# Research Papers — 2025-12 · bundle page 86 of 175 > Full-text markdown of papers #4251–4300 of 8703 published in 2025-12, newest first. AI-readable mirrors of the full 8703-paper 2025-12 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2025-12/page/86 [← Previous bundle](https://aipod.dev/papers/archive/2025-12/page/85.md) [Next bundle →](https://aipod.dev/papers/archive/2025-12/page/87.md) ### Understanding Structured Financial Data with LLMs: A Case Study on Fraud Detection Authors: Xuwei Tan, Yao Ma, Xueru Zhang Published: 2025-12-15 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2512.13040v2 Detecting fraud in financial transactions typically relies on tabular models that demand heavy feature engineering to handle high-dimensional data and offer limited interpretability, making it difficult for humans to understand predictions. Large Language Models (LLMs), in contrast, can produce human-readable explanations and facilitate feature analysis, potentially reducing the manual workload of fraud analysts and informing system refinements. However, they perform poorly when applied directly to tabular fraud detection due to the difficulty of reasoning over many features, the extreme class imbalance, and the absence of contextual information. To bridge this gap, we introduce FinFRE-RAG, a two-stage approach that applies importance-guided feature reduction to serialize a compact subset of numeric/categorical attributes into natural language and performs retrieval-augmented in-context learning over label-aware, instance-level exemplars. Across four public fraud datasets and three families of open-weight LLMs, FinFRE-RAG substantially improves F1/MCC over direct prompting and is competitive with strong tabular baselines in several settings. Although these LLMs still lag behind specialized classifiers, they narrow the performance gap and provide interpretable rationales, highlighting their value as assistive tools in fraud analysis. --- ### Bi-Erasing: A Bidirectional Framework for Concept Removal in Diffusion Models Authors: Hao Chen, Yiwei Wang, Songze Li Published: 2025-12-15 Categories: cs.CV, cs.CR arXiv: https://arxiv.org/abs/2512.13039v2 Concept erasure, which fine-tunes diffusion models to remove undesired or harmful visual concepts, has become a mainstream approach to mitigating unsafe or illegal image generation in text-to-image models.However, existing removal methods typically adopt a unidirectional erasure strategy by either suppressing the target concept or reinforcing safe alternatives, making it difficult to achieve a balanced trade-off between concept removal and generation quality. To address this limitation, we propose a novel Bidirectional Image-Guided Concept Erasure (Bi-Erasing) framework that performs concept suppression and safety enhancement simultaneously. Specifically, based on the joint representation of text prompts and corresponding images, Bi-Erasing introduces two decoupled image branches: a negative branch responsible for suppressing harmful semantics and a positive branch providing visual guidance for safe alternatives. By jointly optimizing these complementary directions, our approach achieves a balance between erasure efficacy and generation usability. In addition, we apply mask-based filtering to the image branches to prevent interference from irrelevant content during the erasure process. Across extensive experiment evaluations, the proposed Bi-Erasing outperforms baseline methods in balancing concept removal effectiveness and visual fidelity. --- ### Progressive Refinement of E-commerce Search Ranking Based on Short-Term Activities of the Buyer Authors: Taoran Sheng, Sathappan Muthiah, Atiq Islam, Jinming Feng Published: 2025-12-15 Categories: cs.IR, cs.LG arXiv: https://arxiv.org/abs/2512.13037v1 In e-commerce shopping, aligning search results with a buyer's immediate needs and preferences presents a significant challenge, particularly in adapting search results throughout the buyer's shopping journey as they move from the initial stages of browsing to making a purchase decision or shift from one intent to another. This study presents a systematic approach to adapting e-commerce search results based on the current context. We start with basic methods and incrementally incorporate more contextual information and state-of-the-art techniques to improve the search outcomes. By applying this evolving contextual framework to items displayed on the search engine results page (SERP), we progressively align search outcomes more closely with the buyer's interests and current search intentions. Our findings demonstrate that this incremental enhancement, from simple heuristic autoregressive features to advanced sequence models, significantly improves ranker performance. The integration of contextual techniques enhances the performance of our production ranker, leading to improved search results in both offline and online A/B testing in terms of Mean Reciprocal Rank (MRR). Overall, the paper details iterative methodologies and their substantial contributions to search result contextualization on e-commerce platforms. --- ### Alada: Alternating Adaptation of Momentum Method for Memory-Efficient Matrix Optimization Authors: Xiaoyu He, Yu Cai, Jin Jia, Canxi Huang, Wenqing Chen, Zibin Zheng Published: 2025-12-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.13034v1 This work proposes Alada, an adaptive momentum method for stochastic optimization over large-scale matrices. Alada employs a rank-one factorization approach to estimate the second moment of gradients, where factors are updated alternatively to minimize the estimation error. Alada achieves sublinear memory overheads and can be readily extended to optimizing tensor-shaped variables.We also equip Alada with a first moment estimation rule, which enhances the algorithm's robustness without incurring additional memory overheads. The theoretical performance of Alada aligns with that of traditional methods such as Adam. Numerical studies conducted on several natural language processing tasks demonstrate the reduction in memory overheads and the robustness in training large models relative to Adam and its variants. --- ### Scaling Bidirectional Spans and Span Violations in Attention Mechanism Authors: Jongwook Kim, Sangheon Yun, Sukjin Yoon Published: 2025-12-15 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2512.13033v1 The canonical $O(N^2)$ Transformer remains the empirical performance frontier in sequence modeling, and its training can be further optimized by addressing geometric inefficiency. We propose an optimization framework that leverages an asymmetric projection to decompose the backward-pass gradients into parallel spans and orthogonal violations, while keeping the canonical forward-pass $QKV$ structure intact. Through consistent experimental validation across various decomposition and projection setups, we provide strong theoretical evidence: the standard attention gradient is suboptimal. We demonstrated that selectively scaling these components, focusing primarily on $0^{th}$ order bidirectional parallel spans, yields the most effective learning signal. On the limited WikiText-2 dataset, and using a crude configuration, this method achieved a $0.56\%$ reduction in validation loss, confirming the framework's fundamental validity and suggesting significant potential gains on larger datasets and deeper training regimes --- ### Comprehensive Evaluation of Rule-Based, Machine Learning, and Deep Learning in Human Estimation Using Radio Wave Sensing: Accuracy, Spatial Generalization, and Output Granularity Trade-offs Authors: Tomoya Tanaka, Tomonori Ikeda, Ryo Yonemoto Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.13031v1 This study presents the first comprehensive comparison of rule-based methods, traditional machine learning models, and deep learning models in radio wave sensing with frequency modulated continuous wave multiple input multiple output radar. We systematically evaluated five approaches in two indoor environments with distinct layouts: a rule-based connected component method; three traditional machine learning models, namely k-nearest neighbors, random forest, and support vector machine; and a deep learning model combining a convolutional neural network and long short term memory. In the training environment, the convolutional neural network long short term memory model achieved the highest accuracy, while traditional machine learning models provided moderate performance. In a new layout, however, all learning based methods showed significant degradation, whereas the rule-based method remained stable. Notably, for binary detection of presence versus absence of people, all models consistently achieved high accuracy across layouts. These results demonstrate that high capacity models can produce fine grained outputs with high accuracy in the same environment, but they are vulnerable to domain shift. In contrast, rule-based methods cannot provide fine grained outputs but exhibit robustness against domain shift. Moreover, regardless of the model type, a clear trade off was revealed between spatial generalization performance and output granularity. --- ### Motus: A Unified Latent Action World Model Authors: Hongzhe Bi, Hengkai Tan, Shenghao Xie, Zeyuan Wang, Shuhe Huang, Haitian Liu, Ruowen Zhao, Yao Feng, Chendong Xiang, Yinze Rong, Hongyan Zhao, Hanyu Liu et al. Published: 2025-12-15 Categories: cs.CV, cs.LG, cs.RO arXiv: https://arxiv.org/abs/2512.13030v2 While a general embodied agent must function as a unified system, current methods are built on isolated models for understanding, world modeling, and control. This fragmentation prevents unifying multimodal generative capabilities and hinders learning from large-scale, heterogeneous data. In this paper, we propose Motus, a unified latent action world model that leverages existing general pretrained models and rich, sharable motion information. Motus introduces a Mixture-of-Transformer (MoT) architecture to integrate three experts (i.e., understanding, video generation, and action) and adopts a UniDiffuser-style scheduler to enable flexible switching between different modeling modes (i.e., world models, vision-language-action models, inverse dynamics models, video generation models, and video-action joint prediction models). Motus further leverages the optical flow to learn latent actions and adopts a recipe with three-phase training pipeline and six-layer data pyramid, thereby extracting pixel-level "delta action" and enabling large-scale action pretraining. Experiments show that Motus achieves superior performance against state-of-the-art methods in both simulation (a +15% improvement over X-VLA and a +45% improvement over Pi0.5) and real-world scenarios(improved by +11~48%), demonstrating unified modeling of all functionalities and priors significantly benefits downstream robotic tasks. --- ### SneakPeek: Future-Guided Instructional Streaming Video Generation Authors: Cheeun Hong, German Barquero, Fadime Sener, Markos Georgopoulos, Edgar Schönfeld, Stefan Popov, Yuming Du, Oscar Mañas, Albert Pumarola Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.13019v1 Instructional video generation is an emerging task that aims to synthesize coherent demonstrations of procedural activities from textual descriptions. Such capability has broad implications for content creation, education, and human-AI interaction, yet existing video diffusion models struggle to maintain temporal consistency and controllability across long sequences of multiple action steps. We introduce a pipeline for future-driven streaming instructional video generation, dubbed SneakPeek, a diffusion-based autoregressive framework designed to generate precise, stepwise instructional videos conditioned on an initial image and structured textual prompts. Our approach introduces three key innovations to enhance consistency and controllability: (1) predictive causal adaptation, where a causal model learns to perform next-frame prediction and anticipate future keyframes; (2) future-guided self-forcing with a dual-region KV caching scheme to address the exposure bias issue at inference time; (3) multi-prompt conditioning, which provides fine-grained and procedural control over multi-step instructions. Together, these components mitigate temporal drift, preserve motion consistency, and enable interactive video generation where future prompt updates dynamically influence ongoing streaming video generation. Experimental results demonstrate that our method produces temporally coherent and semantically faithful instructional videos that accurately follow complex, multi-step task descriptions. --- ### Comprehensive Deployment-Oriented Assessment for Cross-Environment Generalization in Deep Learning-Based mmWave Radar Sensing Authors: Tomoya Tanaka, Tomonori Ikeda, Ryo Yonemoto Published: 2025-12-15 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2512.13018v1 This study presents the first comprehensive evaluation of spatial generalization techniques, which are essential for the practical deployment of deep learning-based radio-frequency (RF) sensing. Focusing on people counting in indoor environments using frequency-modulated continuous-wave (FMCW) multiple-input multiple-output (MIMO) radar, we systematically investigate a broad set of approaches, including amplitude-based statistical preprocessing (sigmoid weighting and threshold zeroing), frequency-domain filtering, autoencoder-based background suppression, data augmentation strategies, and transfer learning. Experimental results collected across two environments with different layouts demonstrate that sigmoid-based amplitude weighting consistently achieves superior cross-environment performance, yielding 50.1% and 55.2% reductions in root-mean-square error (RMSE) and mean absolute error (MAE), respectively, compared with baseline methods. Data augmentation provides additional though modest benefits, with improvements up to 8.8% in MAE. By contrast, transfer learning proves indispensable for large spatial shifts, achieving 82.1% and 91.3% reductions in RMSE and MAE, respectively, with 540 target-domain samples. Taken together, these findings establish a highly practical direction for developing radar sensing systems capable of maintaining robust accuracy under spatial variations by integrating deep learning models with amplitude-based preprocessing and efficient transfer learning. --- ### What Happens Next? Next Scene Prediction with a Unified Video Model Authors: Xinjie Li, Zhimin Chen, Rui Zhao, Florian Schiffers, Zhenyu Liao, Vimal Bhat Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.13015v1 Recent unified models for joint understanding and generation have significantly advanced visual generation capabilities. However, their focus on conventional tasks like text-to-video generation has left the temporal reasoning potential of unified models largely underexplored. To address this gap, we introduce Next Scene Prediction (NSP), a new task that pushes unified video models toward temporal and causal reasoning. Unlike text-to-video generation, NSP requires predicting plausible futures from preceding context, demanding deeper understanding and reasoning. To tackle this task, we propose a unified framework combining Qwen-VL for comprehension and LTX for synthesis, bridged by a latent query embedding and a connector module. This model is trained in three stages on our newly curated, large-scale NSP dataset: text-to-video pre-training, supervised fine-tuning, and reinforcement learning (via GRPO) with our proposed causal consistency reward. Experiments demonstrate our model achieves state-of-the-art performance on our benchmark, advancing the capability of generalist multimodal systems to anticipate what happens next. --- ### JoDiffusion: Jointly Diffusing Image with Pixel-Level Annotations for Semantic Segmentation Promotion Authors: Haoyu Wang, Lei Zhang, Wenrui Liu, Dengyang Jiang, Wei Wei, Chen Ding Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.13014v1 Given the inherently costly and time-intensive nature of pixel-level annotation, the generation of synthetic datasets comprising sufficiently diverse synthetic images paired with ground-truth pixel-level annotations has garnered increasing attention recently for training high-performance semantic segmentation models. However, existing methods necessitate to either predict pseudo annotations after image generation or generate images conditioned on manual annotation masks, which incurs image-annotation semantic inconsistency or scalability problem. To migrate both problems with one stone, we present a novel dataset generative diffusion framework for semantic segmentation, termed JoDiffusion. Firstly, given a standard latent diffusion model, JoDiffusion incorporates an independent annotation variational auto-encoder (VAE) network to map annotation masks into the latent space shared by images. Then, the diffusion model is tailored to capture the joint distribution of each image and its annotation mask conditioned on a text prompt. By doing these, JoDiffusion enables simultaneously generating paired images and semantically consistent annotation masks solely conditioned on text prompts, thereby demonstrating superior scalability. Additionally, a mask optimization strategy is developed to mitigate the annotation noise produced during generation. Experiments on Pascal VOC, COCO, and ADE20K datasets show that the annotated dataset generated by JoDiffusion yields substantial performance improvements in semantic segmentation compared to existing methods. --- ### Deep Learning-Driven Inversion Framework for Shear Modulus Estimation in Magnetic Resonance Elastography (DIME) Authors: Hassan Iftikhar, Rizwan Ahmad, Arunark Kolipaka Published: 2025-12-15 Categories: cs.LG, q-bio.TO arXiv: https://arxiv.org/abs/2512.13010v1 The Multimodal Direct Inversion (MMDI) algorithm is widely used in Magnetic Resonance Elastography (MRE) to estimate tissue shear stiffness. However, MMDI relies on the Helmholtz equation, which assumes wave propagation in a uniform, homogeneous, and infinite medium. Furthermore, the use of the Laplacian operator makes MMDI highly sensitive to noise, which compromises the accuracy and reliability of stiffness estimates. In this study, we propose the Deep-Learning driven Inversion Framework for Shear Modulus Estimation in MRE (DIME), aimed at enhancing the robustness of inversion. DIME is trained on the displacement fields-stiffness maps pair generated through Finite Element Modelling (FEM) simulations. To capture local wave behavior and improve robustness to global image variations, DIME is trained on small image patches. We first validated DIME using homogeneous and heterogeneous datasets simulated with FEM, where DIME produced stiffness maps with low inter-pixel variability, accurate boundary delineation, and higher correlation with ground truth (GT) compared to MMDI. Next, DIME was evaluated in a realistic anatomy-informed simulated liver dataset with known GT and compared directly to MMDI. DIME reproduced ground-truth stiffness patterns with high fidelity (r = 0.99, R^2 = 0.98), while MMDI showed greater underestimation. After validating DIME on synthetic data, we tested the model in in vivo liver MRE data from eight healthy and seven fibrotic subjects. DIME preserved physiologically consistent stiffness patterns and closely matched MMDI, which showed directional bias. Overall, DIME showed higher correlation with ground truth and visually similar stiffness patterns, whereas MMDI displayed a larger bias that can potentially be attributed to directional filtering. These preliminary results highlight the feasibility of DIME for clinical applications in MRE. --- ### TWLR: Text-Guided Weakly-Supervised Lesion Localization and Severity Regression for Explainable Diabetic Retinopathy Grading Authors: Xi Luo, Shixin Xu, Ying Xie, JianZhong Hu, Yuwei He, Yuhui Deng, Huaxiong Huang Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.13008v1 Accurate medical image analysis can greatly