# Research Papers — 2025-12 · bundle page 43 of 175 > Full-text markdown of papers #2101–2150 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/43 [← Previous bundle](https://aipod.dev/papers/archive/2025-12/page/42.md) [Next bundle →](https://aipod.dev/papers/archive/2025-12/page/44.md) ### Helios: A Foundational Language Model for Smart Energy Knowledge Reasoning and Application Authors: Haoyu Jiang, Fanjie Zeng, Boan Qu, Xiaojie Lin, Wei Zhong Published: 2025-12-22 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.19299v2 In the global drive toward carbon neutrality, deeply coordinated smart energy systems underpin industrial transformation. However, the interdisciplinary, fragmented, and fast-evolving expertise in this domain prevents general-purpose LLMs, which lack domain knowledge and physical-constraint awareness, from delivering precise engineering-aligned inference and generation. To address these challenges, we introduce Helios, a large language model tailored to the smart energy domain, together with a comprehensive suite of resources to advance LLM research in this field. Specifically, we develop Enersys, a multi-agent collaborative framework for end-to-end dataset construction, through which we produce: (1) a smart energy knowledge base, EnerBase, to enrich the model's foundational expertise; (2) an instruction fine-tuning dataset, EnerInstruct, to strengthen performance on domain-specific downstream tasks; and (3) an RLHF dataset, EnerReinforce, to align the model with human preferences and industry standards. Leveraging these resources, Helios undergoes large-scale pretraining, SFT, and RLHF. We also release EnerBench, a benchmark for evaluating LLMs in smart energy scenarios, and demonstrate that our approach significantly enhances domain knowledge mastery, task execution accuracy, and alignment with human preferences. --- ### Causal-Guided Detoxify Backdoor Attack of Open-Weight LoRA Models Authors: Linzhi Chen, Yang Sun, Hongru Wei, Yuqi Chen Published: 2025-12-22 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2512.19297v1 Low-Rank Adaptation (LoRA) has emerged as an efficient method for fine-tuning large language models (LLMs) and is widely adopted within the open-source community. However, the decentralized dissemination of LoRA adapters through platforms such as Hugging Face introduces novel security vulnerabilities: malicious adapters can be easily distributed and evade conventional oversight mechanisms. Despite these risks, backdoor attacks targeting LoRA-based fine-tuning remain relatively underexplored. Existing backdoor attack strategies are ill-suited to this setting, as they often rely on inaccessible training data, fail to account for the structural properties unique to LoRA, or suffer from high false trigger rates (FTR), thereby compromising their stealth. To address these challenges, we propose Causal-Guided Detoxify Backdoor Attack (CBA), a novel backdoor attack framework specifically designed for open-weight LoRA models. CBA operates without access to original training data and achieves high stealth through two key innovations: (1) a coverage-guided data generation pipeline that synthesizes task-aligned inputs via behavioral exploration, and (2) a causal-guided detoxification strategy that merges poisoned and clean adapters by preserving task-critical neurons. Unlike prior approaches, CBA enables post-training control over attack intensity through causal influence-based weight allocation, eliminating the need for repeated retraining. Evaluated across six LoRA models, CBA achieves high attack success rates while reducing FTR by 50-70\% compared to baseline methods. Furthermore, it demonstrates enhanced resistance to state-of-the-art backdoor defenses, highlighting its stealth and robustness. --- ### Vibe Reasoning: Eliciting Frontier AI Mathematical Capabilities -- A Case Study on IMO 2025 Problem 6 Authors: Jiaao Wu, Xian Zhang, Fan Yang, Yinpeng Dong Published: 2025-12-22 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.19287v1 We introduce Vibe Reasoning, a human-AI collaborative paradigm for solving complex mathematical problems. Our key insight is that frontier AI models already possess the knowledge required to solve challenging problems -- they simply do not know how, what, or when to apply it. Vibe Reasoning transforms AI's latent potential into manifested capability through generic meta-prompts, agentic grounding, and model orchestration. We demonstrate this paradigm through IMO 2025 Problem 6, a combinatorial optimization problem where autonomous AI systems publicly reported failures. Our solution combined GPT-5's exploratory capabilities with Gemini 3 Pro's proof strengths, leveraging agentic workflows with Python code execution and file-based memory, to derive both the correct answer (2112) and a rigorous mathematical proof. Through iterative refinement across multiple