# Research Papers — 2025-12 · bundle page 2 of 175 > Full-text markdown of papers #51–100 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/2 [← Previous bundle](https://aipod.dev/papers/archive/2025-12.md) [Next bundle →](https://aipod.dev/papers/archive/2025-12/page/3.md) ### RIMRULE: Improving Tool-Using Language Agents via MDL-Guided Rule Learning Authors: Xiang Gao, Yuguang Yao, Qi Zhang, Kaiwen Dong, Avinash Baidya, Ruocheng Guo, Hilaf Hasson, Kamalika Das Published: 2025-12-31 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.00086v3 Large language models (LLMs) often struggle to use tools reliably in domain-specific settings, where APIs may be idiosyncratic, under-documented, or tailored to private workflows. This highlights the need for effective adaptation to task-specific tools. We propose RIMRULE, a neuro-symbolic approach for LLM adaptation based on dynamic rule injection. Compact, interpretable rules are distilled from failure traces and injected into the prompt during inference to improve task performance. These rules are proposed by the LLM itself and consolidated using a Minimum Description Length (MDL) objective that favors generality and conciseness. Each rule is stored in both natural language and a structured symbolic form, supporting efficient retrieval at inference time. Experiments on tool-use benchmarks show that this approach improves accuracy on both seen and unseen tools without modifying LLM weights. It outperforms prompting-based adaptation methods and complements finetuning. Moreover, rules learned from one LLM can be reused to improve others, including long reasoning LLMs, highlighting the portability of symbolic knowledge across architectures. --- ### Exploration in the Limit Authors: Brian M. Cho, Nathan Kallus Published: 2025-12-31 Categories: cs.LG, math.ST, stat.ME arXiv: https://arxiv.org/abs/2601.00084v1 In fixed-confidence best arm identification (BAI), the objective is to quickly identify the optimal option while controlling the probability of error below a desired threshold. Despite the plethora of BAI algorithms, existing methods typically fall short in practical settings, as stringent exact error control requires using loose tail inequalities and/or parametric restrictions. To overcome these limitations, we introduce a relaxed formulation that requires valid error control asymptotically with respect to a minimum sample size. This aligns with many real-world settings that often involve weak signals, high desired significance, and post-experiment inference requirements, all of which necessitate long horizons. This allows us to achieve tighter optimality, while better handling flexible nonparametric outcome distributions and fully leveraging individual-level contexts. We develop a novel asymptotic anytime-valid confidence sequences over arm indices, and we use it to design a new BAI algorithm for our asymptotic framework. Our method flexibly incorporates covariates for variance reduction and ensures approximate error control in fully nonparametric settings. Under mild convergence assumptions, we provide asymptotic bounds on the sample complexity and show the worst-case sample complexity of our approach matches the best-case sample complexity of Gaussian BAI under exact error guarantees and known variances. Experiments suggest our approach reduces average sample complexities while maintaining error control. --- ### Cuffless, calibration-free hemodynamic monitoring with physics-informed machine learning models Authors: Henry Crandall, Tyler Schuessler, Filip Bělík, Albert Fabregas, Barry M. Stults, Alexandra Boyadzhiev, Huanan Zhang, Jim S. Wu, Aylin R. Rodan, Stephen P. Juraschek, Ramakrishna Mukkamala, Alfred K. Cheung et al. Published: 2025-12-31 Categories: physics.med-ph, cs.LG arXiv: https://arxiv.org/abs/2601.00081v1 Wearable technologies have the potential to transform ambulatory and at-home hemodynamic monitoring by providing continuous assessments of cardiovascular health metrics and guiding clinical management. However, existing cuffless wearable devices for blood pressure (BP) monitoring often rely on methods lacking theoretical foundations, such as pulse wave analysis or pulse arrival time, making them vulnerable to physiological and experimental confounders that undermine their accuracy and clinical utility. Here, we developed a smartwatch device with real-time electrical bioimpedance (BioZ) sensing for cuffless hemodynamic monitoring. We elucidate the biophysical relationship between BioZ and BP via a multiscale analytical and computational modeling framework, and identify physiological, anatomical, and experimental parameters that influence the pulsatile BioZ signal at the wrist. A signal-tagged physics-informed neural network incorporating fluid dynamics principles enables calibration-free estimation of BP and radial and axial blood velocity. We successfully tested our approach with healthy individuals at rest and after physical activity including physical and autonomic challenges, and with patients with hypertension and cardiovascular disease in outpatient and intensive care settings. Our findings demonstrate the feasibility of BioZ technology for cuffless BP and blood velocity monitoring, addressing critical limitations of existing cuffless technologies. --- ### IMBWatch -- a Spatio-Temporal Graph Neural Network approach to detect Illicit Massage Business Authors: Swetha Varadarajan, Abhishek Ray, Lumina Albert Published: 2025-12-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.00075v1 Illicit Massage Businesses (IMBs) are a covert and persistent form of organized exploitation that operate under the facade of legitimate wellness services while facilitating human trafficking, sexual exploitation, and coerced labor. Detecting IMBs is difficult due to encoded digital advertisements, frequent changes in personnel and locations, and the reuse of shared infrastructure such as phone numbers and addresses. Traditional approaches, including community tips and regulatory inspections, are largely reactive and ineffective at revealing the broader operational networks traffickers rely on. To address these challenges, we introduce IMBWatch, a spatio-temporal graph neural network (ST-GNN) framework for large-scale IMB detection. IMBWatch constructs dynamic graphs from open-source intelligence, including scraped online advertisements, business license records, and crowdsourced reviews. Nodes represent heterogeneous entities such as businesses, aliases, phone numbers, and locations, while edges capture spatio-temporal and relational patterns, including co-location, repeated phone usage, and synchronized advertising. The framework combines graph convolutional operations with temporal attention mechanisms to model the evolution of IMB networks over time and space, capturing patterns such as intercity worker movement, burner phone rotation, and coordinated advertising surges. Experiments on real-world datasets from multiple U.S. cities show that IMBWatch outperforms baseline models, achieving higher accuracy and F1 scores. Beyond performance gains, IMBWatch offers improved interpretability, providing actionable insights to support proactive and targeted interventions. The framework is scalable, adaptable to other illicit domains, and released with anonymized data and open-source code to support reproducible research. --- ### Automated electrostatic characterization of quantum dot devices in single- and bilayer heterostructures Authors: Merritt P. R. Losert, Dario Denora, Barnaby van Straaten, Michael Chan, Stefan D. Oosterhout, Lucas Stehouwer, Giordano Scappucci, Menno Veldhorst, Justyna P. Zwolak Published: 2025-12-31 Categories: cond-mat.mes-hall, cs.CV, cs.ET, cs.LG, quant-ph arXiv: https://arxiv.org/abs/2601.00067v1 As quantum dot (QD)-based spin qubits advance toward larger, more complex device architectures, rapid, automated device characterization and data analysis tools become critical. The orientation and spacing of transition lines in a charge stability diagram (CSD) contain a fingerprint of a QD device's capacitive environment, making these measurements useful tools for device characterization. However, manually interpreting these features is time-consuming, error-prone, and impractical at scale. Here, we present an automated protocol for extracting underlying capacitive properties from CSDs. Our method integrates machine learning, image processing, and object detection to identify and track charge transitions across large datasets without manual labeling. We demonstrate this method using experimentally measured data from a strained-germanium single-quantum-well (planar) and a strained-germanium double-quantum-well (bilayer) QD device. Unlike for planar QD devices, CSDs in bilayer germanium heterostructure exhibit a larger set of transitions, including interlayer tunneling and distinct loading lines for the vertically stacked QDs, making them a powerful testbed for automation methods. By analyzing the properties of many CSDs, we can statistically estimate physically relevant quantities, like relative lever arms and capacitive couplings. Thus, our protocol enables rapid extraction of useful, nontrivial information about QD devices. --- ### When the Same Coefficients Reach Different Places: Asymmetric Realizability in Transplanting Tokenizers across Large Language Models Authors: Xiaoze Liu, Weichen Yu, Matt Fredrikson, Xiaoqian Wang, Jing Gao Published: 2025-12-31 Categories: cs.LG, cs.CL, cs.CR arXiv: https://arxiv.org/abs/2601.00065v3 Tokenizer transplant in cross-vocabulary model composition reconstructs donor-only embedding rows as weighted combinations over shared lexical anchors and reuses those coefficients on the base. We identify a structural geometric property of this reconstruction: the same coefficient vector reaches different sets in the donor and base anchor spans, an \emph{asymmetric realizability} gap. Across 65 donor-base pairs under OMP, with cross-operator validation on CLP, WECHSEL, and FOCUS, we construct \textit{breaker tokens}: single coefficient vectors that remain statistically inert in the donor anchor span while producing a high-salience reconstruction in the base. The same Gemma-2-2B donor checkpoint admits this construction against 13 different downstream bases drawn from five model families. The planted direction passes weight-merging with a clean reference unchanged. In a deployer case study, standard LoRA fine-tuning suppresses the breaker primarily on prompts whose distribution matches the training corpus and is not a sufficient mitigation against this attack family in our setting. The tested spectral filters miss the asymmetry. We discuss potential misuse in the open-weight composition supply chain. --- ### TeleWorld: Towards Dynamic Multimodal Synthesis with a 4D World Model Authors: Yabo Chen, Yuanzhi Liang, Jiepeng Wang, Tingxi Chen, Junfei Cheng, Zixiao Gu, Yuyang Huang, Zicheng Jiang, Wei Li, Tian Li, Weichen Li, Zuoxin Li et al. Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.00051v1 World models aim to endow AI systems with the ability to represent, generate, and interact with dynamic environments in a coherent and temporally consistent manner. While recent video generation models have demonstrated impressive visual quality, they remain limited in real-time interaction, long-horizon consistency, and persistent memory of dynamic scenes, hindering their evolution into practical world models. In this report, we present TeleWorld, a real-time multimodal 4D world modeling framework that unifies video generation, dynamic scene reconstruction, and long-term world memory within a closed-loop system. TeleWorld introduces a novel generation-reconstruction-guidance paradigm, where generated video streams are continuously reconstructed into a dynamic 4D spatio-temporal representation, which in turn guides subsequent generation to maintain spatial, temporal, and physical consistency. To support long-horizon generation with low latency, we employ an autoregressive diffusion-based video model enhanced with Macro-from-Micro Planning (MMPL)--a hierarchical planning method that reduces error accumulation from frame-level to segment-level-alongside efficient Distribution Matching Distillation (DMD), enabling real-time synthesis under practical computational budgets. Our approach achieves seamless integration of dynamic object modeling and static scene representation within a unified 4D framework, advancing world models toward practical, interactive, and computationally accessible systems. Extensive experiments demonstrate that TeleWorld achieves strong performance in both static and dynamic world understanding, long-term consistency, and real-time generation efficiency, positioning it as a practical step toward interactive, memory-enabled world models for multimodal generation and embodied intelligence. --- ### Group Cross-Correlations with Faintly Constrained Filters Authors: Benedikt Fluhr Published: 2025-12-31 Categories: math.DS, cs.LG, math.GR arXiv: https://arxiv.org/abs/2601.00045v2 Group convolutional layers with respect to some group $G$ are modeled by convolutions or cross-correlations with a filter, and they provide the fundamental building block for group convolutional neural networks. For entirely unconstrained filters and $G$ a non-abelian group, any hidden layer of such a network requires as many nodes as vertices in a fine enough discretization of $G$. In order to reduce the necessary number of nodes, certain constraints on filters were proposed in the literature. We propose weaker constraints retaining this benefit while also resolving an incompatibility previous constraints have for group actions with non-compact stabilizers. Moreover, we generalize previous results to group actions that are not necessarily transitive, and we weaken the common assumption that $G$ is unimodular. --- ### Large Empirical Case Study: Go-Explore adapted for AI Red Team Testing Authors: Manish Bhatt, Adrian Wood, Idan Habler, Ammar Al-Kahfah Published: 2025-12-31 Categories: cs.CR, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2601.00042v2 Production LLM agents with tool-using capabilities require security testing despite their safety training. We adapt Go-Explore to evaluate GPT-4o-mini across 28 experimental runs spanning six research questions. We find that random-seed variance dominates algorithmic parameters, yielding an 8x spread in outcomes; single-seed comparisons are unreliable, while multi-seed averaging materially reduces variance in our setup. Reward shaping consistently harms performance, causing exploration collapse in 94% of runs or producing 18 false positives with zero verified attacks. In our environment, simple state signatures outperform complex ones. For comprehensive security testing, ensembles provide attack-type diversity, whereas single agents optimize coverage within a given attack type. Overall, these results suggest that seed variance and targeted domain knowledge can outweigh algorithmic sophistication when testing safety-trained models. --- ### Deep Learning Approach for the Diagnosis of Pediatric Pneumonia Using Chest X-ray Imaging Authors: Fatemeh Hosseinabadi, Mohammad Mojtaba Rohani Published: 2025-12-31 Categories: eess.IV, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2601.00041v1 Pediatric pneumonia remains a leading cause of morbidity and mortality in children worldwide. Timely and accurate diagnosis is critical but often challenged by limited radiological expertise and the physiological and procedural complexity of pediatric imaging. This study investigates the performance of state-of-the-art convolutional neural network (CNN) architectures ResNetRS, RegNet, and EfficientNetV2 using transfer learning for the automated classification of pediatric chest Xray images as either pneumonia or normal.A curated subset of 1,000 chest X-ray images was extracted from a publicly available dataset originally comprising 5,856 pediatric images. All images were preprocessed and labeled for binary classification. Each model was fine-tuned using pretrained ImageNet weights and evaluated based on accuracy and sensitivity. RegNet achieved the highest classification performance with an accuracy of 92.4 and a sensitivity of 90.1, followed by ResNetRS (accuracy: 91.9, sensitivity: 89.3) and EfficientNetV2 (accuracy: 88.5, sensitivity: 88.1). --- ### SpaceTimePilot: Generative Rendering of Dynamic Scenes Across Space and Time Authors: Zhening Huang, Hyeonho Jeong, Xuelin Chen, Yulia Gryaditskaya, Tuanfeng Y. Wang, Joan Lasenby, Chun-Hao Huang Published: 2025-12-31 Categories: cs.CV, cs.AI, cs.RO arXiv: https://arxiv.org/abs/2512.25075v1 We present SpaceTimePilot, a video diffusion model that disentangles space and time for controllable generative rendering. Given a monocular video, SpaceTimePilot can independently alter the camera viewpoint and the motion sequence within the generative process, re-rendering the scene for continuous and arbitrary exploration across space and time. To achieve this, we introduce an effective animation time-embedding mechanism in the diffusion process, allowing explicit control of the output video's motion sequence with respect to that of the source video. As no datasets provide paired videos of the same dynamic scene with continuous temporal variations, we propose a simple yet effective temporal-warping training scheme that repurposes existing multi-view datasets to mimic temporal differences. This strategy effectively supervises the model to learn temporal control and achieve robust space-time disentanglement. To further enhance the precision of dual control, we introduce two additional components: an improved camera-conditioning mechanism that allows altering the camera from the first frame, and CamxTime, the first synthetic space-and-time full-coverage rendering dataset that provides fully free space-time video trajectories within a scene. Joint training on the temporal-warping scheme and the CamxTime dataset yields more precise temporal control. We evaluate SpaceTimePilot on both real-world and synthetic data, demonstrating clear space-time disentanglement and strong results compared to prior work. Project page: https://zheninghuang.github.io/Space-Time-Pilot/ Code: https://github.com/ZheningHuang/spacetimepilot --- ### GaMO: Geometry-aware Multi-view Diffusion Outpainting for Sparse-View 3D Reconstruction Authors: Yi-Chuan Huang, Hao-Jen Chien, Chin-Yang Lin, Ying-Huan Chen, Yu-Lun Liu Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25073v2 Recent 3D reconstruction methods achieve impressive results with dense multi-view imagery but struggle when only a few views are available. Various approaches, including regularization techniques, semantic priors, and geometric constraints, have been implemented to address this challenge. Recent diffusion-based approaches further improve performance by generating novel views to augment training data. Despite this progress, we identify three critical limitations in current state-of-the-art approaches: (i) inadequate coverage beyond known view peripheries, (ii) geometric inconsistencies across generated views, and (iii) computational inefficiency due to expensive pipelines. We introduce GaMO (Geometry-aware Multi-view Outpainter), a framework that reformulates sparse-view reconstruction through multi-view outpainting. Instead of generating new viewpoints, GaMO expands the field of view from existing camera poses, which inherently preserves geometric consistency while providing broader scene coverage. Our approach employs multi-view conditioning and geometry-aware denoising strategies in a zero-shot manner without training. Extensive experiments on Replica, ScanNet++, and Mip-NeRF 360 demonstrate strong reconstruction