBench (+5.4%), Video-MME (+7.0%), and MMVU (+5.7%). Our work highlights visual knowledge as a missing cornerstone for developing more generalizable MLLMs that can not only see but also truly understand our worlds.

---

### VKnowU: Evaluating Visual Knowledge Understanding in Multimodal LLMs

Authors: Tianxiang Jiang, Sheng Xia, Yicheng Xu, Linquan Wu, Xiangyu Zeng, Limin Wang, Yu Qiao, Yi Wang
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20272v1

While Multimodal Large Language Models (MLLMs) have become adept at recognizing objects, they often lack the intuitive, human-like understanding of the world's underlying physical and social principles. This high-level vision-grounded semantics, which we term visual knowledge, forms a bridge between perception and reasoning, yet remains an underexplored area in current MLLMs. To systematically evaluate this capability, we present VKnowU, a comprehensive benchmark featuring 1,680 questions in 1,249 videos, covering 8 core types of visual knowledge spanning both world-centric (e.g., intuitive physics) and human-centric (e.g., subjective intentions). Evaluation of 23 SOTA MLLMs reveals that leading models still fall short of human performance, with particularly notable gaps in the world-centric. To bridge this gap, we introduce a new dataset, VKnowQA, and VideoKnow+, a baseline model that explicitly incorporates visual knowledge into MLLMs. VideoKnow+ follows a structured See-Think-Answer paradigm and adopts reinforcement learning with visual knowledge reward, achieving a +3.7% improvement on VKnowU and consistent gains on MVBench, Video-MME, and MMVU. Our work highlights visual knowledge as a missing cornerstone for developing more generalizable MLLMs that can not only see but also truly understand our physical and social worlds.

---

### DRL-Guided Neural Batch Sampling for Semi-Supervised Pixel-Level Anomaly Detection

Authors: Amirhossein Khadivi Noghredeh, Abdollah Safari, Fatemeh Ziaeetabar, Firoozeh Haghighi
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20270v2

Anomaly detection in industrial visual inspection is challenging due to the scarcity of defective samples. Most existing methods rely on unsupervised reconstruction using only normal data, often resulting in overfitting and poor detection of subtle defects. We propose a semi-supervised deep reinforcement learning framework that integrates a neural batch sampler, an autoencoder, and a predictor. The RL-based sampler adaptively selects informative patches by balancing exploration and exploitation through a composite reward. The autoencoder generates loss profiles highlighting abnormal regions, while the predictor performs segmentation in the loss-profile space. This interaction enables the system to effectively learn both normal and defective patterns with limited labeled data. Experiments on the MVTec AD dataset demonstrate that our method achieves higher accuracy and better localization of subtle anomalies than recent state-of-the-art approaches while maintaining low complexity, yielding an average improvement of 0.15 in F1_max and 0.06 in AUC, with a maximum gain of 0.37 in F1_max in the best case.

---

### DRL-Guided Neural Batch Sampling for Semi-Supervised Pixel-Level Anomaly Detection

Authors: Amirhossein Khadivi Noghredeh, Abdollah Safari, Fatemeh Ziaeetabar, Firoozeh Haghighi
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20270v1

Anomaly detection in industrial visual inspection is challenging due to the scarcity of defective samples. Most existing methods rely on unsupervised reconstruction using only normal data, often resulting in overfitting and poor detection of subtle defects. We propose a semi-supervised deep reinforcement learning framework that integrates a neural batch sampler, an autoencoder, and a predictor. The RL-based sampler adaptively selects informative patches by balancing exploration and exploitation through a composite reward. The autoencoder generates loss profiles highlighting abnormal regions, while the predictor performs segmentation in the loss-profile space. This interaction enables the system to effectively learn both normal and defective patterns with limited labeled data. Experiments on the MVTec AD dataset demonstrate that our method achieves higher accuracy and better localization of subtle anomalies than recent state-of-the-art approaches while maintaining low complexity, yielding an average improvement of 0.15 in F1_max and 0.06 in AUC, with a maximum gain of 0.37 in F1_max in the best case.

---

### Advancing Image Classification with Discrete Diffusion Classification Modeling

Authors: Omer Belhasin, Shelly Golan, Ran El-Yaniv, Michael Elad
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20263v1

Image classification is a well-studied task in computer vision, and yet it remains challenging under high-uncertainty conditions, such as when input images are corrupted or training data are limited. Conventional classification approaches typically train models to directly predict class labels from input images, but this might lead to suboptimal performance in such scenarios. To address this issue, we propose Discrete Diffusion Classification Modeling (DiDiCM), a novel framework that leverages a diffusion-based procedure to model the posterior distribution of class labels conditioned on the input image. DiDiCM supports diffusion-based predictions either on class probabilities or on discrete class labels, providing flexibility in computation and memory trade-offs. We conduct a comprehensive empirical study demonstrating the superior performance of DiDiCM over standard classifiers, showing that a few diffusion iterations achieve higher classification accuracy on the ImageNet dataset compared to baselines, with accuracy gains increasing as the task becomes more challenging. We release our code at https://github.com/omerb01/didicm .

---

### Modality-Balanced Collaborative Distillation for Multi-Modal Domain Generalization

Authors: Xiaohan Wang, Zhangtao Cheng, Ting Zhong, Leiting Chen, Fan Zhou
Published: 2025-11-25
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2511.20258v3

Weight Averaging (WA) has emerged as a powerful technique for enhancing generalization by promoting convergence to a flat loss landscape, which correlates with stronger out-of-distribution performance. However, applying WA directly to multi-modal domain generalization (MMDG) is challenging: differences in optimization speed across modalities lead WA to overfit to faster-converging ones in early stages, suppressing the contribution of slower yet complementary modalities, thereby hindering effective modality fusion and skewing the loss surface toward sharper, less generalizable minima. To address this issue, we propose MBCD, a unified collaborative distillation framework that retains WA's flatness-inducing advantages while overcoming its shortcomings in multi-modal contexts. MBCD begins with adaptive modality dropout in the student model to curb early-stage bias toward dominant modalities. A gradient consistency constraint then aligns learning signals between uni-modal branches and the fused representation, encouraging coordinated and smoother optimization. Finally, a WA-based teacher conducts cross-modal distillation by transferring fused knowledge to each uni-modal branch, which strengthens cross-modal interactions and steer convergence toward flatter solutions. Extensive experiments on MMDG benchmarks show that MBCD consistently outperforms existing methods, achieving superior accuracy and robustness across diverse unseen domains.

---

### Modality-Balanced Collaborative Distillation for Multi-Modal Domain Generalization

Authors: Xiaohan Wang, Zhangtao Cheng, Ting Zhong, Leiting Chen, Fan Zhou
Published: 2025-11-25
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2511.20258v1

Weight Averaging (WA) has emerged as a powerful technique for enhancing generalization by promoting convergence to a flat loss landscape, which correlates with stronger out-of-distribution performance. However, applying WA directly to multi-modal domain generalization (MMDG) is challenging: differences in optimization speed across modalities lead WA to overfit to faster-converging ones in early stages, suppressing the contribution of slower yet complementary modalities, thereby hindering effective modality fusion and skewing the loss surface toward sharper, less generalizable minima. To address this issue, we propose MBCD, a unified collaborative distillation framework that retains WA's flatness-inducing advantages while overcoming its shortcomings in multi-modal contexts. MBCD begins with adaptive modality dropout in the student model to curb early-stage bias toward dominant modalities. A gradient consistency constraint then aligns learning signals between uni-modal branches and the fused representation, encouraging coordinated and smoother optimization. Finally, a WA-based teacher conducts cross-modal distillation by transferring fused knowledge to each uni-modal branch, which strengthens cross-modal interactions and steer convergence toward flatter solutions. Extensive experiments on MMDG benchmarks show that MBCD consistently outperforms existing methods, achieving superior accuracy and robustness across diverse unseen domains.