attempts, we discovered the necessity of agentic grounding and model orchestration, while human prompts evolved from problem-specific hints to generic, transferable meta-prompts. We analyze why capable AI fails autonomously, how each component addresses specific failure modes, and extract principles for effective vibe reasoning. Our findings suggest that lightweight human guidance can unlock frontier models' mathematical reasoning potential. This is ongoing work; we are developing automated frameworks and conducting broader evaluations to further validate Vibe Reasoning's generality and effectiveness. --- ### GShield: Mitigating Poisoning Attacks in Federated Learning Authors: Sameera K. M., Serena Nicolazzo, Antonino Nocera, Vinod P., Rafidha Rehiman K. A Published: 2025-12-22 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2512.19286v2 Federated Learning (FL) has recently emerged as a revolutionary approach to collaborative training Machine Learning models. In particular, it enables decentralized model training while preserving data privacy, but its distributed nature makes it highly vulnerable to a severe attack known as Data Poisoning. In such scenarios, malicious clients inject manipulated data into the training process, thereby degrading global model performance or causing targeted misclassification. In this paper, we present a novel defense mechanism called GShield, designed to detect and mitigate malicious and low-quality updates, especially under non-independent and identically distributed (non-IID) data scenarios. GShield operates by learning the distribution of benign gradients through clustering and Gaussian modeling during an initial round, enabling it to establish a reliable baseline of trusted client behavior. With this benign profile, GShield selectively aggregates only those updates that align with the expected gradient patterns, effectively isolating adversarial clients and preserving the integrity of the global model. An extensive experimental campaign demonstrates that our proposed defense significantly improves model robustness compared to the state-of-the-art methods while maintaining a high accuracy of performance across both tabular and image datasets. Furthermore, GShield improves the accuracy of the targeted class by 43\% to 65\% after detecting malicious and low-quality clients. --- ### OmniEgoCap: Camera-Agnostic Sequence-Level Egocentric Motion Reconstruction Authors: Kyungwon Cho, Hanbyul Joo Published: 2025-12-22 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.19283v2 The proliferation of commercial egocentric devices offers a unique lens into human behavior, yet reconstructing full-body 3D motion remains difficult due to frequent self-occlusion and the 'out-of-sight' nature of the wearer's limbs. While head and hand trajectories provide sparse anchor points, current methods often overfit to specific hardware optics or rely on expensive, post-hoc optimizations that compromise motion naturalness. In this paper, we present OmniEgoCap, a unified diffusion framework that scales egocentric reconstruction to diverse capture setups. By shifting from short-term windowed estimation to sequence-level inference, our method captures a global perspective and recovers invariant physical attributes, such as height and body proportions, that provide critical constraints for disambiguating head-only cues. To ensure hardware-agnostic generalization, we introduce a geometry-aware visibility augmentation strategy that treats intermittent hand appearances as principled geometric constraints rather than missing data. Our architecture jointly predicts temporally coherent motion and consistent body shape, establishing a new state-of-the-art on public benchmarks and demonstrating robust performance across diverse, in-the-wild environments. --- ### Digital Twin-Driven Zero-Shot Fault Diagnosis of Axial Piston Pumps Using Fluid-Borne Noise Signals Authors: Chang Dong, Jianfeng Tao, Chengliang Liu Published: 2025-12-22 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2512.19280v1 Axial piston pumps are crucial components in fluid power systems, where reliable fault diagnosis is essential for ensuring operational safety and efficiency. Traditional data-driven methods require extensive labeled fault data, which is often impractical to obtain, while model-based approaches suffer from parameter uncertainties. This paper proposes a digital twin (DT)-driven zero-shot fault diagnosis framework utilizing fluid-borne noise (FBN) signals. The framework calibrates a high-fidelity DT model using only healthy-state data, generates synthetic fault signals for training deep learning classifiers, and employs a physics-informed neural network (PINN) as a virtual sensor for flow ripple estimation. Gradient-weighted class activation mapping (Grad-CAM) is integrated to visualize the decision-making process of neural networks, revealing that large