performance across sparse-view settings (3, 6, and 9 input views). Notably, our method is significantly more efficient than existing diffusion-based approaches, reducing the overall runtime to within 10 minutes. Project page: https://yichuanh.github.io/GaMO/ --- ### GaMO: Geometry-aware Multi-view Diffusion Outpainting for Sparse-View 3D Reconstruction Authors: Yi-Chuan Huang, Hao-Jen Chien, Chin-Yang Lin, Ying-Huan Chen, Yu-Lun Liu Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25073v1 Recent advances in 3D reconstruction have achieved remarkable progress in high-quality scene capture from dense multi-view imagery, yet struggle when input views are limited. Various approaches, including regularization techniques, semantic priors, and geometric constraints, have been implemented to address this challenge. Latest diffusion-based methods have demonstrated substantial improvements by generating novel views from new camera poses to augment training data, surpassing earlier regularization and prior-based techniques. Despite this progress, we identify three critical limitations in these state-of-the-art approaches: inadequate coverage beyond known view peripheries, geometric inconsistencies across generated views, and computationally expensive pipelines. We introduce GaMO (Geometry-aware Multi-view Outpainter), a framework that reformulates sparse-view reconstruction through multi-view outpainting. Instead of generating new viewpoints, GaMO expands the field of view from existing camera poses, which inherently preserves geometric consistency while providing broader scene coverage. Our approach employs multi-view conditioning and geometry-aware denoising strategies in a zero-shot manner without training. Extensive experiments on Replica and ScanNet++ demonstrate state-of-the-art reconstruction quality across 3, 6, and 9 input views, outperforming prior methods in PSNR and LPIPS, while achieving a $25\times$ speedup over SOTA diffusion-based methods with processing time under 10 minutes. Project page: https://yichuanh.github.io/GaMO/ --- ### Coordinated Humanoid Manipulation with Choice Policies Authors: Haozhi Qi, Yen-Jen Wang, Toru Lin, Brent Yi, Yi Ma, Koushil Sreenath, Jitendra Malik Published: 2025-12-31 Categories: cs.RO, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2512.25072v1 Humanoid robots hold great promise for operating in human-centric environments, yet achieving robust whole-body coordination across the head, hands, and legs remains a major challenge. We present a system that combines a modular teleoperation interface with a scalable learning framework to address this problem. Our teleoperation design decomposes humanoid control into intuitive submodules, which include hand-eye coordination, grasp primitives, arm end-effector tracking, and locomotion. This modularity allows us to collect high-quality demonstrations efficiently. Building on this, we introduce Choice Policy, an imitation learning approach that generates multiple candidate actions and learns to score them. This architecture enables both fast inference and effective modeling of multimodal behaviors. We validate our approach on two real-world tasks: dishwasher loading and whole-body loco-manipulation for whiteboard wiping. Experiments show that Choice Policy significantly outperforms diffusion policies and standard behavior cloning. Furthermore, our results indicate that hand-eye coordination is critical for success in long-horizon tasks. Our work demonstrates a practical path toward scalable data collection and learning for coordinated humanoid manipulation in unstructured environments. --- ### Edit3r: Instant 3D Scene Editing from Sparse Unposed Images Authors: Jiageng Liu, Weijie Lyu, Xueting Li, Yejie Guo, Ming-Hsuan Yang Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25071v1 We present Edit3r, a feed-forward framework that reconstructs and edits 3D scenes in a single pass from unposed, view-inconsistent, instruction-edited images. Unlike prior methods requiring per-scene optimization, Edit3r directly predicts instruction-aligned 3D edits, enabling fast and photorealistic rendering without optimization or pose estimation. A key challenge in training such a model lies in the absence of multi-view consistent edited images for supervision. We address this with (i) a SAM2-based recoloring strategy that generates reliable, cross-view-consistent supervision, and (ii) an asymmetric input strategy that pairs a recolored reference view with raw auxiliary views, encouraging the network to fuse and align disparate observations. At inference, our model effectively handles images edited by 2D methods such as InstructPix2Pix, despite not being exposed to such edits during training. For large-scale quantitative evaluation, we introduce DL3DV-Edit-Bench, a benchmark built on the DL3DV test split, featuring 20 diverse scenes, 4 edit types and 100 edits in total. Comprehensive quantitative and qualitative results show that Edit3r achieves superior semantic alignment and enhanced 3D consistency compared to recent baselines, while operating at significantly higher inference speed, making it promising for real-time 3D editing applications. --- ### Scaling Open-Ended Reasoning to Predict the Future Authors: Nikhil Chandak, Shashwat Goel, Ameya Prabhu, Moritz Hardt, Jonas Geiping Published: 2025-12-31 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2512.25070v2 High-stakes decision making involves reasoning under uncertainty about the future. In this work, we train language models to make predictions on open-ended forecasting questions. To scale up training data, we synthesize novel forecasting questions from global events reported in daily news, using a fully automated, careful curation recipe. We train the Qwen3 thinking models on our dataset, OpenForesight. To prevent leakage of future information during training and evaluation, we use an offline news corpus, both for data generation and retrieval in our forecasting system. Guided by a small validation set, we show the benefits of retrieval, and an improved reward function for reinforcement learning (RL). Once we obtain our final forecasting system, we perform held-out testing between May to August 2025. Our specialized model, OpenForecaster 8B, matches much larger proprietary models, with our training improving the accuracy, calibration, and consistency of predictions. We find calibration improvements from forecasting training generalize across popular benchmarks. We open-source all our models, code, and data to make research on language model forecasting broadly accessible. --- ### Scaling Open-Ended Reasoning to Predict the Future Authors: Nikhil Chandak, Shashwat Goel, Ameya Prabhu, Moritz Hardt, Jonas Geiping Published: 2025-12-31 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2512.25070v1 High-stakes decision making involves reasoning under uncertainty about the future. In this work, we train language models to make predictions on open-ended forecasting questions. To scale up training data, we synthesize novel forecasting questions from global events reported in daily news, using a fully automated, careful curation recipe. We train the Qwen3 thinking models on our dataset, OpenForesight. To prevent leakage of future information during training and evaluation, we use an offline news corpus, both for data generation and retrieval in our forecasting system. Guided by a small validation set, we show the benefits of retrieval, and an improved reward function for reinforcement learning (RL). Once we obtain our final forecasting system, we perform held-out testing between May to August 2025. Our specialized model, OpenForecaster 8B, matches much larger proprietary models, with our training improving the accuracy, calibration, and consistency of predictions. We find calibration improvements from forecasting training generalize across popular benchmarks. We open-source all our models, code, and data to make research on language model forecasting broadly accessible. --- ### FineTec: Fine-Grained Action Recognition Under Temporal Corruption via Skeleton Decomposition and Sequence Completion Authors: Dian Shao, Mingfei Shi, Like Liu Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25067v1 Recognizing fine-grained actions from temporally corrupted skeleton sequences remains a significant challenge, particularly in real-world scenarios where online pose estimation often yields substantial missing data. Existing methods often struggle to accurately recover temporal dynamics and fine-grained spatial structures, resulting in the loss of subtle motion cues crucial for distinguishing similar actions. To address this, we propose FineTec, a unified framework for Fine-grained action recognition under Temporal Corruption. FineTec first restores a base skeleton sequence from corrupted input using context-aware completion with diverse temporal masking. Next, a skeleton-based spatial decomposition module partitions the skeleton into five semantic regions, further divides them into dynamic and static subgroups based on motion variance, and generates two augmented skeleton sequences via targeted perturbation. These, along with the base sequence, are then processed by a physics-driven estimation module, which utilizes Lagrangian dynamics to estimate joint accelerations. Finally, both the fused skeleton position sequence and the fused acceleration sequence are jointly fed into a GCN-based action recognition head. Extensive experiments on both coarse-grained (NTU-60, NTU-120) and fine-grained (Gym99, Gym288) benchmarks show that FineTec significantly outperforms previous methods under various levels of temporal corruption. Specifically, FineTec achieves top-1 accuracies of 89.1% and 78.1% on the challenging Gym99-severe and Gym288-severe settings, respectively, demonstrating its robustness and generalizability. Code and datasets could be found at