assist clinical diagnosis, but its effectiveness relies on high-quality expert annotations Obtaining pixel-level labels for medical images, particularly fundus images, remains costly and time-consuming. Meanwhile, despite the success of deep learning in medical imaging, the lack of interpretability limits its clinical adoption. To address these challenges, we propose TWLR, a two-stage framework for interpretable diabetic retinopathy (DR) assessment. In the first stage, a vision-language model integrates domain-specific ophthalmological knowledge into text embeddings to jointly perform DR grading and lesion classification, effectively linking semantic medical concepts with visual features. The second stage introduces an iterative severity regression framework based on weakly-supervised semantic segmentation. Lesion saliency maps generated through iterative refinement direct a progressive inpainting mechanism that systematically eliminates pathological features, effectively downgrading disease severity toward healthier fundus appearances. Critically, this severity regression approach achieves dual benefits: accurate lesion localization without pixel-level supervision and providing an interpretable visualization of disease-to-healthy transformations. Experimental results on the FGADR, DDR, and a private dataset demonstrate that TWLR achieves competitive performance in both DR classification and lesion segmentation, offering a more explainable and annotation-efficient solution for automated retinal image analysis. --- ### Light Field Based 6DoF Tracking of Previously Unobserved Objects Authors: Nikolai Goncharov, James L. Gray, Donald G. Dansereau Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.13007v1 Object tracking is an important step in robotics and reautonomous driving pipelines, which has to generalize to previously unseen and complex objects. Existing high-performing methods often rely on pre-captured object views to build explicit reference models, which restricts them to a fixed set of known objects. However, such reference models can struggle with visually complex appearance, reducing the quality of tracking. In this work, we introduce an object tracking method based on light field images that does not depend on a pre-trained model, while being robust to complex visual behavior, such as reflections. We extract semantic and geometric features from light field inputs using vision foundation models and convert them into view-dependent Gaussian splats. These splats serve as a unified object representation, supporting differentiable rendering and pose optimization. We further introduce a light field object tracking dataset containing challenging reflective objects with precise ground truth poses. Experiments demonstrate that our method is competitive with state-of-the-art model-based trackers in these difficult cases, paving the way toward universal object tracking in robotic systems. Code/data available at https://github.com/nagonch/LiFT-6DoF. --- ### Few-Step Distillation for Text-to-Image Generation: A Practical Guide Authors: Yifan Pu, Yizeng Han, Zhiwei Tang, Jiasheng Tang, Fan Wang, Bohan Zhuang, Gao Huang Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.13006v1 Diffusion distillation has dramatically accelerated class-conditional image synthesis, but its applicability to open-ended text-to-image (T2I) generation is still unclear. We present the first systematic study that adapts and compares state-of-the-art distillation techniques on a strong T2I teacher model, FLUX.1-lite. By casting existing methods into a unified framework, we identify the key obstacles that arise when moving from discrete class labels to free-form language prompts. Beyond a thorough methodological analysis, we offer practical guidelines on input scaling, network architecture, and hyperparameters, accompanied by an open-source implementation and pretrained student models. Our findings establish a solid foundation for deploying fast, high-fidelity, and resource-efficient diffusion generators in real-world T2I applications. Code is available on github.com/alibaba-damo-academy/T2I-Distill. --- ### General OOD Detection via Model-aware and Subspace-aware Variable Priority Authors: Min Lu, Hemant Ishwaran Published: 2025-12-15 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2512.13003v2 Out-of-distribution (OOD) detection is essential for determining when a supervised model encounters inputs that differ meaningfully from its training distribution. While widely studied in classification, OOD detection for regression and survival analysis remains limited due to the absence of discrete labels and the challenge of quantifying predictive uncertainty. We introduce a framework for OOD detection that is simultaneously model aware and subspace aware, and that embeds variable prioritization directly into the detection step. The method uses the fitted predictor to construct localized neighborhoods around each test case that emphasize the features driving the model's learned relationship and downweight directions that are less relevant to prediction. It produces OOD scores without relying on global distance metrics or estimating the full feature density. The framework is applicable across outcome types, and in our implementation we use random forests, where the rule structure yields transparent neighborhoods and effective scoring. Experiments on synthetic and real data benchmarks designed to isolate functional shifts show consistent improvements over existing methods. We further demonstrate the approach in an esophageal cancer survival study, where distribution shifts related to lymphadenectomy identify patterns relevant to surgical guidelines. --- ### Calibrating Uncertainty for Zero-Shot Adversarial CLIP Authors: Wenjing Lu, Zerui Tao, Yuning Qiu, Dongping Zhang, Yang Yang, Qibin Zhao Published: 2025-12-15 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2512.12997v3 CLIP delivers strong zero-shot classification but remains highly vulnerable to adversarial attacks. Prior adversarial fine-tuning work primarily matches predicted logits between clean and adversarial examples, which overlooks uncertainty calibration and may degrade the zero-shot generalization. A common expectation in reliable uncertainty estimation is that predictive uncertainty should increase as inputs become more difficult or shift away from the training distribution. However, we frequently observe the opposite in the adversarial setting: perturbations not only degrade accuracy but also suppress uncertainty, leading to severe miscalibration and over-confidence. This reveals a critical reliability gap beyond robustness. To bridge this gap, we propose an adversarial fine-tuning objective for CLIP considering both accuracy and uncertainty. By reparameterizing CLIP outputs as the concentration parameters of a Dirichlet distribution, we propose a unified representation that captures relative semantic structure and confidence magnitude. This enables holistic distribution alignment under perturbations, moving beyond single-logit anchoring and restoring calibrated uncertainty. Experiments across multiple zero-shot benchmarks demonstrate that our method significantly improves uncertainty calibration and achieves competitive adversarial robustness while preserving clean accuracy. --- ### A Bayesian approach to learning mixtures of nonparametric components Authors: Yilei Zhang, Yun Wei, Aritra Guha, XuanLong Nguyen Published: 2025-12-15 Categories: stat.ME, math.ST, stat.CO, stat.ML arXiv: https://arxiv.org/abs/2512.12988v2 Mixture models are widely used in modeling heterogeneous data populations. A standard approach of mixture modeling assumes that the mixture component takes a parametric kernel form. In many applications, making parametric assumptions on the latent subpopulation distributions may be unrealistic, which motivates the need for nonparametric modeling of the mixture components themselves. In this paper, we study finite mixtures with nonparametric mixture components, using a Bayesian nonparametric modeling approach. In particular, it is assumed that the data population is generated according to a finite mixture of latent component distributions, where each component is endowed with a Bayesian nonparametric prior such as the Dirichlet process mixture. We present conditions under which the individual mixture component's distribution can be identified, and establish posterior contraction behavior for the data population's density, as well as densities of the latent mixture components. We develop an efficient MCMC algorithm for posterior inference and demonstrate via simulation studies and real-world data illustrations that it is possible to efficiently learn complex forms of probability distribution for the latent subpopulations. In theory, the posterior contraction rate of the component densities is nearly polynomial, which is a significant improvement over the logarithmic convergence rates of estimating mixing measures via deconvolution. --- ### Tackling Snow-Induced Challenges: Safe Autonomous Lane-Keeping with Robust Reinforcement Learning Authors: Amin Jalal Aghdasian, Farzaneh Abdollahi, Ali Kamali Iglie Published: 2025-12-15 Categories: cs.RO, cs.AI, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2512.12987v1 This paper proposes two new algorithms for the lane keeping system (LKS) in autonomous vehicles (AVs) operating under snowy road conditions. These algorithms use deep reinforcement learning (DRL) to handle uncertainties and slippage. They include Action-Robust Recurrent Deep Deterministic Policy Gradient (AR-RDPG) and end-to-end Action-Robust convolutional neural network Attention Deterministic Policy Gradient (AR-CADPG), two action-robust approaches for decision-making. In the AR-RDPG method, within the perception layer, camera images are first denoised using multi-scale neural networks. Then, the centerline coefficients are extracted by a pre-trained deep convolutional neural network (DCNN). These coefficients, concatenated with the driving