kernels matching the subsequence length in time-domain inputs and small kernels in time-frequency domain inputs enable higher diagnostic accuracy by focusing on physically meaningful features. Experimental validations demonstrate that training on signals from the calibrated DT model yields diagnostic accuracies exceeding 95\% on real-world benchmarks, while uncalibrated models result in significantly lower performance, highlighting the framework's effectiveness in data-scarce scenarios. --- ### Is Visual Realism Enough? Evaluating Gait Biometric Fidelity in Generative AI Human Animation Authors: Ivan DeAndres-Tame, Chengwei Ye, Ruben Tolosana, Ruben Vera-Rodriguez, Shiqi Yu Published: 2025-12-22 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.19275v1 Generative AI (GenAI) models have revolutionized animation, enabling the synthesis of humans and motion patterns with remarkable visual fidelity. However, generating truly realistic human animation remains a formidable challenge, where even minor inconsistencies can make a subject appear unnatural. This limitation is particularly critical when AI-generated videos are evaluated for behavioral biometrics, where subtle motion cues that define identity are easily lost or distorted. The present study investigates whether state-of-the-art GenAI human animation models can preserve the subtle spatio-temporal details needed for person identification through gait biometrics. Specifically, we evaluate four different GenAI models across two primary evaluation tasks to assess their ability to i) restore gait patterns from reference videos under varying conditions of complexity, and ii) transfer these gait patterns to different visual identities. Our results show that while visual quality is mostly high, biometric fidelity remains low in tasks focusing on identification, suggesting that current GenAI models struggle to disentangle identity from motion. Furthermore, through an identity transfer task, we expose a fundamental flaw in appearance-based gait recognition: when texture is disentangled from motion, identification collapses, proving current GenAI models rely on visual attributes rather than temporal dynamics. --- ### 3SGen: Unified Subject, Style, and Structure-Driven Image Generation with Adaptive Task-specific Memory Authors: Xinyang Song, Libin Wang, Weining Wang, Zhiwei Li, Jianxin Sun, Dandan Zheng, Jingdong Chen, Qi Li, Zhenan Sun Published: 2025-12-22 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.19271v1 Recent image generation approaches often address subject, style, and structure-driven conditioning in isolation, leading to feature entanglement and limited task transferability. In this paper, we introduce 3SGen, a task-aware unified framework that performs all three conditioning modes within a single model. 3SGen employs an MLLM equipped with learnable semantic queries to align text-image semantics, complemented by a VAE branch that preserves fine-grained visual details. At its core, an Adaptive Task-specific Memory (ATM) module dynamically disentangles, stores, and retrieves condition-specific priors, such as identity for subjects, textures for styles, and spatial layouts for structures, via a lightweight gating mechanism along with several scalable memory items. This design mitigates inter-task interference and naturally scales to compositional inputs. In addition, we propose 3SGen-Bench, a unified image-driven generation benchmark with standardized metrics for evaluating cross-task fidelity and controllability. Extensive experiments on our proposed 3SGen-Bench and other public benchmarks demonstrate our superior performance across diverse image-driven generation tasks. --- ### Translating Flow to Policy via Hindsight Online Imitation Authors: Yitian Zheng, Zhangchen Ye, Weijun Dong, Shengjie Wang, Yuyang Liu, Chongjie Zhang, Chuan Wen, Yang Gao Published: 2025-12-22 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2512.19269v2 Recent advances in hierarchical robot systems leverage a high-level planner to propose task plans and a low-level policy to generate robot actions. This design allows training the planner on action-free or even non-robot data sources (e.g., videos), providing transferable high-level guidance. Nevertheless, grounding these high-level plans into executable actions remains challenging, especially with the limited availability of high-quality robot data. To this end, we propose to improve the low-level policy through online interactions. Specifically, our approach collects online rollouts, retrospectively annotates the corresponding high-level goals from achieved outcomes, and aggregates these hindsight-relabeled experiences to update a goal-conditioned imitation policy. Our method, Hindsight Flow-conditioned Online Imitation (HinFlow), instantiates this idea with 2D point flows as the high-level planner. Across diverse manipulation tasks in both