https://smartdianlab.github.io/projects-FineTec/. --- ### From Inpainting to Editing: Unlocking Robust Mask-Free Visual Dubbing via Generative Bootstrapping Authors: Xu He, Haoxian Zhang, Hejia Chen, Changyuan Zheng, Liyang Chen, Songlin Tang, Jiehui Huang, Xiaoqiang Liu, Pengfei Wan, Zhiyong Wu Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25066v2 Audio-driven visual dubbing aims to synchronize a video's lip movements with new speech but is fundamentally challenged by the lack of ideal training data: paired videos differing only in lip motion. Existing methods circumvent this via mask-based inpainting. However, masking inevitably destroys spatiotemporal context, leading to identity drift and poor robustness (e.g., to occlusions), while also inducing lip-shape leakage that degrades lip sync. To bridge this gap, we propose X-Dub, a novel two-stage generative bootstrapping framework leveraging powerful Diffusion Transformers to unlock mask-free dubbing. Our core insight is to repurpose a mask-based inpainting model exclusively as a dedicated data generator to synthesize scalable, high-fidelity pseudo-paired data, which is subsequently utilized to train and bootstrap a robust, mask-free editing model as the final video dubber. The final dubber is liberated from masking artifacts and leverages the complete video input for high-fidelity inference. We further introduce timestep-adaptive multi-phase learning to disentangle conflicting objectives (structure, lip motion, and texture) across diffusion phases, facilitating stable convergence and advanced editing quality. Additionally, we present X-DubBench, a benchmark for diverse scenarios. Extensive experiments demonstrate that our method achieves state-of-the-art performance with superior lip sync, visual quality, and robustness. --- ### From Inpainting to Editing: A Self-Bootstrapping Framework for Context-Rich Visual Dubbing Authors: Xu He, Haoxian Zhang, Hejia Chen, Changyuan Zheng, Liyang Chen, Songlin Tang, Jiehui Huang, Xiaoqiang Liu, Pengfei Wan, Zhiyong Wu Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25066v1 Audio-driven visual dubbing aims to synchronize a video's lip movements with new speech, but is fundamentally challenged by the lack of ideal training data: paired videos where only a subject's lip movements differ while all other visual conditions are identical. Existing methods circumvent this with a mask-based inpainting paradigm, where an incomplete visual conditioning forces models to simultaneously hallucinate missing content and sync lips, leading to visual artifacts, identity drift, and poor synchronization. In this work, we propose a novel self-bootstrapping framework that reframes visual dubbing from an ill-posed inpainting task into a well-conditioned video-to-video editing problem. Our approach employs a Diffusion Transformer, first as a data generator, to synthesize ideal training data: a lip-altered companion video for each real sample, forming visually aligned video pairs. A DiT-based audio-driven editor is then trained on these pairs end-to-end, leveraging the complete and aligned input video frames to focus solely on precise, audio-driven lip modifications. This complete, frame-aligned input conditioning forms a rich visual context for the editor, providing it with complete identity cues, scene interactions, and continuous spatiotemporal dynamics. Leveraging this rich context fundamentally enables our method to achieve highly accurate lip sync, faithful identity preservation, and exceptional robustness against challenging in-the-wild scenarios. We further introduce a timestep-adaptive multi-phase learning strategy as a necessary component to disentangle conflicting editing objectives across diffusion timesteps, thereby facilitating stable training and yielding enhanced lip synchronization and visual fidelity. Additionally, we propose ContextDubBench, a comprehensive benchmark dataset for robust evaluation in diverse and challenging practical application scenarios. --- ### Vulcan: Instance-specialized, Verifiable Systems Heuristics Through LLM-driven Search Authors: Rohit Dwivedula, Divyanshu Saxena, Sujay Yadalam, Eric Hayden Campbell, Daehyeok Kim, Aditya Akella Published: 2025-12-31 Categories: cs.OS, cs.AI, cs.DC arXiv: https://arxiv.org/abs/2512.25065v3 Systems resource management tasks rely primarily on hand-designed heuristics. However, growing hardware heterogeneity and workload diversity require heuristics specialized to particular deployment instances, making manual design expensive and difficult to scale. In this paper, we explore how to synthesize systems heuristics using LLMs. The main challenge is ensuring that generated heuristics execute safely, integrate correctly with the surrounding system, and still achieve strong performance. We propose Vulcan, a framework that identifies LLM-friendly interfaces that isolate core decision logic from the rest of the implementation. With Vulcan, LLM-generated code is restricted to simple stateless decision functions, while trusted runtime abstractions provide rich derived statistics for meaningful policy exploration without system-integration bugs. To ensure execution safety, LLMs synthesize heuristics in a restricted language, Anvil, that guarantees important properties by construction. We evaluate Vulcan across three well-studied domains and demonstrate up to 4.9$\times$ higher savings for spot-VM scheduling, up to 2$\times$ lower miss ratios for cache eviction, and up to 14% higher application performance for tiered-memory systems, while ensuring execution safety throughout. --- ### Vulcan: Instance-Optimal Systems Heuristics Through LLM-Driven Search Authors: Rohit Dwivedula, Divyanshu Saxena, Sujay Yadalam, Daehyeok Kim, Aditya Akella Published: 2025-12-31 Categories: cs.OS, cs.AI, cs.DC arXiv: https://arxiv.org/abs/2512.25065v1 Resource-management tasks in modern operating and distributed systems continue to rely primarily on hand-designed heuristics for tasks such as scheduling, caching, or active queue management. Designing performant heuristics is an expensive, time-consuming process that we are forced to continuously go through due to the constant flux of hardware, workloads and environments. We propose a new alternative: synthesizing instance-optimal heuristics -- specialized for the exact workloads and hardware where they will be deployed -- using code-generating large language models (LLMs). To make this synthesis tractable, Vulcan separates policy and mechanism through LLM-friendly, task-agnostic interfaces. With these interfaces, users specify the inputs and objectives of their desired policy, while Vulcan searches for performant policies via evolutionary search over LLM-generated code. This interface is expressive enough to capture a wide range of system policies, yet sufficiently constrained to allow even small, inexpensive LLMs to generate correct and executable code. We use Vulcan to synthesize performant heuristics for cache eviction and memory tiering, and find that these heuristics outperform all human-designed state-of-the-art algorithms by upto 69% and 7.9% in performance for each of these tasks respectively. --- ### Many Minds from One Model: Bayesian-Inspired Transformers for Population Diversity Authors: Diji Yang, Yi Zhang Published: 2025-12-31 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2512.25063v2 Despite their scale and success, modern transformers are usually trained as single-minded systems: optimization produces a deterministic set of parameters, representing a single functional hypothesis about the data. Motivated by the analogy to human populations, in which population-level intelligence emerges from diverse individual behaviors, we propose Population Bayesian Transformers (B-Trans), which enable sampling diverse yet coherent transformer large language model instances (hereafter referred to as a 'mind') from a single pre-trained LLM. B-Trans introduces a Bayesian-inspired posterior proxy by injecting stochasticity directly into normalization layers, avoiding the prohibitive cost of training full Bayesian neural networks. Sampling from this proxy yields a population of minds with diverse behaviors while maintaining general competence. During the generation of each response, we sample a single realization from the random distribution and hold it fixed, ensuring temporal consistency and reasoning coherence. Experiments on zero-shot generation and Reinforcement Learning with Verifiable Rewards (RLVR) demonstrate that B-Trans effectively leverages the stochastic model diversity, yielding superior response diversity while achieving better task performance compared to deterministic baselines. --- ### Many Minds from One Model: Bayesian Transformers for Population Intelligence Authors: Diji Yang, Yi Zhang Published: 2025-12-31 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2512.25063v1 Despite their scale and success, modern transformers are almost universally trained as single-minded systems: optimization produces one deterministic set of parameters, representing a single functional hypothesis about the data. Motivated by the idea that intelligence emerge from many minds, we propose Population Bayesian Transformers (B-Trans), which transform a standard Large Language Model into a Bayesian Transformer model to supports sampling diverse yet coherent model instances from a single set of pre-trained weights. B-Trans introduces a Bayesian-motivated posterior proxy by treating the bias-like offsets in normalization layers as stochastic variables with a Gaussian variational approximation, inducing a distribution over model behavior without the cost of training full Bayesian neural networks. Sampling from this proxy yields a set of model instances with diverse behaviors while maintaining general competence. To preserve coherence within each generation, we freeze the sampled noise at