---

### Interpretable Air Pollution Forecasting by Physics-Guided Spatiotemporal Decoupling

Authors: Zhiguo Zhang, Xiaoliang Ma, Daniel Schlesinger
Published: 2025-11-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2511.20257v2

Accurate and interpretable air pollution forecasting is crucial for public health, but most models face a trade-off between performance and interpretability. This study proposes a physics-guided, interpretable-by-design spatiotemporal learning framework. The model decomposes the spatiotemporal behavior of air pollutant concentrations into two transparent, additive modules. The first is a physics-guided transport kernel with directed weights conditioned on wind and geography (advection). The second is an explainable attention mechanism that learns local responses and attributes future concentrations to specific historical lags and exogenous drivers. Evaluated on a comprehensive dataset from the Stockholm region, our model consistently outperforms state-of-the-art baselines across multiple forecasting horizons. Our model's integration of high predictive performance and spatiotemporal interpretability provides a more reliable foundation for operational air-quality management in real-world applications.

---

### Interpretable Air Pollution Forecasting by Physics-Guided Spatiotemporal Decoupling

Authors: Zhiguo Zhang, Xiaoliang Ma, Daniel Schlesinger
Published: 2025-11-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2511.20257v1

Accurate and interpretable air pollution forecasting is crucial for public health, but most models face a trade-off between performance and interpretability. This study proposes a physics-guided, interpretable-by-design spatiotemporal learning framework. The model decomposes the spatiotemporal behavior of air pollutant concentrations into two transparent, additive modules. The first is a physics-guided transport kernel with directed weights conditioned on wind and geography (advection). The second is an explainable attention mechanism that learns local responses and attributes future concentrations to specific historical lags and exogenous drivers. Evaluated on a comprehensive dataset from the Stockholm region, our model consistently outperforms state-of-the-art baselines across multiple forecasting horizons. Our model's integration of high predictive performance and spatiotemporal interpretability provides a more reliable foundation for operational air-quality management in real-world applications.

---

### The Image as Its Own Reward: Reinforcement Learning with Adversarial Reward for Image Generation

Authors: Weijia Mao, Hao Chen, Zhenheng Yang, Mike Zheng Shou
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20256v1

A reliable reward function is essential for reinforcement learning (RL) in image generation. Most current RL approaches depend on pre-trained preference models that output scalar rewards to approximate human preferences. However, these rewards often fail to capture human perception and are vulnerable to reward hacking, where higher scores do not correspond to better images. To address this, we introduce Adv-GRPO, an RL framework with an adversarial reward that iteratively updates both the reward model and the generator. The reward model is supervised using reference images as positive samples and can largely avoid being hacked. Unlike KL regularization that constrains parameter updates, our learned reward directly guides the generator through its visual outputs, leading to higher-quality images. Moreover, while optimizing existing reward functions can alleviate reward hacking, their inherent biases remain. For instance, PickScore may degrade image quality, whereas OCR-based rewards often reduce aesthetic fidelity. To address this, we take the image itself as a reward, using reference images and vision foundation models (e.g., DINO) to provide rich visual rewards. These dense visual signals, instead of a single scalar, lead to consistent gains across image quality, aesthetics, and task-specific metrics. Finally, we show that combining reference samples with foundation-model rewards enables distribution transfer and flexible style customization. In human evaluation, our method outperforms Flow-GRPO and SD3, achieving 70.0% and 72.4% win rates in image quality and aesthetics, respectively. Code and models have been released.

---

### XiCAD: Camera Activation Detection in the Da Vinci Xi User Interface

Authors: Alexander C. Jenke, Gregor Just, Claas de Boer, Martin Wagner, Sebastian Bodenstedt, Stefanie Speidel
Published: 2025-11-25
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2511.20254v1

Purpose: Robot-assisted minimally invasive surgery relies on endoscopic video as the sole intraoperative visual feedback. The DaVinci Xi system overlays a graphical user interface (UI) that indicates the state of each robotic arm, including the activation of the endoscope arm. Detecting this activation provides valuable metadata such as camera movement information, which can support downstream surgical data science tasks including tool tracking, skill assessment, or camera control automation. Methods: We developed a lightweight pipeline based on a ResNet18 convolutional neural network to automatically identify the position of the camera tile and its activation state within the DaVinci Xi UI. The model was fine-tuned on manually annotated data from the SurgToolLoc dataset and evaluated across three public datasets comprising over 70,000 frames. Results: The model achieved F1-scores between 0.993 and 1.000 for the binary detection of active cameras and correctly localized the camera tile in all cases without false multiple-camera detections. Conclusion: The proposed pipeline enables reliable, real-time extraction of camera activation metadata from surgical videos, facilitating automated preprocessing and analysis for diverse downstream applications. All code, trained models, and annotations are publicly available.

---

### Zoo3D: Zero-Shot 3D Object Detection at Scene Level

Authors: Andrey Lemeshko, Bulat Gabdullin, Nikita Drozdov, Anton Konushin, Danila Rukhovich, Maksim Kolodiazhnyi
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20253v1

3D object detection is fundamental for spatial understanding. Real-world environments demand models capable of recognizing diverse, previously unseen objects, which remains a major limitation of closed-set methods. Existing open-vocabulary 3D detectors relax annotation requirements but still depend on training scenes, either as point clouds or images. We take this a step further by introducing Zoo3D, the first training-free 3D object detection framework. Our method constructs 3D bounding boxes via graph clustering of 2D instance masks, then assigns semantic labels using a novel open-vocabulary module with best-view selection and view-consensus mask generation. Zoo3D operates in two modes: the zero-shot Zoo3D$_0$, which requires no training at all, and the self-supervised Zoo3D$_1$, which refines 3D box prediction by training a class-agnostic detector on Zoo3D$_0$-generated pseudo labels. Furthermore, we extend Zoo3D beyond point clouds to work directly with posed and even unposed images. Across ScanNet200 and ARKitScenes benchmarks, both Zoo3D$_0$ and Zoo3D$_1$ achieve state-of-the-art results in open-vocabulary 3D object detection. Remarkably, our zero-shot Zoo3D$_0$ outperforms all existing self-supervised methods, hence demonstrating the power and adaptability of training-free, off-the-shelf approaches for real-world 3D understanding. Code is available at https://github.com/col14m/zoo3d .

---

### Diversifying Long Prompt Image Generation through Structured Prompt Embedding Space Sampling

Authors: Bo-Kai Ruan, Teng-Fang Hsiao, Ling Lo, Yi-Lun Wu, Hong-Han Shuai
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20251v2

Modern text-to-image models produce impressive visual results from richly specified prompts, yet their behavior under long prompts remains insufficiently understood. In this paper, we study a practical failure mode in which accumulated semantic constraints progressively suppress output variation, causing diversity to collapse even when many visual factors remain unspecified. We show that this phenomenon appears consistently across recent generation models as prompt length increases, and provide a theoretical motivation that connects long-prompt conditioning with reduced sampling entropy in the prompt embedding space. Based on this observation, we introduce PromptMoG, a training-free approach that samples prompt embeddings from a Mixture-of-Gaussians distribution to restore generative flexibility while maintaining semantic fidelity. To support systematic evaluation, we further present LPD-Bench, a structured benchmark of long and semantically dense prompts for measuring both fidelity and diversity under compositional text conditioning. Extensive experiments on four large-scale diffusion models, including SD3.5-Large, Flux.1-Krea-Dev, CogView4, and Qwen-Image, show that PromptMoG consistently improves diversity for long-prompt image generation. The code is publicly available at https://github.com/basiclab/PromptMoG.

---

### PromptMoG: Enhancing Diversity in Long-Prompt Image Generation via Prompt Embedding Mixture-of-Gaussian Sampling

Authors: Bo-Kai Ruan, Teng-Fang Hsiao, Ling Lo, Yi-Lun Wu, Hong-Han Shuai
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20251v1

Recent advances in text-to-image (T2I) generation have achieved remarkable visual outcomes through large-scale rectified flow models. However, how these models behave under long prompts remains underexplored. Long prompts encode rich content, spatial, and stylistic information that enhances fidelity but often suppresses diversity, leading to repetitive and less creative outputs. In this work, we systematically study this fidelity-diversity dilemma and reveal that state-of-the-art models exhibit a clear drop in diversity as prompt length increases. To enable consistent evaluation, we introduce LPD-Bench, a benchmark designed for assessing both fidelity and diversity in long-prompt generation. Building on our analysis, we develop a theoretical framework that increases sampling entropy through prompt reformulation and propose a training-free method, PromptMoG, which samples prompt embeddings from a Mixture-of-Gaussians in the embedding space to enhance diversity while preserving semantics. Extensive experiments on four state-of-the-art models, SD3.5-Large, Flux.1-Krea-Dev, CogView4, and Qwen-Image, demonstrate that PromptMoG consistently improves long-prompt generation diversity without semantic drifting.