simulation and physical world, our method achieves more than $2\times$ performance improvement over the base policy, significantly outperforming the existing methods. Moreover, our framework enables policy acquisition from planners trained on cross-embodiment video data, demonstrating its potential for scalable and transferable robot learning. --- ### Machine Unlearning in the Era of Quantum Machine Learning: An Empirical Study Authors: Carla Crivoi, Radu Tudor Ionescu Published: 2025-12-22 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2512.19253v4 We present the first empirical study of machine unlearning (MU) in hybrid quantum-classical neural networks. While MU has been extensively explored in classical deep learning, its behavior within variational quantum circuits (VQCs) and quantum-augmented architectures remains largely unexplored. First, we adapt a broad suite of unlearning methods to quantum settings, including gradient-based, distillation-based, regularization-based and certified techniques. Second, we introduce two new unlearning strategies tailored to hybrid models. Experiments across Iris, MNIST, and Fashion-MNIST, under both subset removal and full-class deletion, reveal that quantum models can support effective unlearning, but outcomes depend strongly on circuit depth, entanglement structure, and task complexity. Shallow VQCs display high intrinsic stability with minimal memorization, whereas deeper hybrid models exhibit stronger trade-offs between utility, forgetting strength, and alignment with retrain oracle. We find that certain methods, e.g. EU-k, LCA, and Certified Unlearning, consistently provide the best balance across metrics. These findings establish baseline empirical insights into quantum machine unlearning and highlight the need for quantum-aware algorithms and theoretical guarantees, as quantum machine learning systems continue to expand in scale and capability. We publicly release our code at: https://github.com/CrivoiCarla/HQML. --- ### Small Language Models as Compiler Experts: Auto-Parallelization for Heterogeneous Systems Authors: Prathamesh Devadiga Published: 2025-12-22 Categories: cs.LG, cs.PL arXiv: https://arxiv.org/abs/2512.19250v1 Traditional auto-parallelizing compilers, reliant on rigid heuristics, struggle with the complexity of modern heterogeneous systems. This paper presents a comprehensive evaluation of small (approximately 1B parameter) language-model-driven compiler auto-parallelization. We evaluate three models: gemma3, llama3.2, and qwen2.5, using six reasoning strategies across 11 real-world kernels drawn from scientific computing, graph algorithms, and machine learning. Our system is benchmarked against strong compiler baselines, including LLVM Polly, TVM, and Triton. Across 376 total evaluations, the proposed approach achieves an average speedup of 6.81x and a peak performance of 43.25x on convolution operations. We analyze scalability, verify correctness using multiple sanitizers, and confirm robustness across diverse compilers and hardware platforms. Our results demonstrate that small, efficient language models can serve as powerful reasoning engines for complex compiler optimization tasks. --- ### Auto-Prompting with Retrieval Guidance for Frame Detection in Logistics Authors: Do Minh Duc, Quan Xuan Truong, Nguyen Tat Dat, Nguyen Van Vinh Published: 2025-12-22 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.19247v1 Prompt engineering plays a critical role in adapting large language models (LLMs) to complex reasoning and labeling tasks without the need for extensive fine-tuning. In this paper, we propose a novel prompt optimization pipeline for frame detection in logistics texts, combining retrieval-augmented generation (RAG), few-shot prompting, chain-of-thought (CoT) reasoning, and automatic CoT synthesis (Auto-CoT) to generate highly effective task-specific prompts. Central to our approach is an LLM-based prompt optimizer agent that iteratively refines the prompts using retrieved examples, performance feedback, and internal self-evaluation. Our framework is evaluated on a real-world logistics text annotation task, where reasoning accuracy and labeling efficiency are critical. Experimental results show that the optimized prompts - particularly those enhanced via Auto-CoT and RAG - improve real-world inference accuracy by up to 15% compared to baseline zero-shot or static prompts. The system demonstrates consistent improvements across multiple LLMs, including GPT-4o, Qwen 2.5 (72B), and LLaMA 3.1 (70B), validating its generalizability and practical value. These findings suggest that structured prompt optimization is a viable alternative to full fine-tuning, offering scalable solutions for deploying LLMs in domain-specific NLP applications such as logistics. --- ### From Black-Box Tuning to Guided Optimization via Hyperparameters Interaction Analysis Authors: Moncef