the sequence level, enforcing temporal consistency across tokens. B-Trans allows for population-level decision-making, where aggregating predictions across sampled individuals significantly enhances exploration. Experiments across zero-shot generation, Reinforcement Learning with Verifiable Rewards (RLVR), and RL without explicit labels demonstrate that B-Trans effectively leverage the wisdom of crowds, yielding superior semantic diversity while achieving better task performance compared to deterministic baselines. --- ### On the geometry and topology of representations: the manifolds of modular addition Authors: Gabriela Moisescu-Pareja, Gavin McCracken, Harley Wiltzer, Vincent Létourneau, Colin Daniels, Doina Precup, Jonathan Love Published: 2025-12-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.25060v1 The Clock and Pizza interpretations, associated with architectures differing in either uniform or learnable attention, were introduced to argue that different architectural designs can yield distinct circuits for modular addition. In this work, we show that this is not the case, and that both uniform attention and trainable attention architectures implement the same algorithm via topologically and geometrically equivalent representations. Our methodology goes beyond the interpretation of individual neurons and weights. Instead, we identify all of the neurons corresponding to each learned representation and then study the collective group of neurons as one entity. This method reveals that each learned representation is a manifold that we can study utilizing tools from topology. Based on this insight, we can statistically analyze the learned representations across hundreds of circuits to demonstrate the similarity between learned modular addition circuits that arise naturally from common deep learning paradigms. --- ### Reliable and Resilient Collective Communication Library for LLM Training and Serving Authors: Wei Wang, Nengneng Yu, Sixian Xiong, Zaoxing Liu Published: 2025-12-31 Categories: cs.DC, cs.LG, cs.NI arXiv: https://arxiv.org/abs/2512.25059v1 Modern ML training and inference now span tens to tens of thousands of GPUs, where network faults can waste 10--15\% of GPU hours due to slow recovery. Common network errors and link fluctuations trigger timeouts that often terminate entire jobs, forcing expensive checkpoint rollback during training and request reprocessing during inference. We present R$^2$CCL, a fault-tolerant communication library that provides lossless, low-overhead failover by exploiting multi-NIC hardware. R$^2$CCL performs rapid connection migration, bandwidth-aware load redistribution, and resilient collective algorithms to maintain progress under failures. We evaluate R$^2$CCL on two 8-GPU H100 InfiniBand servers and via large-scale ML simulators modeling hundreds of GPUs with diverse failure patterns. Experiments show that R$^2$CCL is highly robust to NIC failures, incurring less than 1\% training and less than 3\% inference overheads. R$^2$CCL outperforms baselines AdapCC and DejaVu by 12.18$\times$ and 47$\times$, respectively. --- ### Context-aware LLM-based AI Agents for Human-centered Energy Management Systems in Smart Buildings Authors: Tianzhi He, Farrokh Jazizadeh Published: 2025-12-31 Categories: cs.AI, cs.HC arXiv: https://arxiv.org/abs/2512.25055v1 This study presents a conceptual framework and a prototype assessment for Large Language Model (LLM)-based Building Energy Management System (BEMS) AI agents to facilitate context-aware energy management in smart buildings through natural language interaction. The proposed framework comprises three modules: perception (sensing), central control (brain), and action (actuation and user interaction), forming a closed feedback loop that captures, analyzes, and interprets energy data to respond intelligently to user queries and manage connected appliances. By leveraging the autonomous data analytics capabilities of LLMs, the BEMS AI agent seeks to offer context-aware insights into energy consumption, cost prediction, and device scheduling, thereby addressing limitations in existing energy management systems. The prototype's performance was evaluated using 120 user queries across four distinct real-world residential energy datasets and different evaluation metrics, including latency, functionality, capability, accuracy, and cost-effectiveness. The generalizability of the framework was demonstrated using ANOVA tests. The results revealed promising performance, measured by response accuracy in device control (86%), memory-related tasks (97%), scheduling and automation (74%), and energy analysis (77%), while more complex cost estimation tasks highlighted areas for improvement with an accuracy of 49%. This benchmarking study moves toward formalizing the assessment of LLM-based BEMS AI agents and identifying future research directions, emphasizing the trade-off between response accuracy and computational efficiency. --- ### AdaGReS:Adaptive Greedy Context Selection via Redundancy-Aware Scoring for Token-Budgeted RAG Authors: Chao Peng, Bin Wang, Zhilei Long, Jinfang Sheng Published: 2025-12-31 Categories: cs.CL, cs.AI, cs.IR arXiv: https://arxiv.org/abs/2512.25052v1 Retrieval-augmented generation (RAG) is highly sensitive to the quality of selected context, yet standard top-k retrieval often returns redundant or near-duplicate chunks that waste token budget and degrade downstream generation. We present AdaGReS, a redundancy-aware context selection framework for token-budgeted RAG that optimizes a set-level objective combining query-chunk relevance and intra-set redundancy penalties. AdaGReS performs greedy selection under a token-budget constraint using marginal gains derived from the objective, and introduces a closed-form, instance-adaptive calibration of the relevance-redundancy trade-off parameter to eliminate manual tuning and adapt to candidate-pool statistics and budget limits. We further provide a theoretical analysis showing that the proposed objective exhibits epsilon-approximate submodularity under practical embedding similarity conditions, yielding near-optimality guarantees for greedy selection. Experiments on open-domain question answering (Natural Questions) and a high-redundancy biomedical (drug) corpus demonstrate consistent improvements in redundancy control and context quality, translating to better end-to-end answer quality and robustness across settings. --- ### Generative Classifiers Avoid Shortcut Solutions Authors: Alexander C. Li, Ananya Kumar, Deepak Pathak Published: 2025-12-31 Categories: cs.LG, cs.AI, cs.CV, cs.NE arXiv: https://arxiv.org/abs/2512.25034v1 Discriminative approaches to classification often learn shortcuts that hold in-distribution but fail even under minor distribution shift. This failure mode stems from an overreliance on features that are spuriously correlated with the label. We show that generative classifiers, which use class-conditional generative models, can avoid this issue by modeling all features, both core and spurious, instead of mainly spurious ones. These generative classifiers are simple to train, avoiding the need for specialized augmentations, strong regularization, extra hyperparameters, or knowledge of the specific spurious correlations to avoid. We find that diffusion-based and autoregressive generative classifiers achieve state-of-the-art performance on five standard image and text distribution shift benchmarks and reduce the impact of spurious correlations in realistic applications, such as medical or satellite datasets. Finally, we carefully analyze a Gaussian toy setting to understand the inductive biases of generative classifiers, as well as the data properties that determine when generative classifiers outperform discriminative ones. --- ### Modeling Language as a Sequence of Thoughts Authors: Nasim Borazjanizadeh, James McClelland Published: 2025-12-31 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.25026v2 Transformer language models can generate strikingly natural text by modeling language as a sequence of tokens, but by relying primarily on surface-level co-occurrence statistics they fail to form globally consistent latent representations of entities and events, which contributes to poor relational generalization (the reversal curse), contextualization errors, and data inefficiency. Cognitive science, by contrast, shows that human comprehension converts linguistic input into compact, event-like representations that persist in memory while verbatim form is short-lived. Motivated by these findings, we introduce the Thought Gestalt (TG) model, a recurrent transformer that models language at two levels of abstraction: tokens and sentence-level "thought" states. TG generates one sentence at a time while cross-attending to a working memory of prior sentence representations. Token and sentence representations are generated using a shared stack of transformer blocks and trained with a single objective, next-token prediction loss. By retaining the computation graph of sentence representations written to working memory, gradients from future token losses flow backward through cross-attention to optimize the parameters that generate earlier sentence vectors. In scaling experiments, TG consistently improves data and parameter efficiency compared to matched GPT-2 runs and other baselines, with scaling fits indicating GPT-2 requires ~5-8% more data and ~33-42% more parameters to match TG's test loss. TG also reduces errors in relational-direction generalization on a father-son reversal curse probe. --- ### Modeling Language as a Sequence of Thoughts Authors: Nasim Borazjanizadeh, James McClelland Published: 2025-12-31 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.25026v1 Transformer language models can generate strikingly natural text by modeling language as a sequence of tokens. Yet, by relying primarily on