---

### Uplifting Table Tennis: A Robust, Real-World Application for 3D Trajectory and Spin Estimation

Authors: Daniel Kienzle, Katja Ludwig, Julian Lorenz, Shin'ichi Satoh, Rainer Lienhart
Published: 2025-11-25
Categories: cs.CV, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2511.20250v1

Obtaining the precise 3D motion of a table tennis ball from standard monocular videos is a challenging problem, as existing methods trained on synthetic data struggle to generalize to the noisy, imperfect ball and table detections of the real world. This is primarily due to the inherent lack of 3D ground truth trajectories and spin annotations for real-world video. To overcome this, we propose a novel two-stage pipeline that divides the problem into a front-end perception task and a back-end 2D-to-3D uplifting task. This separation allows us to train the front-end components with abundant 2D supervision from our newly created TTHQ dataset, while the back-end uplifting network is trained exclusively on physically-correct synthetic data. We specifically re-engineer the uplifting model to be robust to common real-world artifacts, such as missing detections and varying frame rates. By integrating a ball detector and a table keypoint detector, our approach transforms a proof-of-concept uplifting method into a practical, robust, and high-performing end-to-end application for 3D table tennis trajectory and spin analysis.

---

### HistoSpeckle-Net: Mutual Information-Guided Deep Learning for high-fidelity reconstruction of complex OrganAMNIST images via perturbed Multimode Fibers

Authors: Jawaria Maqbool, M. Imran Cheema
Published: 2025-11-25
Categories: cs.CV, physics.optics
arXiv: https://arxiv.org/abs/2511.20245v1

Existing deep learning methods in multimode fiber (MMF) imaging often focus on simpler datasets, limiting their applicability to complex, real-world imaging tasks. These models are typically data-intensive, a challenge that becomes more pronounced when dealing with diverse and complex images. In this work, we propose HistoSpeckle-Net, a deep learning architecture designed to reconstruct structurally rich medical images from MMF speckles. To build a clinically relevant dataset, we develop an optical setup that couples laser light through a spatial light modulator (SLM) into an MMF, capturing output speckle patterns corresponding to input OrganAMNIST images. Unlike previous MMF imaging approaches, which have not considered the underlying statistics of speckles and reconstructed images, we introduce a distribution-aware learning strategy. We employ a histogram-based mutual information loss to enhance model robustness and reduce reliance on large datasets. Our model includes a histogram computation unit that estimates smooth marginal and joint histograms for calculating mutual information loss. It also incorporates a unique Three-Scale Feature Refinement Module, which leads to multiscale Structural Similarity Index Measure (SSIM) loss computation. Together, these two loss functions enhance both the structural fidelity and statistical alignment of the reconstructed images. Our experiments on the complex OrganAMNIST dataset demonstrate that HistoSpeckle-Net achieves higher fidelity than baseline models such as U-Net and Pix2Pix. It gives superior performance even with limited training samples and across varying fiber bending conditions. By effectively reconstructing complex anatomical features with reduced data and under fiber perturbations, HistoSpeckle-Net brings MMF imaging closer to practical deployment in real-world clinical environments.

---

### Reinforcement Learning to Initialize Newton-Raphson for AC Power Flow with Quantum Annealing-Based Environment Updates

Authors: Zeynab Kaseb, Matthias Moller, Lindsay Spoor, Jerry J. Guo, Yu Xiang, Peter Palensky, Pedro P. Vergara
Published: 2025-11-25
Categories: eess.SY, cs.ET, cs.LG
arXiv: https://arxiv.org/abs/2511.20237v2

The Newton-Raphson (NR) method is widely used for solving power flow (PF) equations due to its quadratic convergence. However, its performance deteriorates under poor initialization or extreme operating scenarios, e.g., high levels of renewable energy penetration. We propose the use of reinforcement learning (RL) to optimize the initialization of NR, and introduce a quantum-enhanced RL environment update mechanism that addresses the combinatorially large action space at each RL timestep by formulating the voltage adjustment task as a Quadratic Unconstrained Binary Optimization (QUBO) problem, solved with an Ising machine. RL initialization is benchmarked against flat start and start from the DC (linearized) PF solution on a standard 4-bus system, Iwamoto's ill-conditioned 11-bus system, and the IEEE 118-bus system under normal and stressed loading and reactive power limits, with verified operational solutions. On all systems, a supervised initializer refined by RL requires fewer NR iterations than flat and DC starts and than the same initializer without RL, for all seeds. For example, on the 118-bus system under normal and stressed loading, it reached 2.04 and 2.86 NR iterations, compared with 3.02 and 5.13 from DC start and 2.61 and 3.09 without RL. In wall-clock time, this pays off only for an initializer integrated into the solver and reused for many solves on a fixed topology. On the 4-bus system, a quantum-enhanced RL agent with a quantum-inspired annealer moved challenging initial states that required 29 and 44 NR iterations to initializations that required three NR iterations within one RL timestep.

---

### Quantum-Enhanced Reinforcement Learning for Accelerating Newton-Raphson Convergence with Ising Machines: A Case Study for Power Flow Analysis

Authors: Zeynab Kaseb, Matthias Moller, Lindsay Spoor, Jerry J. Guo, Yu Xiang, Peter Palensky, Pedro P. Vergara
Published: 2025-11-25
Categories: eess.SY, cs.ET, cs.LG
arXiv: https://arxiv.org/abs/2511.20237v1

The Newton-Raphson (NR) method is widely used for solving power flow (PF) equations due to its quadratic convergence. However, its performance deteriorates under poor initialization or extreme operating scenarios, e.g., high levels of renewable energy penetration. Traditional NR initialization strategies often fail to address these challenges, resulting in slow convergence or even divergence. We propose the use of reinforcement learning (RL) to optimize the initialization of NR, and introduce a novel quantum-enhanced RL environment update mechanism to mitigate the significant computational cost of evaluating power system states over a combinatorially large action space at each RL timestep by formulating the voltage adjustment task as a quadratic unconstrained binary optimization problem. Specifically, quantum/digital annealers are integrated into the RL environment update to evaluate state transitions using a problem Hamiltonian designed for PF. Results demonstrate significant improvements in convergence speed, a reduction in NR iteration counts, and enhanced robustness under different operating conditions.

---

### Actionable and diverse counterfactual explanations incorporating domain knowledge and plausibility constraints

Authors: Szymon Bobek, Łukasz Bałec, Grzegorz J. Nalepa
Published: 2025-11-25
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2511.20236v3

Counterfactual explanations improve the actionable interpretability of machine learning models by identifying minimal changes required to achieve a desired outcome. However, existing methods often neglect dependencies among features, which can lead to unrealistic or impractical modifications. This limitation reduces the usefulness of counterfactual explanations in real-world decision-support systems. Motivated by applications in cybersecurity for email marketing, we propose DANCE (Diverse, Actionable, and Knowledge-Constrained Explanations), a method for generating counterfactuals that incorporate feature dependencies and domain constraints. DANCE models relationships between features using linear and probabilistic structures that can be learned from data or specified by experts. These dependencies are enforced during the search process to improve plausibility and feasibility. The method jointly optimizes plausibility, diversity, proximity, and sparsity within a unified objective. We evaluate DANCE on 140 datasets from OpenML and demonstrate that it achieves competitive or superior performance compared to existing approaches across multiple evaluation criteria. Additionally, we validate the method in a real-world industrial setting in collaboration with an email marketing platform, showing that it produces domain-consistent and actionable recommendations.