Garouani, Ayah Barhrhouj Published: 2025-12-22 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.19246v1 Hyperparameters tuning is a fundamental, yet computationally expensive, step in optimizing machine learning models. Beyond optimization, understanding the relative importance and interaction of hyperparameters is critical to efficient model development. In this paper, we introduce MetaSHAP, a scalable semi-automated eXplainable AI (XAI) method, that uses meta-learning and Shapley values analysis to provide actionable and dataset-aware tuning insights. MetaSHAP operates over a vast benchmark of over 09 millions evaluated machine learning pipelines, allowing it to produce interpretable importance scores and actionable tuning insights that reveal how much each hyperparameter matters, how it interacts with others and in which value ranges its influence is concentrated. For a given algorithm and dataset, MetaSHAP learns a surrogate performance model from historical configurations, computes hyperparameters interactions using SHAP-based analysis, and derives interpretable tuning ranges from the most influential hyperparameters. This allows practitioners not only to prioritize which hyperparameters to tune, but also to understand their directionality and interactions. We empirically validate MetaSHAP on a diverse benchmark of 164 classification datasets and 14 classifiers, demonstrating that it produces reliable importance rankings and competitive performance when used to guide Bayesian optimization. --- ### VisionDirector: Vision-Language Guided Closed-Loop Refinement for Generative Image Synthesis Authors: Meng Chu, Senqiao Yang, Haoxuan Che, Suiyun Zhang, Xichen Zhang, Shaozuo Yu, Haokun Gui, Zhefan Rao, Dandan Tu, Rui Liu, Jiaya Jia Published: 2025-12-22 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.19243v3 Generative models can now produce photorealistic imagery, yet they still struggle with the long, multi-goal prompts that professional designers issue. To expose this gap and better evaluate models' performance in real-world settings, we introduce Long Goal Bench (LGBench), a 2,000-task suite (1,000 T2I and 1,000 I2I) whose average instruction contains 18 to 22 tightly coupled goals spanning global layout, local object placement, typography, and logo fidelity. We find that even state-of-the-art models satisfy fewer than 72 percent of the goals and routinely miss localized edits, confirming the brittleness of current pipelines. To address this, we present VisionDirector, a training-free vision-language supervisor that (i) extracts structured goals from long instructions, (ii) dynamically decides between one-shot generation and staged edits, (iii) runs micro-grid sampling with semantic verification and rollback after every edit, and (iv) logs goal-level rewards. We further fine-tune the planner with Group Relative Policy Optimization, yielding shorter edit trajectories (3.1 versus 4.2 steps) and stronger alignment. VisionDirector achieves new state of the art on GenEval (plus 7 percent overall) and ImgEdit (plus 0.07 absolute) while producing consistent qualitative improvements on typography, multi-object scenes, and pose editing. --- ### ChemATP: A Training-Free Chemical Reasoning Framework for Large Language Models Authors: Mingxu Zhang, Dazhong Shen, Qi Zhang, Ying Sun Published: 2025-12-22 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.19240v1 Large Language Models (LLMs) exhibit strong general reasoning but struggle in molecular science due to the lack of explicit chemical priors in standard string representations. Current solutions face a fundamental dilemma. Training-based methods inject priors into parameters, but this static coupling hinders rapid knowledge updates and often compromises the model's general reasoning capabilities. Conversely, existing training-free methods avoid these issues but rely on surface-level prompting, failing to provide the fine-grained atom-level priors essential for precise chemical reasoning. To address this issue, we introduce ChemATP, a framework that decouples chemical knowledge from the reasoning engine. By constructing the first atom-level textual knowledge base, ChemATP enables frozen LLMs to explicitly retrieve and reason over this information dynamically. This architecture ensures interpretability and adaptability while preserving the LLM's intrinsic general intelligence. Experiments show that ChemATP significantly outperforms training-free baselines and rivals state-of-the-art training-based models, demonstrating that explicit prior injection is a competitive alternative to implicit parameter updates. --- ### Identifying Features Associated with Bias Against 93 Stigmatized Groups in Language Models and Guardrail Model Safety Mitigation Authors: Anna-Maria Gueorguieva, Aylin Caliskan Published: 2025-12-22 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2512.19238v1 Large