surface-level co-occurrence statistics, they fail to form globally consistent latent representations of entities and events, lack of which contributes to brittleness in relational direction (e.g., reversal curse), contextualization errors, and data inefficiency. On the other hand, cognitive science shows that human comprehension involves converting the input linguistic stream into compact, event-like representations that persist in memory while verbatim form is short-lived. Motivated by this view, we introduce Thought Gestalt (TG) model, a recurrent Transformer that models language at two levels of abstraction - tokens and sentence-level "thought" states. TG generates the tokens of one sentence at a time while cross-attending to a memory of prior sentence representations. In TG, token and sentence representations are generated using the same set of model parameters and trained with a single objective, the next-token cross-entropy: by retaining the computation graph of sentence representations written to memory, gradients from future token losses flow backward through cross-attention to optimize the parameters generating earlier sentence vectors. In scaling experiments, TG consistently improves efficiency over matched GPT-2 runs, among other baselines, with scaling fits indicating GPT-2 requires ~5-8% more data and ~33-42% more parameters to match TG's loss. TG also reduces errors on relational direction generalization on a father-son reversal curse probe. --- ### ResponseRank: Data-Efficient Reward Modeling through Preference Strength Learning Authors: Timo Kaufmann, Yannick Metz, Daniel Keim, Eyke Hüllermeier Published: 2025-12-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.25023v1 Binary choices, as often used for reinforcement learning from human feedback (RLHF), convey only the direction of a preference. A person may choose apples over oranges and bananas over grapes, but which preference is stronger? Strength is crucial for decision-making under uncertainty and generalization of preference models, but hard to measure reliably. Metadata such as response times and inter-annotator agreement can serve as proxies for strength, but are often noisy and confounded. We propose ResponseRank to address the challenge of learning from noisy strength signals. Our method uses relative differences in proxy signals to rank responses to pairwise comparisons by their inferred preference strength. To control for systemic variation, we compare signals only locally within carefully constructed strata. This enables robust learning of utility differences consistent with strength-derived rankings while making minimal assumptions about the strength signal. Our contributions are threefold: (1) ResponseRank, a novel method that robustly learns preference strength by leveraging locally valid relative strength signals; (2) empirical evidence of improved sample efficiency and robustness across diverse tasks: synthetic preference learning (with simulated response times), language modeling (with annotator agreement), and RL control tasks (with simulated episode returns); and (3) the Pearson Distance Correlation (PDC), a novel metric that isolates cardinal utility learning from ordinal accuracy. --- ### Convergence of the generalization error for deep gradient flow methods for PDEs Authors: Chenguang Liu, Antonis Papapantoleon, Jasper Rou Published: 2025-12-31 Categories: math.NA, cs.LG, q-fin.CP, stat.ML arXiv: https://arxiv.org/abs/2512.25017v2 The aim of this article is to provide a firm mathematical foundation for the application of deep gradient flow methods (DGFMs) for the solution of (high-dimensional) partial differential equations (PDEs). We decompose the generalization error of DGFMs into an approximation and a training error. We first show that the solution of PDEs that satisfy reasonable and verifiable assumptions can be approximated by neural networks, thus the approximation error tends to zero as the number of neurons tends to infinity. Then, we derive the gradient flow that the training process follows in the ``wide network limit'' and analyze the limit of this flow as the training time tends to infinity. These results combined show that the generalization error of DGFMs tends to zero as the number of neurons and the training time tend to infinity. --- ### Convergence of the generalization error for deep gradient flow methods for PDEs Authors: Chenguang Liu, Antonis Papapantoleon, Jasper Rou Published: 2025-12-31 Categories: math.NA, cs.LG, q-fin.CP, stat.ML arXiv: https://arxiv.org/abs/2512.25017v1 The aim of this article is to provide a firm mathematical foundation for the application of deep gradient flow methods (DGFMs) for the solution of (high-dimensional) partial differential equations (PDEs). We decompose the generalization error of DGFMs into an approximation and a training error. We first show that the solution of PDEs that satisfy reasonable and verifiable assumptions can be approximated by neural networks, thus the approximation error tends to zero as the number of neurons tends to infinity. Then, we derive the gradient flow that the training process follows in the ``wide network limit'' and analyze the limit of this flow as the training time tends to infinity. These results combined show that the generalization error of DGFMs tends to zero as the number of neurons and the training time tend to infinity. --- ### MAMA-Memeia! Multi-Aspect Multi-Agent Collaboration for Depressive Symptoms Identification in Memes Authors: Siddhant Agarwal, Adya Dhuler, Polly Ruhnke, Melvin Speisman, Md Shad Akhtar, Shweta Yadav Published: 2025-12-31 Categories: cs.CL arXiv: https://arxiv.org/abs/2512.25015v1 Over the past years, memes have evolved from being exclusively a medium of humorous exchanges to one that allows users to express a range of emotions freely and easily. With the ever-growing utilization of memes in expressing depressive sentiments, we conduct a study on identifying depressive symptoms exhibited by memes shared by users of online social media platforms. We introduce RESTOREx as a vital resource for detecting depressive symptoms in memes on social media through the Large Language Model (LLM) generated and human-annotated explanations. We introduce MAMAMemeia, a collaborative multi-agent multi-aspect discussion framework grounded in the clinical psychology method of Cognitive Analytic Therapy (CAT) Competencies. MAMAMemeia improves upon the current state-of-the-art by 7.55% in macro-F1 and is established as the new benchmark compared to over 30 methods. --- ### Diffusion Language Models are Provably Optimal Parallel Samplers Authors: Haozhe Jiang, Nika Haghtalab, Lijie Chen Published: 2025-12-31 Categories: cs.LG, cs.CC arXiv: https://arxiv.org/abs/2512.25014v1 Diffusion language models (DLMs) have emerged as a promising alternative to autoregressive models for faster inference via parallel token generation. We provide a rigorous foundation for this advantage by formalizing a model of parallel sampling and showing that DLMs augmented with polynomial-length chain-of-thought (CoT) can simulate any parallel sampling algorithm using an optimal number of sequential steps. Consequently, whenever a target distribution can be generated using a small number of sequential steps, a DLM can be used to generate the distribution using the same number of optimal sequential steps. However, without the ability to modify previously revealed tokens, DLMs with CoT can still incur large intermediate footprints. We prove that enabling remasking (converting unmasked tokens to masks) or revision (converting unmasked tokens to other unmasked tokens) together with CoT further allows DLMs to simulate any parallel sampling algorithm with optimal space complexity. We further justify the advantage of revision by establishing a strict expressivity gap: DLMs with revision or remasking are strictly more expressive than those without. Our results not only provide a theoretical justification for the promise of DLMs as the most efficient parallel sampler, but also advocate for enabling revision in DLMs. --- ### FoundationSLAM: Unleashing the Power of Depth Foundation Models for End-to-End Dense Visual SLAM Authors: Yuchen Wu, Jiahe Li, Fabio Tosi, Matteo Poggi, Jin Zheng, Xiao Bai Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25008v2 We present FoundationSLAM, a learning-based monocular dense SLAM system that addresses the absence of geometric consistency in previous flow-based approaches for accurate and robust tracking and mapping. Our core idea is to bridge flow estimation with geometric reasoning by leveraging the guidance from foundation depth models. To this end, we first develop a Hybrid Flow Network that produces geometry-aware correspondences, enabling consistent depth and pose inference across diverse keyframes. To enforce global consistency, we propose a Bi-Consistent Bundle Adjustment Layer that jointly optimizes keyframe pose and depth under multi-view constraints. Furthermore, we introduce a Reliability-Aware Refinement mechanism that dynamically adapts the flow update process by distinguishing between reliable and uncertain regions, forming a closed feedback loop between matching and optimization. Extensive experiments demonstrate that FoundationSLAM achieves superior trajectory accuracy and dense reconstruction quality across multiple challenging datasets, while running in real-time at 18 FPS, demonstrating strong generalization to various scenarios and practical applicability of our method. --- ### FoundationSLAM: Unleashing the Power of Depth Foundation Models for End-to-End Dense Visual SLAM Authors: Yuchen Wu, Jiahe Li, Fabio Tosi, Matteo Poggi, Jin Zheng, Xiao Bai Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25008v1 We present FoundationSLAM, a learning-based monocular dense SLAM system that addresses the absence of geometric consistency in previous