characteristics, are used as input to the control layer. The AR-CADPG method presents an end-to-end approach in which a convolutional neural network (CNN) and an attention mechanism are integrated within a DRL framework. Both methods are first trained in the CARLA simulator and validated under various snowy scenarios. Real-world experiments on a Jetson Nano-based autonomous vehicle confirm the feasibility and stability of the learned policies. Among the two models, the AR-CADPG approach demonstrates superior path-tracking accuracy and robustness, highlighting the effectiveness of combining temporal memory, adversarial resilience, and attention mechanisms in AVs. --- ### VoroLight: Learning Voronoi Surface Meshes via Sphere Intersection Authors: Jiayin Lu, Ying Jiang, Yumeng He, Yin Yang, Chenfanfu Jiang Published: 2025-12-15 Categories: cs.CG, cs.CV, cs.GR, cs.LG, math.OC arXiv: https://arxiv.org/abs/2512.12984v2 Voronoi diagrams naturally produce convex, watertight, and topologically consistent cells, making them an appealing representation for 3D shape reconstruction. However, standard differentiable Voronoi approaches typically optimize generator positions in stable configurations, which can lead to locally uneven surface geometry. We present VoroLight, a differentiable framework that promotes controlled Voronoi degeneracy for smooth surface reconstruction. Instead of optimizing generator positions alone, VoroLight associates each Voronoi surface vertex with a trainable sphere and introduces a sphere--intersection loss that encourages higher-order equidistance among face-incident generators. This formulation improves surface regularity while preserving intrinsic Voronoi properties such as watertightness and convexity. Because losses are defined directly on surface vertices, VoroLight supports multimodal shape supervision from implicit fields, point clouds, meshes, and multi--view images. By introducing additional interior generators optimized under a centroidal Voronoi tessellation objective, the framework naturally extends to volumetric Voronoi meshes with consistent surface--interior topology. Across diverse input modalities, VoroLight achieves competitive reconstruction fidelity while producing smoother and more geometrically regular Voronoi surfaces. Project page: https://jiayinlu19960224.github.io/vorolight/ --- ### Scaling Up AI-Generated Image Detection with Generator-Aware Prototypes Authors: Ziheng Qin, Yuheng Ji, Renshuai Tao, Yuxuan Tian, Yuyang Liu, Yipu Wang, Xiaolong Zheng Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.12982v2 The pursuit of a universal AI-generated image (AIGI) detector often relies on aggregating data from numerous generators to improve generalization. However, this paper identifies a paradoxical phenomenon we term the Benefit then Conflict dilemma, where detector performance stagnates and eventually degrades as source diversity expands. Our systematic analysis, diagnoses this failure by identifying two core issues: severe data-level heterogeneity, which causes the feature distributions of real and synthetic images to increasingly overlap, and a critical model-level bottleneck from fixed, pretrained encoders that cannot adapt to the rising complexity. To address these challenges, we propose Generator-Aware Prototype Learning (GAPL), a framework that constrain representation with a structured learning paradigm. GAPL learns a compact set of canonical forgery prototypes to create a unified, low-variance feature space, effectively countering data heterogeneity.To resolve the model bottleneck, it employs a two-stage training scheme with Low-Rank Adaptation, enhancing its discriminative power while preserving valuable pretrained knowledge. This approach establishes a more robust and generalizable decision boundary. Through extensive experiments, we demonstrate that GAPL achieves state-of-the-art performance, showing superior detection accuracy across a wide variety of GAN and diffusion-based generators. Code is available at https://github.com/UltraCapture/GAPL --- ### CoDeQ: End-to-End Joint Model Compression with Dead-Zone Quantizer for High-Sparsity and Low-Precision Networks Authors: Jonathan Wenshøj, Tong Chen, Bob Pepin, Raghavendra Selvan Published: 2025-12-15 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2512.12981v1 While joint pruning--quantization is theoretically superior to sequential application, current joint methods rely on auxiliary procedures outside the training loop for finding compression parameters. This reliance adds engineering complexity and hyperparameter tuning, while also lacking a direct data-driven gradient signal, which might result in sub-optimal compression. In this paper, we introduce CoDeQ, a simple, fully differentiable method for joint pruning--quantization. Our approach builds on a key observation: the dead-zone of a scalar quantizer is equivalent to magnitude pruning, and can be used to induce sparsity directly within the quantization operator. Concretely, we parameterize the dead-zone width and learn it via backpropagation, alongside the quantization parameters. This design provides explicit control of sparsity, regularized by a single global hyperparameter, while decoupling sparsity selection from bit-width selection. The result is a method for Compression with Dead-zone Quantizer (CoDeQ) that supports both fixed-precision and mixed-precision quantization (controlled by an optional second hyperparameter). It simultaneously determines the sparsity pattern and quantization parameters in a single end-to-end optimization. Consequently, CoDeQ does not require any auxiliary procedures, making the method architecture-agnostic and straightforward to implement. On ImageNet with ResNet-18, CoDeQ reduces bit operations to ~5% while maintaining close to full precision accuracy in both fixed and mixed-precision regimes. --- ### VLCache: Computing 2% Vision Tokens and Reusing 98% for Vision-Language Inference Authors: Shengling Qin, Hao Yu, Chenxin Wu, Zheng Li, Yizhong Cao, Zhengyang Zhuge, Yuxin Zhou, Wentao Yao, Yi Zhang, Zhengheng Wang, Shuai Bai, Jianwei Zhang et al. Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.12977v2 This paper presents VLCache, a cache reuse framework that exploits both Key-Value (KV) cache and encoder cache from prior multimodal inputs to eliminate costly recomputation when the same multimodal inputs recur. Unlike previous heuristic approaches, we formally identify the cumulative reuse error effect and demonstrate how to minimize the non-prefix cache reuse error effectively. We further analyze the varying importance of model layers and propose a dynamic, layer-aware recomputation strategy to balance accuracy and efficiency. Experimental results show that VLCache achieves an accuracy on par with full recomputation, while requiring only 2-5% of the tokens to compute, yielding 1.2x-16x TTFT speedups. We develop an experimental implementation of the proposed VLCache pipeline based on SGLang, enabling significantly faster inference in practical deployments. --- ### Authors Should Label Their Own Documents Authors: Marcus Ma, Cole Johnson, Nolan Bridges, Jackson Trager, Georgios Chochlakis, Shrikanth Narayanan Published: 2025-12-15 Categories: cs.CL arXiv: https://arxiv.org/abs/2512.12976v3 Third-party annotation is the status quo for labeling text, but egocentric information such as sentiment and belief can at best only be approximated by a third-person proxy. We introduce author labeling, an annotation technique where the writer of the document itself annotates the data at the moment of creation. We collaborate with a commercial chatbot with over 20,000 users to deploy an author labeling annotation system. This system identifies task-relevant queries, generates on-the-fly labeling questions, and records authors' answers in real time. We train and deploy an online-learning model architecture for product recommendation with author-labeled data to improve performance. We train our model to minimize the prediction error on questions generated for a set of predetermined subjective beliefs using author-labeled responses. Our model achieves a 537% improvement in click-through rate compared to an industry advertising baseline running concurrently. We then compare the quality and practicality of author labeling to three traditional annotation approaches for sentiment analysis and find author labeling to be higher quality, faster to acquire, and cheaper. These findings reinforce existing literature that annotations, especially for egocentric and subjective beliefs, are significantly higher quality when labeled by the author rather than a third party. To facilitate broader scientific adoption, we release an author labeling service for the research community at https://academic.echogroup.ai. --- ### Application of Deep Learning in Biological Data Compression Authors: Chunyu Zou Published: 2025-12-15 Categories: cs.LG, cs.IT arXiv: https://arxiv.org/abs/2512.12975v1 Cryogenic electron microscopy (Cryo-EM) has become an essential tool for capturing high-resolution biological structures. Despite its advantage in visualizations, the large storage size of Cryo-EM data file poses significant challenges for researchers and educators. This paper investigates the application of deep learning, specifically implicit neural representation (INR), to compress Cryo-EM biological data. The proposed approach first extracts the binary map of each file according to the density threshold. The density map is highly repetitive, ehich can be effectively compressed by GZIP. The neural network then trains to encode spatial density information, allowing the storage of network parameters and learnable latent vectors. To improve reconstruction accuracy, I further incorporate the positional encoding to enhance spatial representation and a weighted Mean Squared Error (MSE) loss function to balance density distribution variations. Using this approach, my aim is to provide a practical and efficient biological data compression solution that can be used for