---

### Actionable and diverse counterfactual explanations incorporating domain knowledge and causal constraints

Authors: Szymon Bobek, Łukasz Bałec, Grzegorz J. Nalepa
Published: 2025-11-25
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2511.20236v1

Counterfactual explanations enhance the actionable interpretability of machine learning models by identifying the minimal changes required to achieve a desired outcome of the model. However, existing methods often ignore the complex dependencies in real-world datasets, leading to unrealistic or impractical modifications. Motivated by cybersecurity applications in the email marketing domain, we propose a method for generating Diverse, Actionable, and kNowledge-Constrained Explanations (DANCE), which incorporates feature dependencies and causal constraints to ensure plausibility and real-world feasibility of counterfactuals. Our method learns linear and nonlinear constraints from data or integrates expert-provided dependency graphs, ensuring counterfactuals are plausible and actionable. By maintaining consistency with feature relationships, the method produces explanations that align with real-world constraints. Additionally, it balances plausibility, diversity, and sparsity, effectively addressing key limitations in existing algorithms. The work is developed based on a real-life case study with Freshmail, the largest email marketing company in Poland and supported by a joint R&D project Sendguard. Furthermore, we provide an extensive evaluation using 140 public datasets, which highlights its ability to generate meaningful, domain-relevant counterfactuals that outperform other existing approaches based on widely used metrics. The source code for reproduction of the results can be found in a GitHub repository we provide.

---

### Leveraging weights signals -- Predicting and improving generalizability in reinforcement learning

Authors: Olivier Moulin, Vincent Francois-lavet, Paul Elbers, Mark Hoogendoorn
Published: 2025-11-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2511.20234v1

Generalizability of Reinforcement Learning (RL) agents (ability to perform on environments different from the ones they have been trained on) is a key problem as agents have the tendency to overfit to their training environments. In order to address this problem and offer a solution to increase the generalizability of RL agents, we introduce a new methodology to predict the generalizability score of RL agents based on the internal weights of the agent's neural networks. Using this prediction capability, we propose some changes in the Proximal Policy Optimization (PPO) loss function to boost the generalization score of the agents trained with this upgraded version. Experimental results demonstrate that our improved PPO algorithm yields agents with stronger generalizability compared to the original version.

---

### REFLEX: Self-Refining Explainable Fact-Checking via Verdict-Anchored Style Control

Authors: Chuyi Kong, Wei Gao, Jing Ma, Hongzhan Lin, Yuxi Sun
Published: 2025-11-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2511.20233v5

The prevalence of fake news on social media demands automated fact-checking systems to provide accurate verdicts with faithful explanations. However, existing large language model (LLM)-based approaches ignore deceptive misinformation styles in LLM-generated explanations, resulting in unfaithful rationales that can mislead human judgments. They rely heavily on external knowledge sources, introducing hallucinations and even high latency that undermine reliability and responsiveness, which is crucial for real-time use. To address these challenges, we propose REason-guided Fact-checking with Latent EXplanations (REFLEX), a self-refining paradigm that explicitly controls reasoning style anchored on verdict. REFLEX utilizes self-disagreement veracity signals between the backbone model and its fine-tuned variant to construct steering vectors, naturally disentangling fact from style. Experiments on the real-world dataset show REFLEX achieves state-of-the-art performance under LLaMA-series models with only 465 self-refined samples. Moreover, owing to its transferability, REFLEX yields up to a 7.54% gain on in-the-wild data. Our results further demonstrate that our method effectively mitigates faithful hallucination, thereby guiding the model toward more accurate verdicts than previous works in explainable fact-checking.

---

### REFLEX: Self-Refining Explainable Fact-Checking via Disentangling Truth into Style and Substance

Authors: Chuyi Kong, Gao Wei, Jing Ma, Hongzhan Lin, Zhiyuan Fan
Published: 2025-11-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2511.20233v1

The prevalence of misinformation on social media threatens public trust, demanding automated fact-checking systems that provide accurate verdicts with interpretable explanations. However, existing large language model-based (LLM-based) approaches often rely heavily on external knowledge sources, introducing substantial latency and even hallucinations that undermine reliability, interpretability, and responsiveness, which is crucial for real-time use. To address these challenges, we propose REason-guided Fact-checking with Latent EXplanations REFLEX paradigm, a plug-and-play, self-refining paradigm that leverages the internal knowledge in backbone model to improve both verdict accuracy and explanation quality. REFLEX reformulates fact-checking as a role-play dialogue and jointly trains verdict prediction and explanation generation. It adaptively extracts contrastive activation pairs between the backbone model and its fine-tuned variant to construct steering vectors that disentangle truth into style and substance naturally. These activation-level signals guide inference and suppress noisy explanations, enabling more faithful and efficient reasoning. Experiments on real-world datasets show that REFLEX outperforms previous methods that steer toward a single truth direction and underscores the challenge traditional approaches face when handling the subtle, human-unknown truth in fact-checking tasks. Remarkably, with only 465 self-refined training samples, RELFEX achieves state-of-the-art performance. Furthermore, models trained with explanatory objectives can effectively guide those without them, yielding up to a 7.57% improvement, highlighting that internal explanation signals play a dual role in both interpreting and enhancing factual reasoning.

---

### DiCaP: Distribution-Calibrated Pseudo-labeling for Semi-Supervised Multi-Label Learning

Authors: Bo Han, Zhuoming Li, Xiaoyu Wang, Yaxin Hou, Hui Liu, Junhui Hou, Yuheng Jia
Published: 2025-11-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2511.20225v2

Semi-supervised multi-label learning (SSMLL) aims to address the challenge of limited labeled data in multi-label learning (MLL) by leveraging unlabeled data to improve the model's performance. While pseudo-labeling has become a dominant strategy in SSMLL, most existing methods assign equal weights to all pseudo-labels regardless of their quality, which can amplify the impact of noisy or uncertain predictions and degrade the overall performance. In this paper, we theoretically verify that the optimal weight for a pseudo-label should reflect its correctness likelihood. Empirically, we observe that on the same dataset, the correctness likelihood distribution of unlabeled data remains stable, even as the number of labeled training samples varies. Building on this insight, we propose Distribution-Calibrated Pseudo-labeling (DiCaP), a correctness-aware framework that estimates posterior precision to calibrate pseudo-label weights. We further introduce a dual-thresholding mechanism to separate confident and ambiguous regions: confident samples are pseudo-labeled and weighted accordingly, while ambiguous ones are explored by unsupervised contrastive learning. Experiments conducted on multiple benchmark datasets verify that our method achieves consistent improvements, surpassing state-of-the-art methods by up to 4.27%.

---

### DiCaP: Distribution-Calibrated Pseudo-labeling for Semi-Supervised Multi-Label Learning

Authors: Bo Han, Zhuoming Li, Xiaoyu Wang, Yaxin Hou, Hui Liu, Junhui Hou, Yuheng Jia
Published: 2025-11-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2511.20225v1

Semi-supervised multi-label learning (SSMLL) aims to address the challenge of limited labeled data in multi-label learning (MLL) by leveraging unlabeled data to improve the model's performance. While pseudo-labeling has become a dominant strategy in SSMLL, most existing methods assign equal weights to all pseudo-labels regardless of their quality, which can amplify the impact of noisy or uncertain predictions and degrade the overall performance. In this paper, we theoretically verify that the optimal weight for a pseudo-label should reflect its correctness likelihood. Empirically, we observe that on the same dataset, the correctness likelihood distribution of unlabeled data remains stable, even as the number of labeled training samples varies. Building on this insight, we propose Distribution-Calibrated Pseudo-labeling (DiCaP), a correctness-aware framework that estimates posterior precision to calibrate pseudo-label weights. We further introduce a dual-thresholding mechanism to separate confident and ambiguous regions: confident samples are pseudo-labeled and weighted accordingly, while ambiguous ones are explored by unsupervised contrastive learning. Experiments conducted on multiple benchmark datasets verify that our method achieves consistent improvements, surpassing state-of-the-art methods by up to 4.27%.