flow-based approaches for accurate and robust tracking and mapping. Our core idea is to bridge flow estimation with geometric reasoning by leveraging the guidance from foundation depth models. To this end, we first develop a Hybrid Flow Network that produces geometry-aware correspondences, enabling consistent depth and pose inference across diverse keyframes. To enforce global consistency, we propose a Bi-Consistent Bundle Adjustment Layer that jointly optimizes keyframe pose and depth under multi-view constraints. Furthermore, we introduce a Reliability-Aware Refinement mechanism that dynamically adapts the flow update process by distinguishing between reliable and uncertain regions, forming a closed feedback loop between matching and optimization. Extensive experiments demonstrate that FoundationSLAM achieves superior trajectory accuracy and dense reconstruction quality across multiple challenging datasets, while running in real-time at 18 FPS, demonstrating strong generalization to various scenarios and practical applicability of our method. --- ### Bi-C2R: Bidirectional Continual Compatible Representation for Re-indexing Free Lifelong Person Re-identification Authors: Zhenyu Cui, Jiahuan Zhou, Yuxin Peng Published: 2025-12-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.25000v1 Lifelong person Re-IDentification (L-ReID) exploits sequentially collected data to continuously train and update a ReID model, focusing on the overall performance of all data. Its main challenge is to avoid the catastrophic forgetting problem of old knowledge while training on new data. Existing L-ReID methods typically re-extract new features for all historical gallery images for inference after each update, known as "re-indexing". However, historical gallery data typically suffers from direct saving due to the data privacy issue and the high re-indexing costs for large-scale gallery images. As a result, it inevitably leads to incompatible retrieval between query features extracted by the updated model and gallery features extracted by those before the update, greatly impairing the re-identification performance. To tackle the above issue, this paper focuses on a new task called Re-index Free Lifelong person Re-IDentification (RFL-ReID), which requires performing lifelong person re-identification without re-indexing historical gallery images. Therefore, RFL-ReID is more challenging than L-ReID, requiring continuous learning and balancing new and old knowledge in diverse streaming data, and making the features output by the new and old models compatible with each other. To this end, we propose a Bidirectional Continuous Compatible Representation (Bi-C2R) framework to continuously update the gallery features extracted by the old model to perform efficient L-ReID in a compatible manner. We verify our proposed Bi-C2R method through theoretical analysis and extensive experiments on multiple benchmarks, which demonstrate that the proposed method can achieve leading performance on both the introduced RFL-ReID task and the traditional L-ReID task. --- ### Basic Inequalities for First-Order Optimization with Applications to Statistical Risk Analysis Authors: Seunghoon Paik, Kangjie Zhou, Matus Telgarsky, Ryan J. Tibshirani Published: 2025-12-31 Categories: math.ST, cs.LG, math.NA, math.OC, stat.ML arXiv: https://arxiv.org/abs/2512.24999v2 In this work, we introduce $\textit{basic inequalities}$ for first-order iterative optimization algorithms, forming a simple yet versatile framework which connects implicit and explicit regularization. Building on related comparison inequalities for optimization iterates that already exist in the literature, we extend and unify these arguments to produce a general framework, which can be used as a tool for statistical analysis. In more detail, let $f$ denote the objective function to be optimized. Given a first-order iterative algorithm initialized at $θ_0$, with current iterate $θ_T$, the basic inequality upper bounds $f(θ_T) - f(z)$ for any reference point $z$ in terms of the accumulated step sizes, and the distances between $θ_0$, $θ_T$, and $z$. These distances are measured in a geometry inherent to the optimization algorithm, which then translates into a notion of regularization being applied across the path of iterates. In addition to refining existing results on gradient descent, we provide new results for mirror descent and other first-order methods. We then show how to use these basic inequalities to derive elementary yet useful bounds on the prediction risk of early-stopped gradient descent and exponentiated gradient descent iterates in generalized linear models. We also supplement these findings with numerical experiments. --- ### Basic Inequalities for First-Order Optimization with Applications to Statistical Risk Analysis Authors: Seunghoon Paik, Kangjie Zhou, Matus Telgarsky, Ryan J. Tibshirani Published: 2025-12-31 Categories: math.ST, cs.LG, math.NA, math.OC, stat.ML arXiv: https://arxiv.org/abs/2512.24999v1 We introduce \textit{basic inequalities} for first-order iterative optimization algorithms, forming a simple and versatile framework that connects implicit and explicit regularization. While related inequalities appear in the literature, we isolate and highlight a specific form and develop it as a well-rounded tool for statistical analysis. Let $f$ denote the objective function to be optimized. Given a first-order iterative algorithm initialized at $θ_0$ with current iterate $θ_T$, the basic inequality upper bounds $f(θ_T)-f(z)$ for any reference point $z$ in terms of the accumulated step sizes and the distances between $θ_0$, $θ_T$, and $z$. The bound translates the number of iterations into an effective regularization coefficient in the loss function. We demonstrate this framework through analyses of training dynamics and prediction risk bounds. In addition to revisiting and refining known results on gradient descent, we provide new results for mirror descent with Bregman divergence projection, for generalized linear models trained by gradient descent and exponentiated gradient descent, and for randomized predictors. We illustrate and supplement these theoretical findings with experiments on generalized linear models. --- ### Classifying long legal documents using short random chunks Authors: Luis Adrián Cabrera-Diego Published: 2025-12-31 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.24997v1 Classifying legal documents is a challenge, besides their specialized vocabulary, sometimes they can be very long. This means that feeding full documents to a Transformers-based models for classification might be impossible, expensive or slow. Thus, we present a legal document classifier based on DeBERTa V3 and a LSTM, that uses as input a collection of 48 randomly-selected short chunks (max 128 tokens). Besides, we present its deployment pipeline using Temporal, a durable execution solution, which allow us to have a reliable and robust processing workflow. The best model had a weighted F-score of 0.898, while the pipeline running on CPU had a processing median time of 498 seconds per 100 files. --- ### Efficiently Estimating Data Efficiency for Language Model Fine-tuning Authors: Gyung Hyun Je, Colin Raffel Published: 2025-12-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.24991v1 While large language models (LLMs) demonstrate reasonable zero-shot capability across many downstream tasks, fine-tuning is a common practice to improve their performance. However, a task's data efficiency--i.e., the number of fine-tuning examples needed to achieve a desired level of performance--is often unknown, resulting in costly cycles of incremental annotation and retraining. Indeed, we demonstrate across a curated set of 30 specialized tasks that performant LLMs may struggle zero-shot but can attain stronger performance after fine-tuning. This motivates the need for methods to predict a task's data efficiency without requiring incremental annotation. After introducing a concrete metric that quantifies a task's data efficiency, we propose using the gradient cosine similarity of low-confidence examples to predict data efficiency based on a small number of labeled samples. We validate our approach on a diverse set of tasks with varying data efficiencies, attaining 8.6% error in overall data efficiency prediction and typically eliminating hundreds of unnecessary annotations on each task. Our experiment results and implementation code are available on GitHub. --- ### PhysTalk: Language-driven Real-time Physics in 3D Gaussian Scenes Authors: Luca Collorone, Mert Kiray, Indro Spinelli, Fabio Galasso, Benjamin Busam Published: 2025-12-31 Categories: cs.GR, cs.CV arXiv: https://arxiv.org/abs/2512.24986v1 Realistic visual simulations are omnipresent, yet their creation requires computing time, rendering, and expert animation knowledge. Open-vocabulary visual effects generation from text inputs emerges as a promising solution that can unlock immense creative potential. However, current pipelines lack both physical realism and effective language interfaces, requiring slow offline optimization. In contrast, PhysTalk takes a 3D Gaussian Splatting (3DGS) scene as input and translates arbitrary user prompts into real time, physics based, interactive 4D animations. A large language model (LLM) generates executable code that directly modifies 3DGS parameters through lightweight proxies and particle dynamics. Notably, PhysTalk is the first framework to couple 3DGS directly with a physics simulator without relying on time consuming mesh extraction. While remaining open vocabulary, this design enables interactive 3D Gaussian animation via collision aware, physics based manipulation of arbitrary, multi material objects. Finally, PhysTalk is train-free and computationally lightweight: this makes 4D animation broadly accessible and shifts these workflows from a "render and wait" paradigm toward an interactive dialogue with a modern, physics-informed pipeline. --- ### DarkQA: Benchmarking Vision-Language Models on Visual-Primitive Question Answering in Low-Light Indoor Scenes Authors: Yohan Park, Hyunwoo Ha, Wonjun Jo, Tae-Hyun Oh Published: 2025-12-31 Categories: cs.CV, cs.AI, cs.LG, cs.RO arXiv: https://arxiv.org/abs/2512.24985v4 Vision Language Models (VLMs) are increasingly adopted as central reasoning modules for embodied agents. Existing benchmarks evaluate their capabilities under ideal, well-lit conditions, yet robust 24/7 operation demands performance under a wide range of visual degradations, including low-light conditions at night or in dark environments, a core necessity that has been largely overlooked. To address this underexplored challenge, we present DarkQA, an open-source benchmark for evaluating perceptual primitives under multi-level low-light conditions in embodied scenarios. DarkQA evaluates single-view egocentric observations across controlled degradation levels, isolating low-light perceptual failures before they are entangled with complex embodied tasks. The benchmark contains 9.4K deterministically generated and verifiable question-image pairs spanning five visual-primitive families. A key design feature of DarkQA is its physical fidelity: visual degradations are modeled in linear RAW space, simulating physics-based illumination drop and sensor noise followed by an ISP-inspired rendering pipeline; we further validate the synthesis against real paired low-light camera data. We evaluate representative VLMs and Low-Light Image Enhancement (LLIE) preprocessing methods. Results show consistent VLM degradation under low illumination and sensor noise, while LLIE provides severity-dependent but unstable recovery. We demonstrate the utility of DarkQA by evaluating a wide range of state-of-the-art VLMs and Low-Light Image Enhancement (LLIE) models, and systematically reveal VLMs' limitations when operating under these challenging visual conditions. Our code and benchmark dataset will be released upon acceptance. Project website: https://darkqa-benchmark.github.io --- ### DarkEQA: Benchmarking Vision-Language Models for Embodied Question Answering in Low-Light Indoor Environments Authors: Yohan Park, Hyunwoo Ha, Wonjun Jo, Tae-Hyun Oh Published: 2025-12-31 Categories: cs.CV, cs.AI, cs.LG, cs.RO arXiv: https://arxiv.org/abs/2512.24985v1 Vision Language Models (VLMs) are increasingly adopted as central reasoning modules for embodied agents. Existing benchmarks evaluate their capabilities under ideal, well-lit conditions, yet robust 24/7 operation demands performance under a wide range of visual degradations, including low-light conditions at night or in dark environments--a core necessity that has been largely overlooked. To address this underexplored challenge, we present DarkEQA, an open-source benchmark for evaluating EQA-relevant perceptual primitives under multi-level low-light conditions. DarkEQA isolates the perception bottleneck by evaluating question answering from egocentric observations under controlled degradations, enabling attributable robustness analysis. A key design feature of DarkEQA is its physical fidelity: visual degradations are modeled in linear RAW space, simulating physics-based illumination drop and sensor noise followed by an ISP-inspired rendering pipeline. We demonstrate the utility of DarkEQA by evaluating a wide range of state-of-the-art VLMs and Low-Light Image Enhancement (LLIE) models. Our analysis systematically reveals VLMs' limitations when operating under these challenging visual conditions. Our code and benchmark dataset will be released upon acceptance. --- ### A Modal Logic for Possibilistic Reasoning with Fuzzy Formal Contexts Authors: Prosenjit Howlader, Churn-Jung Liau Published: 2025-12-31 Categories: cs.LO, cs.AI arXiv: https://arxiv.org/abs/2512.24980v1 We introduce a two-sort weighted modal logic for possibilistic reasoning with fuzzy formal contexts. The syntax of the logic includes two types of weighted modal operators corresponding to classical necessity ($\Box$) and sufficiency ($\boxminus$) modalities and its formulas are interpreted in fuzzy formal contexts based on possibility theory. We present its axiomatization that is \emph{sound} with respect to the class of all fuzzy context models. In addition, both the necessity and sufficiency fragments of the logic are also individually complete with respect to the class of all fuzzy context models. We highlight the expressive power of the logic with some illustrative examples. As a formal context is the basic construct of formal concept analysis (FCA), we generalize three main notions in FCA, i.e., formal concepts, object oriented concepts, and property oriented concepts, to their corresponding $c$-cut concepts in fuzzy formal contexts. Then, we show that our logical language can represent all three of these generalized notions. Finally, we demonstrate the possibility of extending our logic to reasoning with multi-relational fuzzy contexts, in which the Boolean combinations of different fuzzy relations are allowed. --- ### SymSeqBench: a unified framework for the generation and analysis of rule-based symbolic sequences and datasets Authors: Barna Zajzon, Younes Bouhadjar, Maxime Fabre, Felix Schmidt, Noah Ostendorf, Emre Neftci, Abigail Morrison, Renato Duarte Published: 2025-12-31 Categories: q-bio.NC, cs.AI, cs.LG, cs.NE arXiv: https://arxiv.org/abs/2512.24977v1 Sequential structure is a key feature of multiple domains of natural cognition and behavior, such as language, movement and decision-making. Likewise, it is also a central property of tasks to which we would like to apply artificial intelligence. It is therefore of great importance to develop frameworks that allow us to evaluate sequence learning and processing in a domain agnostic fashion, whilst simultaneously providing a link to formal theories of computation and computability. To address this need, we introduce two complementary software tools: SymSeq, designed to rigorously generate and analyze structured symbolic sequences, and SeqBench, a comprehensive benchmark suite of rule-based sequence processing tasks to evaluate the performance of artificial learning systems in cognitively relevant domains. In combination, SymSeqBench offers versatility in investigating sequential structure across diverse knowledge domains, including experimental psycholinguistics, cognitive psychology, behavioral analysis, neuromorphic computing and artificial intelligence. Due to its basis in Formal Language Theory (FLT), SymSeqBench provides researchers in multiple domains with a convenient and practical way to apply the concepts of FLT to conceptualize and standardize their experiments, thus advancing our understanding of cognition and behavior through shared computational frameworks and formalisms. The tool is modular, openly available and accessible to the research community. --- ### Attribution-Guided Distillation of Matryoshka Sparse Autoencoders Authors: Cristina P. Martin-Linares, Jonathan P. Ling Published: 2025-12-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.24975v1 Sparse autoencoders (SAEs) aim to disentangle model activations into monosemantic, human-interpretable features. In practice, learned features are often redundant and vary across training runs and sparsity levels, which makes interpretations difficult to transfer and reuse. We introduce Distilled Matryoshka Sparse Autoencoders (DMSAEs), a training pipeline that distills a compact core of consistently useful features and reuses it to train new SAEs. DMSAEs run an iterative distillation cycle: train a Matryoshka SAE with a shared core, use gradient X activation to measure each feature's contribution to next-token loss in the most nested reconstruction, and keep only the smallest subset that explains a fixed fraction of the attribution. Only the core encoder weight vectors are transferred across cycles; the core decoder and all non-core latents are reinitialized each time. On Gemma-2-2B layer 12 residual stream activations, seven cycles of distillation (500M tokens, 65k width) yielded a distilled core of 197 features that were repeatedly selected. Training using this distilled core improves several SAEBench metrics and demonstrates that consistent sets of latent features can be transferred across sparsity levels --- ### Evaluating the Impact of Compression Techniques on the Robustness of CNNs under Natural Corruptions Authors: Itallo Patrick Castro Alves Da Silva, Emanuel Adler Medeiros Pereira, Erick de Andrade Barboza, Baldoino Fonseca dos Santos Neto, Marcio de Medeiros Ribeiro Published: 2025-12-31 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.24971v1 Compressed deep learning models are crucial for deploying computer vision systems on resource-constrained devices. However, model compression may affect robustness, especially under natural corruption. Therefore, it is important to consider robustness evaluation while validating computer vision systems. This paper presents a comprehensive evaluation of compression techniques - quantization, pruning, and weight clustering applied individually and in combination to convolutional neural networks (ResNet-50, VGG-19, and MobileNetV2). Using the CIFAR-10-C and CIFAR 100-C datasets, we analyze the trade-offs between robustness, accuracy, and compression ratio. Our results show that certain compression strategies not only preserve but can also improve robustness, particularly on networks with more complex architectures. Utilizing multiobjective assessment, we determine the best configurations, showing that customized technique combinations produce beneficial multi-objective results. This study provides insights into selecting compression methods for robust and efficient deployment of models in corrupted real-world environments. ---