language models (LLMs) have been shown to exhibit social bias, however, bias towards non-protected stigmatized identities remain understudied. Furthermore, what social features of stigmas are associated with bias in LLM outputs is unknown. From psychology literature, it has been shown that stigmas contain six shared social features: aesthetics, concealability, course, disruptiveness, origin, and peril. In this study, we investigate if human and LLM ratings of the features of stigmas, along with prompt style and type of stigma, have effect on bias towards stigmatized groups in LLM outputs. We measure bias against 93 stigmatized groups across three widely used LLMs (Granite 3.0-8B, Llama-3.1-8B, Mistral-7B) using SocialStigmaQA, a benchmark that includes 37 social scenarios about stigmatized identities; for example deciding wether to recommend them for an internship. We find that stigmas rated by humans to be highly perilous (e.g., being a gang member or having HIV) have the most biased outputs from SocialStigmaQA prompts (60% of outputs from all models) while sociodemographic stigmas (e.g. Asian-American or old age) have the least amount of biased outputs (11%). We test if the amount of biased outputs could be decreased by using guardrail models, models meant to identify harmful input, using each LLM's respective guardrail model (Granite Guardian 3.0, Llama Guard 3.0, Mistral Moderation API). We find that bias decreases significantly by 10.4%, 1.4%, and 7.8%, respectively. However, we show that features with significant effect on bias remain unchanged post-mitigation and that guardrail models often fail to recognize the intent of bias in prompts. This work has implications for using LLMs in scenarios involving stigmatized groups and we suggest future work towards improving guardrail models for bias mitigation. --- ### DeliveryBench: Can Agents Earn Profit in Real World? Authors: Lingjun Mao, Jiawei Ren, Kun Zhou, Jixuan Chen, Ziqiao Ma, Lianhui Qin Published: 2025-12-22 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.19234v1 LLMs and VLMs are increasingly deployed as embodied agents, yet existing benchmarks largely revolve around simple short-term tasks and struggle to capture rich realistic constraints that shape real-world decision making. To close this gap, we propose DeliveryBench, a city-scale embodied benchmark grounded in the real-world profession of food delivery. Food couriers naturally operate under long-horizon objectives (maximizing net profit over hours) while managing diverse constraints, e.g., delivery deadline, transportation expense, vehicle battery, and necessary interactions with other couriers and customers. DeliveryBench instantiates this setting in procedurally generated 3D cities with diverse road networks, buildings, functional locations, transportation modes, and realistic resource dynamics, enabling systematic evaluation of constraint-aware, long-horizon planning. We benchmark a range of VLM-based agents across nine cities and compare them with human players. Our results reveal a substantial performance gap to humans, and find that these agents are short-sighted and frequently break basic commonsense constraints. Additionally, we observe distinct personalities across models (e.g., adventurous GPT-5 vs. conservative Claude), highlighting both the brittleness and the diversity of current VLM-based embodied agents in realistic, constraint-dense environments. Our code, data, and benchmark are available at https://deliverybench.github.io. --- ### Regression generation adversarial network based on dual data evaluation strategy for industrial application Authors: Zesen Wang, Yonggang Li, Lijuan Lan Published: 2025-12-22 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.19232v1 Soft sensing infers hard-to-measure data through a large number of easily obtainable variables. However, in complex industrial scenarios, the issue of insufficient data volume persists, which diminishes the reliability of soft sensing. Generative Adversarial Networks (GAN) are one of the effective solutions for addressing insufficient samples. Nevertheless, traditional GAN fail to account for the mapping relationship between labels and features, which limits further performance improvement. Although some studies have proposed solutions, none have considered both performance and efficiency simultaneously. To address these problems, this paper proposes the multi-task learning-based regression GAN framework that integrates regression information into both the discriminator and generator, and implements a shallow sharing mechanism between the discriminator and regressor. This approach significantly enhances the quality of generated samples while improving the algorithm's operational efficiency. Moreover, considering the importance of training samples and generated samples, a dual data evaluation strategy is designed to make GAN generate more