educational and research purpose, while maintaining a reasonable compression ratio and reconstruction quality from file to file. --- ### Cycles Communities from the Perspective of Dendrograms and Gradient Sampling Authors: Sixtus Dakurah Published: 2025-12-15 Categories: q-bio.QM, cs.CG, stat.ML arXiv: https://arxiv.org/abs/2512.12974v1 Identifying and comparing topological features, particularly cycles, across different topological objects remains a fundamental challenge in persistent homology and topological data analysis. This work introduces a novel framework for constructing cycle communities through two complementary approaches. First, a dendrogram-based methodology leverages merge-tree algorithms to construct hierarchical representations of homology classes from persistence intervals. The Wasserstein distance on merge trees is introduced as a metric for comparing dendrograms, establishing connections to hierarchical clustering frameworks. Through simulation studies, the discriminative power of dendrogram representations for identifying cycle communities is demonstrated. Second, an extension of Stratified Gradient Sampling simultaneously learns multiple filter functions that yield cycle barycenter functions capable of faithfully reconstructing distinct sets of cycles. The set of cycles each filter function can reconstruct constitutes cycle communities that are non-overlapping and partition the space of all cycles. Together, these approaches transform the problem of cycle matching into both a hierarchical clustering and topological optimization framework, providing principled methods to identify similar topological structures both within and across groups of topological objects. --- ### Towards Open Standards for Systemic Complexity in Digital Forensics Authors: Paola Di Maio Published: 2025-12-15 Categories: cs.AI, cs.CY arXiv: https://arxiv.org/abs/2512.12970v1 The intersection of artificial intelligence (AI) and digital forensics (DF) is becoming increasingly complex, ubiquitous, and pervasive, with overlapping techniques and technologies being adopted in all types of scientific and technical inquiry. Despite incredible advances, forensic sciences are not exempt from errors and remain vulnerable to fallibility. To mitigate the limitations of errors in DF, the systemic complexity is identified and addressed with the adoption of human-readable artifacts and open standards. A DF AI model schema based on the state of the art is outlined. --- ### QwenLong-L1.5: Post-Training Recipe for Long-Context Reasoning and Memory Management Authors: Weizhou Shen, Ziyi Yang, Chenliang Li, Zhiyuan Lu, Miao Peng, Huashan Sun, Yingcheng Shi, Shengyi Liao, Shaopeng Lai, Bo Zhang, Dayiheng Liu, Fei Huang et al. Published: 2025-12-15 Categories: cs.CL arXiv: https://arxiv.org/abs/2512.12967v1 We introduce QwenLong-L1.5, a model that achieves superior long-context reasoning capabilities through systematic post-training innovations. The key technical breakthroughs of QwenLong-L1.5 are as follows: (1) Long-Context Data Synthesis Pipeline: We develop a systematic synthesis framework that generates challenging reasoning tasks requiring multi-hop grounding over globally distributed evidence. By deconstructing documents into atomic facts and their underlying relationships, and then programmatically composing verifiable reasoning questions, our approach creates high-quality training data at scale, moving substantially beyond simple retrieval tasks to enable genuine long-range reasoning capabilities. (2) Stabilized Reinforcement Learning for Long-Context Training: To overcome the critical instability in long-context RL, we introduce task-balanced sampling with task-specific advantage estimation to mitigate reward bias, and propose Adaptive Entropy-Controlled Policy Optimization (AEPO) that dynamically regulates exploration-exploitation trade-offs. (3) Memory-Augmented Architecture for Ultra-Long Contexts: Recognizing that even extended context windows cannot accommodate arbitrarily long sequences, we develop a memory management framework with multi-stage fusion RL training that seamlessly integrates single-pass reasoning with iterative memory-based processing for tasks exceeding 4M tokens. Based on Qwen3-30B-A3B-Thinking, QwenLong-L1.5 achieves performance comparable to GPT-5 and Gemini-2.5-Pro on long-context reasoning benchmarks, surpassing its baseline by 9.90 points on average. On ultra-long tasks (1M~4M tokens), QwenLong-L1.5's memory-agent framework yields a 9.48-point gain over the agent baseline. Additionally, the acquired long-context reasoning ability translates to enhanced performance in general domains like scientific reasoning, memory tool using, and extended dialogue. --- ### SCAdapter: Content-Style Disentanglement for Diffusion Style Transfer Authors: Luan Thanh Trinh, Kenji Doi, Atsuki Osanai Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.12963v1 Diffusion models have emerged as the leading approach for style transfer, yet they struggle with photo-realistic transfers, often producing painting-like results or missing detailed stylistic elements. Current methods inadequately address unwanted influence from original content styles and style reference content features. We introduce SCAdapter, a novel technique leveraging CLIP image space to effectively separate and integrate content and style features. Our key innovation systematically extracts pure content from content images and style elements from style references, ensuring authentic transfers. This approach is enhanced through three components: Controllable Style Adaptive Instance Normalization (CSAdaIN) for precise multi-style blending, KVS Injection for targeted style integration, and a style transfer consistency objective maintaining process coherence. Comprehensive experiments demonstrate SCAdapter significantly outperforms state-of-the-art methods in both conventional and diffusion-based baselines. By eliminating DDIM inversion and inference-stage optimization, our method achieves at least $2\times$ faster inference than other diffusion-based approaches, making it both more effective and efficient for practical applications. --- ### Leveraging Compression to Construct Transferable Bitrate Ladders Authors: Krishna Srikar Durbha, Hassene Tmar, Ping-Hao Wu, Ioannis Katsavounidis, Alan C. Bovik Published: 2025-12-15 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2512.12952v1 Over the past few years, per-title and per-shot video encoding techniques have demonstrated significant gains as compared to conventional techniques such as constant CRF encoding and the fixed bitrate ladder. These techniques have demonstrated that constructing content-gnostic per-shot bitrate ladders can provide significant bitrate gains and improved Quality of Experience (QoE) for viewers under various network conditions. However, constructing a convex hull for every video incurs a significant computational overhead. Recently, machine learning-based bitrate ladder construction techniques have emerged as a substitute for convex hull construction. These methods operate by extracting features from source videos to train machine learning (ML) models to construct content-adaptive bitrate ladders. Here, we present a new ML-based bitrate ladder construction technique that accurately predicts the VMAF scores of compressed videos, by analyzing the compression procedure and by making perceptually relevant measurements on the source videos prior to compression. We evaluate the performance of our proposed framework against leading prior methods on a large corpus of videos. Since training ML models on every encoder setting is time-consuming, we also investigate how per-shot bitrate ladders perform under different encoding settings. We evaluate the performance of all models against the fixed bitrate ladder and the best possible convex hull constructed using exhaustive encoding with Bjontegaard-delta metrics. --- ### Building from Scratch: A Multi-Agent Framework with Human-in-the-Loop for Multilingual Legal Terminology Mapping Authors: Lingyi Meng, Maolin Liu, Hao Wang, Yilan Cheng, Qi Yang, Idlkaid Mohanmmed Published: 2025-12-15 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.12950v1 Accurately mapping legal terminology across languages remains a significant challenge, especially for language pairs like Chinese and Japanese, which share a large number of homographs with different meanings. Existing resources and standardized tools for these languages are limited. To address this, we propose a human-AI collaborative approach for building a multilingual legal terminology database, based on a multi-agent framework. This approach integrates advanced large language models and legal domain experts throughout the entire process-from raw document preprocessing, article-level alignment, to terminology extraction, mapping, and quality assurance. Unlike a single automated pipeline, our approach places greater emphasis on how human experts participate in this multi-agent system. Humans and AI agents take on different roles: AI agents handle specific, repetitive tasks, such as OCR, text segmentation, semantic alignment, and initial terminology extraction, while human experts provide crucial oversight, review, and supervise the outputs with contextual knowledge and legal judgment. We tested the effectiveness of this framework using a trilingual parallel corpus comprising 35 key Chinese statutes, along with their English and Japanese translations. The experimental results show that this human-in-the-loop, multi-agent workflow not only improves the precision and consistency of multilingual legal terminology mapping but also offers greater scalability compared to traditional manual methods. --- ### Understanding When Graph Convolutional Networks Help: A Diagnostic Study on Label Scarcity and Structural Properties Authors: Nischal Subedi, Ember Kerstetter, Winnie Li, Silo Murphy Published: 2025-12-15 Categories: cs.LG, stat.AP, stat.ML arXiv: https://arxiv.org/abs/2512.12947v1 Graph Convolutional Networks (GCNs) have become a standard approach for semi-supervised node classification, yet practitioners lack clear guidance on when GCNs provide meaningful improvements over simpler baselines. We present a diagnostic study using the Amazon Computers co-purchase data to understand when and why GCNs help. Through systematic experiments with simulated label scarcity, feature ablation, and per-class analysis, we find that GCN performance depends critically on the interaction between graph homophily and feature quality. GCNs provide the largest gains under extreme label scarcity, where they leverage neighborhood structure to compensate for limited supervision. Surprisingly, GCNs can match their original performance even when node features are replaced with random noise, suggesting that structure alone carries sufficient signal on highly homophilous graphs. However, GCNs hurt performance when homophily is low and features are already strong, as noisy neighbors corrupt good predictions. Our quadrant analysis reveals that GCNs help in three of four conditions and only hurt when low homophily meets strong features. These findings offer practical guidance for practitioners deciding whether to adopt graph-based methods. --- ### SLIM-VDB: A Real-Time 3D Probabilistic Semantic Mapping Framework Authors: Anja Sheppard, Parker Ewen, Joey Wilson, Advaith V. Sethuraman, Benard Adewole, Anran Li, Yuzhen Chen, Ram Vasudevan, Katherine A. Skinner Published: 2025-12-15 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2512.12945v1 This paper introduces SLIM-VDB, a new lightweight semantic mapping system with probabilistic semantic fusion for closed-set or open-set dictionaries. Advances in data structures from the computer graphics community, such as OpenVDB, have demonstrated significantly improved computational and memory efficiency in volumetric scene representation. Although OpenVDB has been used for geometric mapping in robotics applications, semantic mapping for scene understanding with OpenVDB remains unexplored. In addition, existing semantic mapping systems lack support for integrating both fixed-category and open-language label predictions within a single framework. In this paper, we propose a novel 3D semantic mapping system that leverages the OpenVDB data structure and integrates a unified Bayesian update framework for both closed- and open-set semantic fusion. Our proposed framework, SLIM-VDB, achieves significant reduction in both memory and integration times compared to current state-of-the-art semantic mapping approaches, while maintaining comparable mapping accuracy. An open-source C++ codebase with a Python interface is available at https://github.com/umfieldrobotics/slim-vdb. --- ### UAGLNet: Uncertainty-Aggregated Global-Local Fusion Network with Cooperative CNN-Transformer for Building Extraction Authors: Siyuan Yao, Dongxiu Liu, Taotao Li, Shengjie Li, Wenqi Ren, Xiaochun Cao Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.12941v1 Building extraction from remote sensing images is a challenging task due to the complex structure variations of the buildings. Existing methods employ convolutional or self-attention blocks to capture the multi-scale features in the segmentation models, while the inherent gap of the feature pyramids and insufficient global-local feature integration leads to inaccurate, ambiguous extraction results. To address this issue, in this paper, we present an Uncertainty-Aggregated Global-Local Fusion Network (UAGLNet), which is capable to exploit high-quality global-local visual semantics under the guidance of uncertainty modeling. Specifically, we propose a novel cooperative encoder, which adopts hybrid CNN and transformer layers at different stages to capture the local and global visual semantics, respectively. An intermediate cooperative interaction block (CIB) is designed to narrow the gap between the local and global features when the network becomes deeper. Afterwards, we propose a Global-Local Fusion (GLF) module to complementarily fuse the global and local representations. Moreover, to mitigate the segmentation ambiguity in uncertain regions, we propose an Uncertainty-Aggregated Decoder (UAD) to explicitly estimate the pixel-wise uncertainty to enhance the segmentation accuracy. Extensive experiments demonstrate that our method achieves superior performance to other state-of-the-art methods. Our code is available at https://github.com/Dstate/UAGLNet --- ### Continuous Edit Distance, Geodesics and Barycenters of Time-varying Persistence Diagrams Authors: Sebastien Tchitchek, Mohamed Kissi, Julien Tierny Published: 2025-12-15 Categories: cs.CG, cs.CV, cs.GR, cs.LG arXiv: https://arxiv.org/abs/2512.12939v1 We introduce the Continuous Edit Distance (CED), a geodesic and elastic distance for time-varying persistence diagrams (TVPDs). The CED extends edit-distance ideas to TVPDs by combining local substitution costs with penalized deletions/insertions, controlled by two parameters: \(α\) (trade-off between temporal misalignment and diagram discrepancy) and \(β\) (gap penalty). We also provide an explicit construction of CED-geodesics. Building on these ingredients, we present two practical barycenter solvers, one stochastic and one greedy, that monotonically decrease the CED Frechet energy. Empirically, the CED is robust to additive perturbations (both temporal and spatial), recovers temporal shifts, and supports temporal pattern search. On real-life datasets, the CED achieves clustering performance comparable or better than standard elastic dissimilarities, while our clustering based on CED-barycenters yields superior classification results. Overall, the CED equips TVPD analysis with a principled distance, interpretable geodesics, and practical barycenters, enabling alignment, comparison, averaging, and clustering directly in the space of TVPDs. A C++ implementation is provided for reproducibility at the following address https://github.com/sebastien-tchitchek/ContinuousEditDistance. --- ### Content Adaptive based Motion Alignment Framework for Learned Video Compression Authors: Tiange Zhang, Xiandong Meng, Siwei Ma Published: 2025-12-15 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.12936v1 Recent advances in end-to-end video compression have shown promising results owing to their unified end-to-end learning optimization. However, such generalized frameworks often lack content-specific adaptation, leading to suboptimal compression performance. To address this, this paper proposes a content adaptive based motion alignment framework that improves performance by adapting encoding strategies to diverse content characteristics. Specifically, we first introduce a two-stage flow-guided deformable warping mechanism that refines motion compensation with coarse-to-fine offset prediction and mask modulation, enabling precise feature alignment. Second, we propose a multi-reference quality aware strategy that adjusts distortion weights based on reference quality, and applies it to hierarchical training to reduce error propagation. Third, we integrate a training-free module that downsamples frames by motion magnitude and resolution to obtain smooth motion estimation. Experimental results on standard test datasets demonstrate that our framework CAMA achieves significant improvements over state-of-the-art Neural Video Compression models, achieving a 24.95% BD-rate (PSNR) savings over our baseline model DCVC-TCM, while also outperforming reproduced DCVC-DC and traditional codec HM-16.25. --- ### Unified Interactive Multimodal Moment Retrieval via Cascaded Embedding-Reranking and Temporal-Aware Score Fusion Authors: Toan Le Ngo Thanh, Phat Ha Huu, Tan Nguyen Dang Duy, Thong Nguyen Le Minh, Anh Nguyen Nhu Tinh Published: 2025-12-15 Categories: cs.CV, cs.AI, cs.IR arXiv: https://arxiv.org/abs/2512.12935v1 The exponential growth of video content has created an urgent need for efficient multimodal moment retrieval systems. However, existing approaches face three critical challenges: (1) fixed-weight fusion strategies fail across cross modal noise and ambiguous queries, (2) temporal modeling struggles to capture coherent event sequences while penalizing unrealistic gaps, and (3) systems require manual modality selection, reducing usability. We propose a unified multimodal moment retrieval system with three key innovations. First, a cascaded dual-embedding pipeline combines BEIT-3 and SigLIP for broad retrieval, refined by BLIP-2 based reranking to balance recall and precision. Second, a temporal-aware scoring mechanism applies exponential decay penalties to large temporal gaps via beam search, constructing coherent event sequences rather than isolated frames. Third, Agent-guided query decomposition (GPT-4o) automatically interprets ambiguous queries, decomposes them into modality specific sub-queries (visual/OCR/ASR), and performs adaptive score fusion eliminating manual modality selection. Qualitative analysis demonstrates that our system effectively handles ambiguous queries, retrieves temporally coherent sequences, and dynamically adapts fusion strategies, advancing interactive moment search capabilities. --- ### Investigating Data Pruning for Pretraining Biological Foundation Models at Scale Authors: Yifan Wu, Jiyue Jiang, Xichen Ye, Yiqi Wang, Chang Zhou, Yitao Xu, Jiayang Chen, He Hu, Weizhong Zhang, Cheng Jin, Jiao Yuan, Yu Li Published: 2025-12-15 Categories: cs.LG, cs.AI, cs.CE arXiv: https://arxiv.org/abs/2512.12932v2 Biological foundation models (BioFMs), pretrained on large-scale biological sequences, have recently shown strong potential in providing meaningful representations for diverse downstream bioinformatics tasks. However, such models often rely on millions to billions of training sequences and billions of parameters, resulting in prohibitive computational costs and significant barriers to reproducibility and accessibility, particularly for academic labs. To address these challenges, we investigate the feasibility of data pruning for BioFM pretraining and propose a post-hoc