---

### DuoTok: Source-Aware Dual-Track Music Tokenization for Vocal-Accompaniment Generation

Authors: Rui Lin, Zhiyue Wu, Jiahe Lei, Kangdi Wang, Weixiong Chen, Junyu Dai, Tao Jiang
Published: 2025-11-25
Categories: cs.SD, cs.AI
arXiv: https://arxiv.org/abs/2511.20224v3

Multi-track music generation requires tokens that preserve acoustic fidelity, support sequence modeling, and maintain cross-track structure. Reconstruction-oriented codecs retain acoustic detail but are difficult to model, while semantic tokenizers may sacrifice fidelity or cross-track alignment. We present DuoTok, a source-aware dual-track music tokenizer for vocal-accompaniment generation based on staged disentanglement. DuoTok first learns a semantic audio representation through self-supervised pretraining, then shapes source-aware structure using feature replacement noise and multi-task supervision: spectral reconstruction, music source separation regularization, and an ASR head for lyric alignment. It freezes the encoder and learns hard-routed codebooks for vocals and accompaniment, while a diffusion decoder restores fine acoustic detail from discrete tokens. Across public benchmarks, DuoTok achieves a favorable predictability-fidelity trade-off at ultra-low bitrate. Under held-constant dual-track language modeling, it improves both unconditional vocal-accompaniment modeling and vocal-conditioned accompaniment prediction. Controlled diagnostics show larger predictability costs under cross-track corruption and greater gains from longer temporal context, supporting stronger use of cross-track and temporal structure rather than merely easier local prediction. DuoTok also maintains competitive reconstruction quality and preserves control-relevant musical attributes in its discrete space. These results suggest that tokenizer design is a core modeling problem for multi-track music generation, beyond compression alone.

---

### DUO-TOK: Dual-Track Semantic Music Tokenizer for Vocal-Accompaniment Generation

Authors: Rui Lin, Zhiyue Wu, Jiahe Le, Kangdi Wang, Weixiong Chen, Junyu Dai, Tao Jiang
Published: 2025-11-25
Categories: cs.SD, cs.AI
arXiv: https://arxiv.org/abs/2511.20224v1

Duo-Tok is a source-aware dual-codebook tokenizer for vocal-accompaniment music that targets the growing tension between reconstruction quality and language-model (LM) learnability in modern lyrics-to-song systems. Existing codecs either prioritize high-fidelity reconstruction with difficult-to-model acoustic tokens or compress aggressively into semantic tokens that are LM-friendly but lossy, and they rarely make the tokenizer itself aware of dual-track structure. Duo-Tok follows a four-stage, SSL-centered pipeline: we first pretrain a BEST-RQ-style encoder on large-scale audio, then stabilize and factorize the representation with Gaussian replacement noise and multi-task supervision, before freezing the encoder to learn SimVQ-based dual codebooks with hard routing for vocals and accompaniment, and finally training latent diffusion decoders on top of the discrete tokens. Duo-Tok at 0.75 kbps shifts the empirical reconstruction-generation Pareto frontier, achieving the best music-tagging AP and the lowest vocabulary-normalized LM perplexity among compared codecs while maintaining reconstruction quality comparable to state-of-the-art music tokenizers.

---

### V-Attack: Targeting Disentangled Value Features for Controllable Adversarial Attacks on LVLMs

Authors: Sen Nie, Jie Zhang, Jianxin Yan, Shiguang Shan, Xilin Chen
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20223v2

Adversarial attacks have evolved from simply disrupting predictions on conventional task-specific models to the more complex goal of manipulating image semantics on Large Vision-Language Models (LVLMs). However, existing methods struggle with controllability and fail to precisely manipulate the semantics of specific concepts in the image. We attribute this limitation to semantic entanglement in the patch-token representations on which adversarial attacks typically operate: global context aggregated by self-attention in the vision encoder dominates individual patch features, making them unreliable handles for precise local semantic manipulation. Our systematic investigation reveals a key insight: value features (V) computed within the transformer attention block serve as much more precise handles for manipulation. We show that V suppresses global-context channels, allowing it to retain high-entropy, disentangled local semantic information. Building on this discovery, we propose V-Attack, a novel method designed for precise local semantic attacks. V-Attack targets the value features and introduces two core components: (1) a Self-Value Enhancement module to refine V's intrinsic semantic richness, and (2) a Text-Guided Value Manipulation module that leverages text prompts to locate source concept and optimize it toward a target concept. By bypassing the entangled patch features, V-Attack achieves highly effective semantic control. Extensive experiments across diverse LVLMs, including LLaVA, InternVL, DeepseekVL and GPT-4o, show that V-Attack improves the attack success rate by an average of 36% over state-of-the-art methods, exposing critical vulnerabilities in modern visual-language understanding. Our code and data are available https://github.com/Summu77/V-Attack.

---

### V-Attack: Targeting Disentangled Value Features for Controllable Adversarial Attacks on LVLMs

Authors: Sen Nie, Jie Zhang, Jianxin Yan, Shiguang Shan, Xilin Chen
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20223v1

Adversarial attacks have evolved from simply disrupting predictions on conventional task-specific models to the more complex goal of manipulating image semantics on Large Vision-Language Models (LVLMs). However, existing methods struggle with controllability and fail to precisely manipulate the semantics of specific concepts in the image. We attribute this limitation to semantic entanglement in the patch-token representations on which adversarial attacks typically operate: global context aggregated by self-attention in the vision encoder dominates individual patch features, making them unreliable handles for precise local semantic manipulation. Our systematic investigation reveals a key insight: value features (V) computed within the transformer attention block serve as much more precise handles for manipulation. We show that V suppresses global-context channels, allowing it to retain high-entropy, disentangled local semantic information. Building on this discovery, we propose V-Attack, a novel method designed for precise local semantic attacks. V-Attack targets the value features and introduces two core components: (1) a Self-Value Enhancement module to refine V's intrinsic semantic richness, and (2) a Text-Guided Value Manipulation module that leverages text prompts to locate source concept and optimize it toward a target concept. By bypassing the entangled patch features, V-Attack achieves highly effective semantic control. Extensive experiments across diverse LVLMs, including LLaVA, InternVL, DeepseekVL and GPT-4o, show that V-Attack improves the attack success rate by an average of 36% over state-of-the-art methods, exposing critical vulnerabilities in modern visual-language understanding. Our code and data are available https://github.com/Summu77/V-Attack.

---

### Decoupling and Damping: Structurally-Regularized Gradient Matching for Multimodal Graph Condensation

Authors: Lian Shen, Zhendan Chen, Meijia Song, Yinhui jiang, Ziming Su, Juan Liu, Xiangrong Liu
Published: 2025-11-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2511.20222v2

In multimodal graph learning, graph structures that integrate information from multiple sources, such as vision and text, can more comprehensively model complex entity relationships. However, the continuous growth of their data scale poses a significant computational bottleneck for training. Graph condensation methods provide a feasible path forward by synthesizing compact and representative datasets. Nevertheless, existing condensation approaches generally suffer from performance limitations in multimodal scenarios, mainly due to two reasons: (1) semantic misalignment between different modalities leads to gradient conflicts; (2) the message-passing mechanism of graph neural networks further structurally amplifies such gradient noise. Based on this, we propose Structural Regularized Gradient Matching (SR-GM), a condensation framework for multimodal graphs. This method alleviates gradient conflicts between modalities through a gradient decoupling mechanism and introduces a structural damping regularizer to suppress the propagation of gradient noise in the topology, thereby transforming the graph structure from a noise amplifier into a training stabilizer. Extensive experiments on four multimodal graph datasets demonstrate the effectiveness of SR-GM, highlighting its state-of-the-art performance and cross-architecture generalization capabilities in multimodal graph dataset condensation.

---

### Decoupling and Damping: Structurally-Regularized Gradient Matching for Multimodal Graph Condensation

Authors: Lian Shen, Zhendan Chen, Yinhui jiang, Meijia Song, Ziming Su, Juan Liu, Xiangrong Liu
Published: 2025-11-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2511.20222v1

In critical web applications such as e-commerce and recommendation systems, multimodal graphs integrating rich visual and textual attributes are increasingly central, yet their large scale introduces substantial computational burdens for training Graph Neural Networks (GNNs). While Graph Condensation (GC) offers a promising solution by synthesizing smaller datasets, existing methods falter in the multimodal setting. We identify a dual challenge causing this failure: (1) conflicting gradients arising from semantic misalignments between modalities, and (2) the GNN's message-passing architecture pathologically amplifying this gradient noise across the graph structure. To address this, we propose Structurally-Regularized Gradient Matching (SR-GM), a novel condensation framework tailored for multimodal graphs. SR-GM introduces two synergistic components: first, a gradient decoupling mechanism that resolves inter-modality conflicts at their source via orthogonal projection; and second, a structural damping regularizer that acts directly on the gradient field. By leveraging the graph's Dirichlet energy, this regularizer transforms the topology from a noise amplifier into a stabilizing force during optimization. Extensive experiments demonstrate that SR-GM significantly improves accuracy and accelerates convergence compared to baseline methods. Ablation studies confirm that addressing both gradient conflict and structural amplification in tandem is essential for achieving superior performance. Moreover, the condensed multimodal graphs exhibit strong cross-architecture generalization and promise to accelerate applications like Neural Architecture Search. This research provides a scalable methodology for multimodal graph-based learning in resource-constrained environments.