influence-guided data pruning framework tailored to biological domains. Our approach introduces a subset-based self-influence formulation that enables efficient estimation of sample importance at low computational cost, and builds upon it two simple yet effective selection strategies, namely Top-k Influence (Top I) and Coverage-Centric Influence (CCI). We empirically validate our method on two representative BioFMs, RNA-FM and ESM-C. For RNA, our framework consistently outperforms random selection baselines under an extreme pruning rate of over 99 percent, demonstrating its effectiveness. Furthermore, we show the generalizability of our framework on protein-related tasks using ESM-C. In particular, our coreset even outperforms random subsets that are ten times larger in both RNA and protein settings, revealing substantial redundancy in biological sequence datasets. These findings underscore the potential of influence-guided data pruning to substantially reduce the computational cost of BioFM pretraining, paving the way for more efficient, accessible, and sustainable biological AI research. --- ### SeVeDo: A Heterogeneous Transformer Accelerator for Low-Bit Inference via Hierarchical Group Quantization and SVD-Guided Mixed Precision Authors: Yuseon Choi, Sangjin Kim, Jungjun Oh, Byeongcheol Kim, Hoi-Jun Yoo Published: 2025-12-15 Categories: cs.LG, cs.AR arXiv: https://arxiv.org/abs/2512.12930v2 Low-bit quantization is a promising technique for efficient transformer inference by reducing computational and memory overhead. However, aggressive bitwidth reduction remains challenging due to activation outliers, leading to accuracy degradation. Existing methods, such as outlier-handling and group quantization, achieve high accuracy but incur substantial energy consumption. To address this, we propose SeVeDo, an energy-efficient SVD-based heterogeneous accelerator that structurally separates outlier-sensitive components into a high-precision low-rank path, while the remaining computations are executed in a low-bit residual datapath with group quantization. To further enhance efficiency, Hierarchical Group Quantization (HGQ) combines coarse-grained floating-point scaling with fine-grained shifting, effectively reducing dequantization cost. Also, SVD-guided mixed precision (SVD-MP) statically allocates higher bitwidths to precision-sensitive components identified through low-rank decomposition, thereby minimizing floating-point operation cost. Experimental results show that SeVeDo achieves a peak energy efficiency of 13.8TOPS/W, surpassing conventional designs, with 12.7TOPS/W on ViT-Base and 13.4TOPS/W on Llama2-7B benchmarks. --- ### MADTempo: An Interactive System for Multi-Event Temporal Video Retrieval with Query Augmentation Authors: Huu-An Vu, Van-Khanh Mai, Trong-Tam Nguyen, Quang-Duc Dam, Tien-Huy Nguyen, Thanh-Huong Le Published: 2025-12-15 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.12929v1 The rapid expansion of video content across online platforms has accelerated the need for retrieval systems capable of understanding not only isolated visual moments but also the temporal structure of complex events. Existing approaches often fall short in modeling temporal dependencies across multiple events and in handling queries that reference unseen or rare visual concepts. To address these challenges, we introduce MADTempo, a video retrieval framework developed by our team, AIO_Trinh, that unifies temporal search with web-scale visual grounding. Our temporal search mechanism captures event-level continuity by aggregating similarity scores across sequential video segments, enabling coherent retrieval of multi-event queries. Complementarily, a Google Image Search-based fallback module expands query representations with external web imagery, effectively bridging gaps in pretrained visual embeddings and improving robustness against out-of-distribution (OOD) queries. Together, these components advance the temporal reasoning and generalization capabilities of modern video retrieval systems, paving the way for more semantically aware and adaptive retrieval across large-scale video corpora. --- ### Sharpness-aware Dynamic Anchor Selection for Generalized Category Discovery Authors: Zhimao Peng, Enguang Wang, Fei Yang, Xialei Liu, Ming-Ming Cheng Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.12925v1 Generalized category discovery (GCD) is an important and challenging task in open-world learning. Specifically, given some labeled data of known classes, GCD aims to cluster unlabeled data that contain both known and unknown classes. Current GCD methods based on parametric classification adopt the DINO-like pseudo-labeling strategy, where the sharpened probability output of one view is used as supervision information for the other view. However, large pre-trained models have a preference for some specific visual patterns, resulting in encoding spurious correlation for unlabeled data and generating noisy pseudo-labels. To address this issue, we propose a novel method, which contains two modules: Loss Sharpness Penalty (LSP) and Dynamic Anchor Selection (DAS). LSP enhances the robustness of model parameters to small perturbations by minimizing the worst-case loss sharpness of the model, which suppressing the encoding of trivial features, thereby reducing overfitting of noise samples and improving the quality of pseudo-labels. Meanwhile, DAS selects representative samples for the unknown classes based on KNN density and class probability during the model training and assigns hard pseudo-labels to them, which not only alleviates the confidence difference between known and unknown classes but also enables the model to quickly learn more accurate feature distribution for the unknown classes, thus further improving the clustering accuracy. Extensive experiments demonstrate that the proposed method can effectively mitigate the noise of pseudo-labels, and achieve state-of-the-art results on multiple GCD benchmarks. --- ### Interpretable Hypothesis-Driven Trading:A Rigorous Walk-Forward Validation Framework for Market Microstructure Signals Authors: Gagan Deep, Akash Deep, William Lamptey Published: 2025-12-15 Categories: q-fin.TR, q-fin.CP, stat.ML arXiv: https://arxiv.org/abs/2512.12924v1 We develop a rigorous walk-forward validation framework for algorithmic trading designed to mitigate overfitting and lookahead bias. Our methodology combines interpretable hypothesis-driven signal generation with reinforcement learning and strict out-of-sample testing. The framework enforces strict information set discipline, employs rolling window validation across 34 independent test periods, maintains complete interpretability through natural language hypothesis explanations, and incorporates realistic transaction costs and position constraints. Validating five market microstructure patterns across 100 US equities from 2015 to 2024, the system yields modest annualized returns (0.55%, Sharpe ratio 0.33) with exceptional downside protection (maximum drawdown -2.76%) and market-neutral characteristics (beta = 0.058). Performance exhibits strong regime dependence, generating positive returns during high-volatility periods (0.60% quarterly, 2020-2024) while underperforming in stable markets (-0.16%, 2015-2019). We report statistically insignificant aggregate results (p-value 0.34) to demonstrate a reproducible, honest validation protocol that prioritizes interpretability and extends naturally to advanced hypothesis generators, including large language models. The key empirical finding reveals that daily OHLCV-based microstructure signals require elevated information arrival and trading activity to function effectively. The framework provides complete mathematical specifications and open-source implementation, establishing a template for rigorous trading system evaluation that addresses the reproducibility crisis in quantitative finance research. For researchers, practitioners, and regulators, this work demonstrates that interpretable algorithmic trading strategies can be rigorously validated without sacrificing transparency or regulatory compliance. --- ### LLM-based Personalized Portfolio Recommender: Integrating Large Language Models and Reinforcement Learning for Intelligent Investment Strategy Optimization Authors: Bangyu Li, Boping Gu, Ziyang Ding Published: 2025-12-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.12922v1 In modern financial markets, investors increasingly seek personalized and adaptive portfolio strategies that reflect their individual risk preferences and respond to dynamic market conditions. Traditional rule-based or static optimization approaches often fail to capture the nonlinear interactions among investor behavior, market volatility, and evolving financial objectives. To address these limitations, this paper introduces the LLM-based Personalized Portfolio Recommender , an integrated framework that combines Large Language Models, reinforcement learning, and individualized risk preference modeling to support intelligent investment decision-making. --- ### Cisco Integrated AI Security and Safety Framework Report Authors: Amy Chang, Tiffany Saade, Sanket Mendapara, Adam Swanda, Ankit Garg Published: 2025-12-15 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2512.12921v1 Artificial intelligence (AI) systems are being readily and rapidly adopted, increasingly permeating critical domains: from consumer platforms and enterprise software to networked systems with embedded agents. While this has unlocked potential for human productivity gains, the attack surface has expanded accordingly: threats now span content safety failures (e.g., harmful or deceptive outputs), model and data integrity compromise (e.g., poisoning, supply-chain tampering), runtime manipulations (e.g., prompt injection, tool and agent misuse), and ecosystem risks (e.g., orchestration abuse, multi-agent collusion). Existing frameworks such as MITRE ATLAS, National Institute of Standards and Technology (NIST) AI 100-2 Adversarial Machine Learning (AML) taxonomy, and OWASP Top 10s for Large Language Models (LLMs) and Agentic AI Applications provide valuable viewpoints, but each covers only slices of this multi-dimensional space. This paper presents Cisco's Integrated AI Security and Safety Framework ("AI Security Framework"), a unified, lifecycle-aware taxonomy and operationalization framework that can be used to classify, integrate, and operationalize the full range of AI risks. It integrates AI security and AI safety across modalities, agents, pipelines, and the broader ecosystem. The AI Security Framework is designed to be practical for threat identification, red-teaming, risk prioritization, and it is comprehensive in scope and can be extensible to emerging deployments in multimodal contexts, humanoids, wearables, and sensory infrastructures. We analyze gaps in prevailing frameworks, discuss design principles for our framework, and demonstrate how the taxonomy provides structure for understanding how modern AI systems fail, how adversaries exploit these failures, and how organizations can build defenses across the AI lifecycle that evolve alongside capability advancements. --- ### Satisfiability Modulo Theory Meets Inductive Logic Programming Authors: Nijesh Upreti, Vaishak Belle Published: 2025-12-15 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.12918v1 Inductive Logic Programming (ILP) provides interpretable rule learning in relational domains, yet remains limited in its ability to induce and reason with numerical constraints. Classical ILP systems operate over discrete predicates and typically rely on discretisation or hand-crafted numerical predicates, making it difficult to infer thresholds or arithmetic relations that must hold jointly across examples. Recent work has begun to address these limitations through tighter integrations of ILP with Satisfiability Modulo Theories (SMT) or specialised numerical inference mechanisms. In this paper we investigate a modular alternative that couples the ILP system PyGol with the SMT solver Z3. Candidate clauses proposed by PyGol are interpreted as quantifier-free formulas over background theories such as linear or nonlinear real arithmetic, allowing numerical parameters to be instantiated and verified by the SMT solver while preserving ILP's declarative relational bias. This supports the induction of hybrid rules that combine symbolic predicates with learned numerical constraints, including thresholds, intervals, and multi-literal arithmetic relations. We formalise this SMT-ILP setting and evaluate it on a suite of synthetic datasets designed to probe linear, relational, nonlinear, and multi-hop reasoning. The results illustrate how a modular SMT-ILP architecture can extend the expressivity of symbolic rule learning, complementing prior numerical ILP approaches while providing a flexible basis for future extensions toward richer theory-aware induction. --- ### CTIGuardian: A Few-Shot Framework for Mitigating Privacy Leakage in Fine-Tuned LLMs Authors: Shashie Dilhara Batan Arachchige, Benjamin Zi Hao Zhao, Hassan Jameel Asghar, Dinusha Vatsalan, Dali Kaafar Published: 2025-12-15 Categories: cs.CR, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2512.12914v1 Large Language Models (LLMs) are often fine-tuned to adapt their general-purpose knowledge to specific tasks and domains such as cyber threat intelligence (CTI). Fine-tuning is mostly done through proprietary datasets that may contain sensitive information. Owners expect their fine-tuned model to not inadvertently leak this information to potentially adversarial end users. Using CTI as a use case, we demonstrate that data-extraction attacks can recover sensitive information from fine-tuned models on CTI reports, underscoring the need for mitigation. Retraining the full model to eliminate this leakage is computationally expensive and impractical. We propose an alternative approach, which we call privacy alignment, inspired by safety alignment in LLMs. Just like safety alignment teaches the model to abide by safety constraints through a few examples, we enforce privacy alignment through few-shot supervision, integrating a privacy classifier and a privacy redactor, both handled by the same underlying LLM. We evaluate our system, called CTIGuardian, using GPT-4o mini and Mistral-7B Instruct models, benchmarking against Presidio, a named entity recognition (NER) baseline. Results show that CTIGuardian provides a better privacy-utility trade-off than NER based models. While we demonstrate its effectiveness on a CTI use case, the framework is generic enough to be applicable to other sensitive domains. --- ### Evaluating Singular Value Thresholds for DNN Weight Matrices based on Random Matrix Theory Authors: Kohei Nishikawa, Koki Shimizu, Hiroki Hashiguchi Published: 2025-12-15 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2512.12911v2 This study evaluates thresholds for removing singular values from singular value decomposition-based low-rank approximations of deep neural network weight matrices. Each weight matrix is modeled as the sum of signal and noise matrices. The low-rank approximation is obtained by removing noise-related singular values using a threshold based on random matrix theory. To assess the adequacy of this threshold, we propose an evaluation metric based on the cosine similarity between the singular vectors of the signal and original weight matrices. The proposed metric is used in numerical experiments to compare two threshold estimation methods. --- ### Machine Learning Architectures for the Estimation of Predicted Occupancy Grids in Road Traffic Authors: Parthasarathy Nadarajan, Michael Botsch, Sebastian Sardina Published: 2025-12-15 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.12907v1 This paper introduces a novel machine learning architecture for an efficient estimation of the probabilistic space-time representation of complex traffic scenarios. A detailed representation of the future traffic scenario is of significant importance for autonomous driving and for all active safety systems. In order to predict the future space-time representation of the traffic scenario, first the type of traffic scenario is identified and then the machine learning algorithm maps the current state of the scenario to possible future states. The input to the machine learning algorithms is the current state representation of a traffic scenario, termed as the Augmented Occupancy Grid (AOG). The output is the probabilistic space-time representation which includes uncertainties regarding the behaviour of the traffic participants and is termed as the Predicted Occupancy Grid (POG). The novel architecture consists of two Stacked Denoising Autoencoders (SDAs) and a set of Random Forests. It is then compared with the other two existing architectures that comprise of SDAs and DeconvNet. The architectures are validated with the help of simulations and the comparisons are made both in terms of accuracy and computational time. Also, a brief overview on the applications of POGs in the field of active safety is presented. --- ### Predictive Sample Assignment for Semantically Coherent Out-of-Distribution Detection Authors: Zhimao Peng, Enguang Wang, Xialei Liu, Ming-Ming Cheng Published: 2025-12-15 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.12906v1 Semantically coherent out-of-distribution detection (SCOOD) is a recently proposed realistic OOD detection setting: given labeled in-distribution (ID) data and mixed in-distribution and out-of-distribution unlabeled data as the training data, SCOOD aims to enable the trained model to accurately identify OOD samples in the testing data. Current SCOOD methods mainly adopt various clustering-based in-distribution sample filtering (IDF) strategies to select clean ID samples from unlabeled data, and take the remaining samples as auxiliary OOD data, which inevitably introduces a large number of noisy samples in training. To address the above issue, we propose a concise SCOOD framework based on predictive sample assignment (PSA). PSA includes a dual-threshold ternary sample assignment strategy based on the predictive energy score that can significantly improve the purity of the selected ID and OOD sample sets by assigning unconfident unlabeled data to an additional discard sample set, and a concept contrastive representation learning loss to further expand the distance between ID and OOD samples in the representation space to assist ID/OOD discrimination. In addition, we also introduce a retraining strategy to help the model fully fit the selected auxiliary ID/OOD samples. Experiments on two standard SCOOD benchmarks demonstrate that our approach outperforms the state-of-the-art methods by a significant margin. --- ### PAC-Bayes Bounds for Multivariate Linear Regression and Linear Autoencoders Authors: Ruixin Guo, Ruoming Jin, Xinyu Li, Yang Zhou Published: 2025-12-15 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2512.12905v1 Linear Autoencoders (LAEs) have shown strong performance in state-of-the-art recommender systems. However, this success remains largely empirical, with limited theoretical understanding. In this paper, we investigate the generalizability -- a theoretical measure of model performance in statistical learning -- of multivariate linear regression and LAEs. We first propose a PAC-Bayes bound for multivariate linear regression, extending the earlier bound for single-output linear regression by Shalaeva et al., and establish sufficient conditions for its convergence. We then show that LAEs, when evaluated under a relaxed mean squared error, can be interpreted as constrained multivariate linear regression models on bounded data, to which our bound adapts. Furthermore, we develop theoretical methods to improve the computational efficiency of optimizing the LAE bound, enabling its practical evaluation on large models and real-world datasets. Experimental results demonstrate that our bound is tight and correlates well with practical ranking metrics such as Recall@K and NDCG@K. ---