---

### Patch-Level Glioblastoma Subregion Classification with a Contrastive Learning-Based Encoder

Authors: Juexin Zhang, Qifeng Zhong, Ying Weng, Ke Chen
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20221v1

The significant molecular and pathological heterogeneity of glioblastoma, an aggressive brain tumor, complicates diagnosis and patient stratification. While traditional histopathological assessment remains the standard, deep learning offers a promising path toward objective and automated analysis of whole slide images. For the BraTS-Path 2025 Challenge, we developed a method that fine-tunes a pre-trained Vision Transformer (ViT) encoder with a dedicated classification head on the official training dataset. Our model's performance on the online validation set, evaluated via the Synapse platform, yielded a Matthews Correlation Coefficient (MCC) of 0.7064 and an F1-score of 0.7676. On the final test set, the model achieved an MCC of 0.6509 and an F1-score of 0.5330, which secured our team second place in the BraTS-Pathology 2025 Challenge. Our results establish a solid baseline for ViT-based histopathological analysis, and future efforts will focus on bridging the performance gap observed on the unseen validation data.

---

### Communication-Efficient Learning for Satellite Constellations

Authors: Ruxandra-Stefania Tudose, Moritz H. W. Grüss, Grace Ra Kim, Karl H. Johansson, Nicola Bastianello
Published: 2025-11-25
Categories: cs.LG, eess.SY, math.OC
arXiv: https://arxiv.org/abs/2511.20220v1

Satellite constellations in low-Earth orbit are now widespread, enabling positioning, Earth imaging, and communications. In this paper we address the solution of learning problems using these satellite constellations. In particular, we focus on a federated approach, where satellites collect and locally process data, with the ground station aggregating local models. We focus on designing a novel, communication-efficient algorithm that still yields accurate trained models. To this end, we employ several mechanisms to reduce the number of communications with the ground station (local training) and their size (compression). We then propose an error feedback mechanism that enhances accuracy, which yields, as a byproduct, an algorithm-agnostic error feedback scheme that can be more broadly applied. We analyze the convergence of the resulting algorithm, and compare it with the state of the art through simulations in a realistic space scenario, showcasing superior performance.

---

### Text-guided Controllable Diffusion for Realistic Camouflage Images Generation

Authors: Yuhang Qian, Haiyan Chen, Wentong Li, Ningzhong Liu, Jie Qin
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20218v1

Camouflage Images Generation (CIG) is an emerging research area that focuses on synthesizing images in which objects are harmoniously blended and exhibit high visual consistency with their surroundings. Existing methods perform CIG by either fusing objects into specific backgrounds or outpainting the surroundings via foreground object-guided diffusion. However, they often fail to obtain natural results because they overlook the logical relationship between camouflaged objects and background environments. To address this issue, we propose CT-CIG, a Controllable Text-guided Camouflage Images Generation method that produces realistic and logically plausible camouflage images. Leveraging Large Visual Language Models (VLM), we design a Camouflage-Revealing Dialogue Mechanism (CRDM) to annotate existing camouflage datasets with high-quality text prompts. Subsequently, the constructed image-prompt pairs are utilized to finetune Stable Diffusion, incorporating a lightweight controller to guide the location and shape of camouflaged objects for enhanced camouflage scene fitness. Moreover, we design a Frequency Interaction Refinement Module (FIRM) to capture high-frequency texture features, facilitating the learning of complex camouflage patterns. Extensive experiments, including CLIPScore evaluation and camouflage effectiveness assessment, demonstrate the semantic alignment of our generated text prompts and CT-CIG's ability to produce photorealistic camouflage images.

---

### CostNav: A Navigation Benchmark for Real-World Economic-Cost Evaluation of Physical AI Agents

Authors: Haebin Seong, Sungmin Kim, Yongjun Cho, Myunchul Joe, Geunwoo Kim, Yubeen Park, Sunhoo Kim, Samwoo Seong, Yoonshik Kim, Suhwan Choi, Jaeyoon Jung, Jiyong Youn et al.
Published: 2025-11-25
Categories: cs.AI, cs.CE, cs.CV, cs.LG, cs.RO
arXiv: https://arxiv.org/abs/2511.20216v7

Current navigation benchmarks focus on task success but do not capture the economic constraints essential for commercializing autonomous delivery systems. We introduce CostNav, an Economic Navigation Benchmark that evaluates physical AI agents on a cost-revenue and break-even analysis, pairing Isaac Sim's collision and cargo dynamics with industry-standard data such as Securities and Exchange Commission (SEC) filings and Abbreviated Injury Scale (AIS) injury reports. To our knowledge, CostNav is the first physics-grounded economic benchmark to use regulatory and financial data to quantify the gap between navigation metrics and commercial deployment, revealing that high task-success rates alone do not ensure economic viability. Evaluating seven baselines (two rule-based and five imitation-learning methods), we find no method economically viable: all yield negative contribution margins. CANVAS, using only an RGB camera and GPS, attains the highest task success and the least-negative margin among methods with non-zero Service-Level Agreement (SLA) compliance (-\$28.40/run), outperforming LiDAR-equipped Nav2 w/ GPS (-\$37.34/run). A sim-trained policy evaluated on a real delivery robot yields SLA compliance close to its simulation result, indicating that policy performance in CostNav's simulation transfers to real-world deployment. We challenge the community to achieve economic viability on CostNav, which scores methods by cost-revenue outcomes. All resources are available at https://github.com/worv-ai/CostNav.

---

### CostNav: A Navigation Benchmark for Cost-Aware Evaluation of Embodied Agents

Authors: Haebin Seong, Sungmin Kim, Minchan Kim, Yongjun Cho, Myunchul Joe, Suhwan Choi, Jaeyoon Jung, Jiyong Youn, Yoonshik Kim, Samwoo Seong, Yubeen Park, Youngjae Yu et al.
Published: 2025-11-25
Categories: cs.AI, cs.CE, cs.CV, cs.LG, cs.RO
arXiv: https://arxiv.org/abs/2511.20216v1

Existing navigation benchmarks focus on task success metrics while overlooking economic viability -- critical for commercial deployment of autonomous delivery robots. We introduce \emph{CostNav}, a \textbf{Micro-Navigation Economic Testbed} that evaluates embodied agents through comprehensive cost-revenue analysis aligned with real-world business operations. CostNav models the complete economic lifecycle including hardware, training, energy, maintenance costs, and delivery revenue with service-level agreements, using industry-derived parameters. \textbf{To our knowledge, CostNav is the first work to quantitatively expose the gap between navigation research metrics and commercial viability}, revealing that optimizing for task success fundamentally differs from optimizing for economic deployment. Our cost model uses parameters derived from industry data sources (energy rates, delivery service pricing), and we project from a reduced-scale simulation to realistic deliveries. Under this projection, the baseline achieves 43.0\% SLA compliance but is \emph{not} commercially viable: yielding a loss of \$30.009 per run with no finite break-even point, because operating costs are dominated by collision-induced maintenance, which accounts for 99.7\% of per-run costs and highlights collision avoidance as a key optimization target. We demonstrate a learning-based on-device navigation baseline and establish a foundation for evaluating rule-based navigation, imitation learning, and cost-aware RL training. CostNav bridges the gap between navigation research and commercial deployment, enabling data-driven decisions about economic trade-offs across navigation paradigms.

---

### OmniAlpha: Aligning Transparency-Aware Generation via Multi-Task Unified Reinforcement Learning

Authors: Hao Yu, Jinglin Wang, Jiabo Zhan, Rui Chen, Zile Wang, Huaisong Zhang, Hongyu Li, Xinrui Chen, Yongxian Wei, Chun Yuan
Published: 2025-11-25
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2511.20211v2

Transparency-aware generation requires modeling not only RGB appearance but also alpha-based opacity and cross-layer composition, which are essential for tasks such as image matting, object removal, layer decomposition, and multi-layer content creation. However, existing RGBA-related methods remain largely fragmented, with separate pipelines designed for individual tasks. While a unified model is desirable, supervised fine-tuning alone is insufficient, as localized regression objectives cannot directly optimize the compositional fidelity, alpha-boundary precision, and structural consistency required for high-quality RGBA generation. To address this, we propose OmniAlpha, a unified multi-task reinforcement learning framework for transparency-aware generation and manipulation. OmniAlpha combines an end-to-end alpha-aware VAE and a sequence-to-sequence Diffusion Transformer, with a bi-directional layer axis in positional encoding to jointly model multiple RGBA inputs and outputs within a single forward pass. Built on a multi-task SFT cold start, it further performs GRPO-style post-training with layer-aware rewards defined on decoded RGBA outputs, enabling direct optimization of cross-layer coherence and fine transparency details. Experiments across five categories of transparency-aware tasks show that OmniAlpha consistently outperforms its unified SFT baseline and achieves strong performance against specialized expert models, including a 9.07% relative reduction in RGB L1 on layer decomposition and 74%/68% improvements over conventional matting tools on SAD/Grad for automatic matting.

---

### OmniAlpha: A Sequence-to-Sequence Framework for Unified Multi-Task RGBA Generation

Authors: Hao Yu, Jiabo Zhan, Zile Wang, Jinglin Wang, Huaisong Zhang, Hongyu Li, Xinrui Chen, Yongxian Wei, Chun Yuan
Published: 2025-11-25
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2511.20211v1

Generative models have excelled in RGB synthesis, but real-world applications require RGBA manipulation. This has led to a fragmented landscape: specialized, single-task models handle alpha but lack versatility, while unified multi-task frameworks are confined to the RGB domain. To bridge this critical gap, we propose OmniAlpha, the first unified, multi-task generative framework for sequence-to-sequence RGBA image generation and editing. Its architecture features MSRoPE-BiL, a novel RoPE method with a bi-directionally extendable layer axis for its Diffusion Transformer (DiT) backbone, enabling the concurrent processing of multiple input and target RGBA layers. To power this framework, we introduce AlphaLayers, a new dataset of 1,000 high-quality, multi-layer triplets, built via a novel automated synthesis and filter pipeline. Jointly training OmniAlpha on this dataset across a comprehensive suite of 21 diverse tasks, extensive experiments demonstrate that our unified approach consistently outperforms strong, specialized baselines. Most notably, OmniAlpha achieves a dramatic 84.8% relative reduction in SAD for mask-free matting on AIM-500 and wins over 90% of human preferences in layer-conditioned completion. Our work proves that a unified, multi-task model can learn a superior shared representation for RGBA, paving the way for more powerful, layer-aware generative systems.

---

### Adaptive SGD with Line-Search and Polyak Stepsizes: Nonconvex Convergence and Accelerated Rates

Authors: Haotian Wu
Published: 2025-11-25
Categories: math.OC, stat.ML
arXiv: https://arxiv.org/abs/2511.20207v4

We extend the convergence analysis of AdaSLS and AdaSPS in (Jiang and Stich, 2024) to the nonconvex setting, presenting a unified convergence analysis of stochastic gradient descent with adaptive Armijo line-search (AdaSLS) and Polyak stepsize (AdaSPS) for nonconvex optimization. Our contributions include: (1) an $\mathcal{O}(1/\sqrt{T})$ convergence rate for general nonconvex smooth functions, (2) an $\mathcal{O}(1/T)$ rate under quasar-convexity and interpolation, and (3) an $\mathcal{O}(1/T)$ rate under the strong growth condition for general nonconvex functions.

---

### Adaptive SGD with Line-Search and Polyak Stepsizes: Nonconvex Convergence and Accelerated Rates

Authors: Haotian Wu
Published: 2025-11-25
Categories: math.OC, stat.ML
arXiv: https://arxiv.org/abs/2511.20207v1

We extend the convergence analysis of AdaSLS and AdaSPS in (Jiang and Stich, 2024) to the nonconvex setting, presenting a unified convergence analysis of stochastic gradient descent with adaptive Armijo line-search (AdaSLS) and Polyak stepsize (AdaSPS) for nonconvex optimization. Our contributions include: (1) an $\mathcal{O}(1/\sqrt{T})$ convergence rate for general nonconvex smooth functions, (2) an $\mathcal{O}(1/T)$ rate under quasar-convexity and interpolation, and (3) an $\mathcal{O}(1/T)$ rate under the strong growth condition for general nonconvex functions.

---

### Robust 3D Brain MRI Inpainting with Random Masking Augmentation

Authors: Juexin Zhang, Ying Weng, Ke Chen
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20202v1

The ASNR-MICCAI BraTS-Inpainting Challenge was established to mitigate dataset biases that limit deep learning models in the quantitative analysis of brain tumor MRI. This paper details our submission to the 2025 challenge, a novel deep learning framework for synthesizing healthy tissue in 3D scans. The core of our method is a U-Net architecture trained to inpaint synthetically corrupted regions, enhanced with a random masking augmentation strategy to improve generalization. Quantitative evaluation confirmed the efficacy of our approach, yielding an SSIM of 0.873$\pm$0.004, a PSNR of 24.996$\pm$4.694, and an MSE of 0.005$\pm$0.087 on the validation set. On the final online test set, our method achieved an SSIM of 0.919$\pm$0.088, a PSNR of 26.932$\pm$5.057, and an RMSE of 0.052$\pm$0.026. This performance secured first place in the BraTS-Inpainting 2025 challenge and surpassed the winning solutions from the 2023 and 2024 competitions on the official leaderboard.

---

### GHR-VQA: Graph-guided Hierarchical Relational Reasoning for Video Question Answering

Authors: Dionysia Danai Brilli, Dimitrios Mallis, Vassilis Pitsikalis, Petros Maragos
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20201v1

We propose GHR-VQA, Graph-guided Hierarchical Relational Reasoning for Video Question Answering (Video QA), a novel human-centric framework that incorporates scene graphs to capture intricate human-object interactions within video sequences. Unlike traditional pixel-based methods, each frame is represented as a scene graph and human nodes across frames are linked to a global root, forming the video-level graph and enabling cross-frame reasoning centered on human actors. The video-level graphs are then processed by Graph Neural Networks (GNNs), transforming them into rich, context-aware embeddings for efficient processing. Finally, these embeddings are integrated with question features in a hierarchical network operating across different abstraction levels, enhancing both local and global understanding of video content. This explicit human-rooted structure enhances interpretability by decomposing actions into human-object interactions and enables a more profound understanding of spatiotemporal dynamics. We validate our approach on the Action Genome Question Answering (AGQA) dataset, achieving significant performance improvements, including a 7.3% improvement in object-relation reasoning over the state of the art.

---

### Interactive AI NPCs Powered by LLMs: Technical Report for the CPDC Challenge 2025

Authors: Yitian Huang, Yuxuan Lei, Jianxun Lian, Hao Liao
Published: 2025-11-25
Categories: cs.AI
arXiv: https://arxiv.org/abs/2511.20200v1

This report presents the solution and results of our team MSRA\_SC in the Commonsense Persona-Grounded Dialogue Challenge (CPDC 2025). We propose a simple yet effective framework that unifies improvements across both GPU Track and API Track. Our method centers on two key components. First, Context Engineering applies dynamic tool pruning and persona clipping for input compression, combined with post-processing techniques such as parameter normalization and function merging. Together with manually refined prompts, this design improves tool call stability, execution reliability, and role-playing guidance. Second, in the GPU Track, we further adopt GRPO training, replacing supervised fine-tuning with reinforcement learning directly optimized by reward signals. This mitigates small-sample overfitting and significantly enhances task-oriented dialogue performance. In the final evaluation, our team ranks 1st in Task 2 API, 2nd in Task 1 API, and 3rd in both Task 3 API and GPU track, demonstrating the effectiveness of our approach. Our code is publicly available at https://gitlab.aicrowd.com/nikoo_yu/cpdc-2025-winning-solution

---

### Towards Benign Memory Forgetting for Selective Multimodal Large Language Model Unlearning

Authors: Zhen Zeng, Leijiang Gu, Zhangling Duan, Feng Li, Cees G. M. Snoek, Meng Wang, Zenglin Shi
Published: 2025-11-25
Categories: cs.AI
arXiv: https://arxiv.org/abs/2511.20196v2

Multimodal large language models (MLLMs) can inadvertently memorize privacy-sensitive information during training. While existing unlearning methods can remove such content, they often severely degrade the model's foundational capabilities, such as general image understanding. This critical shortfall motivates our investigation into benign memory forgetting, the precise removal of targeted, privacy-sensitive knowledge while rigorously preserving unrelated capabilities. To pioneer and evaluate progress toward this objective, we introduce S-MLLMUn Bench, the first benchmark designed to jointly and quantitatively assess an unlearning method's efficacy in knowledge erasure and the preservation of image understanding. Furthermore, we propose the Sculpted Memory Forgetting Adapter (SMFA), a new framework that enables benign memory forgetting. SMFA confines forgetting to designated memory regions, maintaining overall model performance. By initially fine-tuning the model to replace sensitive outputs with refusals, SMFA generates a memory forgetting adapter, followed by a retaining anchor-guided masking mechanism that safeguards unrelated knowledge. Extensive experiments on S-MLLMUn Bench demonstrate that existing methods fail to achieve benign forgetting, whereas our proposed SMFA serves as an effective baseline, successfully achieving targeted knowledge erasure without compromising the model's foundational visual capabilities. Code and data are available at https://github.com/zeng-zhen/S-MLLMUn.

---

### Towards Benign Memory Forgetting for Selective Multimodal Large Language Model Unlearning

Authors: Zhen Zeng, Leijiang Gu, Zhangling Duan, Feng Li, Zenglin Shi, Cees G. M. Snoek, Meng Wang
Published: 2025-11-25
Categories: cs.AI
arXiv: https://arxiv.org/abs/2511.20196v1

Multimodal Large Language Models (MLLMs) achieve remarkable capabilities but can inadvertently memorize privacy-sensitive information. Although existing unlearning methods can remove such knowledge, they fail to achieve benign forgetting because they often degrade the model's general image understanding performance. To address this, we propose the Sculpted Memory Forgetting Adapter (SMFA), which confines forgetting to targeted memory regions while preserving overall capabilities. SMFA first fine-tunes the model to replace sensitive responses with refusals, yielding a memory forgetting adapter, and then applies a retaining anchor-guided masking mechanism to prevent interference with unrelated knowledge and understanding ability. To systematically evaluate selective MLLM unlearning, we introduce S-MLLMUn Bench, the first benchmark designed to jointly assess the removal of sensitive knowledge and retention of general visual understanding. Extensive experiments show that, unlike prior methods, SMFA achieves precise and controllable unlearning while maintaining the model's foundational image understanding.

---

### In-Context Compositional Learning via Sparse Coding Transformer

Authors: Wei Chen, Jingxi Yu, Zichen Miao, Qiang Qiu
Published: 2025-11-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2511.20194v1

Transformer architectures have achieved remarkable success across language, vision, and multimodal tasks, and there is growing demand for them to address in-context compositional learning tasks. In these tasks, models solve the target problems by inferring compositional rules from context examples, which are composed of basic components structured by underlying rules. However, some of these tasks remain challenging for Transformers, which are not inherently designed to handle compositional tasks and offer limited structural inductive bias. In this work, inspired by the principle of sparse coding, we propose a reformulation of the attention to enhance its capability for compositional tasks. In sparse coding, data are represented as sparse combinations of dictionary atoms with coefficients that capture their compositional rules. Specifically, we reinterpret the attention block as a mapping of inputs into outputs through projections onto two sets of learned dictionary atoms: an encoding dictionary and a decoding dictionary. The encoding dictionary decomposes the input into a set of coefficients, which represent the compositional structure of the input. To enhance structured representations, we impose sparsity on these coefficients. The sparse coefficients are then used to linearly combine the decoding dictionary atoms to generate the output. Furthermore, to assist compositional generalization tasks, we propose estimating the coefficients of the target problem as a linear combination of the coefficients obtained from the context examples. We demonstrate the effectiveness of our approach on the S-RAVEN and RAVEN datasets. For certain compositional generalization tasks, our method maintains performance even when standard Transformers fail, owing to its ability to learn and apply compositional rules.

---

### SFA: Scan, Focus, and Amplify toward Guidance-aware Answering for Video TextVQA

Authors: Haibin He, Qihuang Zhong, Juhua Liu, Bo Du, Peng Wang, Jing Zhang
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20190v1

Video text-based visual question answering (Video TextVQA) task aims to answer questions about videos by leveraging the visual text appearing within the videos. This task poses significant challenges, requiring models to accurately perceive and comprehend scene text that varies in scale, orientation, and clarity across frames, while effectively integrating temporal and semantic context to generate precise answers. Moreover, the model must identify question-relevant textual cues and filter out redundant or irrelevant information to ensure answering is guided by the most relevant and informative cues. To address these challenges, we propose SFA, a training-free framework and the first Video-LLM-based method tailored for Video TextVQA, motivated by the human process of answering questions. By adaptively scanning video frames, selectively focusing on key regions, and directly amplifying them, SFA effectively guides the Video-LLM's attention toward essential cues, enabling it to generate more accurate answers. SFA achieves new state-of-the-art results across several public Video TextVQA datasets and surpasses previous methods by a substantial margin, demonstrating its effectiveness and generalizability.

---

### Learning Subgroups with Maximum Treatment Effects without Causal Heuristics

Authors: Lincen Yang, Zhong Li, Matthijs van Leeuwen, Saber Salehkaleybar
Published: 2025-11-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2511.20189v1

Discovering subgroups with the maximum average treatment effect is crucial for targeted decision making in domains such as precision medicine, public policy, and education. While most prior work is formulated in the potential outcome framework, the corresponding structural causal model (SCM) for this task has been largely overlooked. In practice, two approaches dominate. The first estimates pointwise conditional treatment effects and then fits a tree on those estimates, effectively turning subgroup estimation into the harder problem of accurate pointwise estimation. The second constructs decision trees or rule sets with ad-hoc 'causal' heuristics, typically without rigorous justification for why a given heuristic may be used or whether such heuristics are necessary at all. We address these issues by studying the problem directly under the SCM framework. Under the assumption of a partition-based model, we show that optimal subgroup discovery reduces to recovering the data-generating models and hence a standard supervised learning problem (regression or classification). This allows us to adopt any partition-based methods to learn the subgroup from data. We instantiate the approach with CART, arguably one of the most widely used tree-based methods, to learn the subgroup with maximum treatment effect. Finally, on a large collection of synthetic and semi-synthetic datasets, we compare our method against a wide range of baselines and find that our approach, which avoids such causal heuristics, more accurately identifies subgroups with maximum treatment effect. Our source code is available at https://github.com/ylincen/causal-subgroup.

---

### Exo2EgoSyn: Unlocking Foundation Video Generation Models for Exocentric-to-Egocentric Video Synthesis

Authors: Mohammad Mahdi, Yuqian Fu, Nedko Savov, Jiancheng Pan, Danda Pani Paudel, Luc Van Gool
Published: 2025-11-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2511.20186v2

Foundation video generation models such as WAN 2.2 exhibit strong text- and image-conditioned synthesis abilities but remain constrained to the same-view generation setting. In this work, we introduce Exo2EgoSyn, an adaptation of WAN 2.2 that unlocks Exocentric-to-Egocentric(Exo2Ego) cross-view video synthesis. Our framework consists of three key modules. Ego-Exo View Alignment(EgoExo-Align) enforces latent-space alignment between exocentric and egocentric first-frame representations, reorienting the generative space from the given exo view toward the ego view. Multi-view Exocentric Video Conditioning (MultiExoCon) aggregates multi-view exocentric videos into a unified conditioning signal, extending WAN2.2 beyond its vanilla single-image or text conditioning. Furthermore, Pose-Aware Latent Injection (PoseInj) injects relative exo-to-ego camera pose information into the latent state, guiding geometry-aware synthesis across viewpoints. Together, these modules enable high-fidelity ego view video generation from third-person observations without retraining from scratch. Experiments on ExoEgo4D validate that Exo2EgoSyn significantly improves Ego2Exo synthesis, paving the way for scalable cross-view video generation with foundation models. Source code and models will be released publicly.

---