diverse samples, thereby increasing the generalization of subsequent modeling. The superiority of method is validated through four classic industrial soft sensing cases: wastewater treatment plants, surface water, $CO_2$ absorption towers, and industrial gas turbines. --- ### Generation of Programmatic Rules for Document Forgery Detection Using Large Language Models Authors: Valentin Schmidberger, Manuel Eberhardinger, Setareh Maghsudi, Johannes Maucher Published: 2025-12-22 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.19228v1 Document forgery poses a growing threat to legal, economic, and governmental processes, requiring increasingly sophisticated verification mechanisms. One approach involves the use of plausibility checks, rule-based procedures that assess the correctness and internal consistency of data, to detect anomalies or signs of manipulation. Although these verification procedures are essential for ensuring data integrity, existing plausibility checks are manually implemented by software engineers, which is time-consuming. Recent advances in code generation with large language models (LLMs) offer new potential for automating and scaling the generation of these checks. However, adapting LLMs to the specific requirements of an unknown domain remains a significant challenge. This work investigates the extent to which LLMs, adapted on domain-specific code and data through different fine-tuning strategies, can generate rule-based plausibility checks for forgery detection on constrained hardware resources. We fine-tune open-source LLMs, Llama 3.1 8B and OpenCoder 8B, on structured datasets derived from real-world application scenarios and evaluate the generated plausibility checks on previously unseen forgery patterns. The results demonstrate that the models are capable of generating executable and effective verification procedures. This also highlights the potential of LLMs as scalable tools to support human decision-making in security-sensitive contexts where comprehensibility is required. --- ### Selective Phase-Aware Training of nnU-Net for Robust Breast Cancer Segmentation in Multi-Center DCE-MRI Authors: Beyza Zayim, Aissiou Ikram, Boukhiar Naima Published: 2025-12-22 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2512.19225v1 Breast cancer remains the most common cancer among women and is a leading cause of female mortality. Dynamic contrast-enhanced MRI (DCE-MRI) is a powerful imaging tool for evaluating breast tumors, yet the field lacks a standardized benchmark for analyzing treatment responses and guiding personalized care. We participated in the MAMA-MIA Challenge's Primary Tumor Segmentation task and this work presents a proposed selective, phase-aware training framework for the nnU-Net architecture, emphasizing quality-focused data selection to strengthen model robustness and generalization. We employed the No New Net (nnU-Net) framework with a selective training strategy that systematically analyzed the impact of image quality and center-specific variability on segmentation performance. Controlled experiments on the DUKE, NACT, ISPY1, and ISPY2 datasets revealed that including ISPY scans with motion artifacts and reduced contrast impaired segmentation performance, even with advanced preprocessing, such as contrast-limited adaptive histogram equalization (CLAHE). In contrast, training on DUKE and NACT data, which exhibited clearer contrast and fewer motion artifacts despite varying resolutions, with early phase images (0000-0002) provided more stable training conditions. Our results demonstrate the importance of phase-sensitive and quality-aware training strategies in achieving reliable segmentation performance in heterogeneous clinical datasets, highlighting the limitations of the expansion of naive datasets and motivating the need for future automation of quality-based data selection strategies. --- ### Phase-space entropy at acquisition reflects downstream learnability Authors: Xiu-Cheng Wang, Jun-Jie Zhanga, Nan Cheng, Long-Gang Pang, Taijiao Du, Deyu Meng Published: 2025-12-22 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.19223v2 Modern learning systems work with data that vary widely across domains, but they all ultimately depend on how much structure is already present in the measurements before any model is trained. This raises a basic question: is there a general, modality-agnostic way to quantify how acquisition itself preserves or destroys the information that downstream learners could use? Here we propose an acquisition-level scalar $ΔS_{\mathcal B}$ based on instrument-resolved phase space. Unlike pixelwise distortion or purely spectral errors that often saturate under aggressive undersampling, $ΔS_{\mathcal B}$ directly quantifies how acquisition mixes or removes joint space--frequency structure at the instrument scale. We show theoretically that \(ΔS_{\mathcal B}\) correctly identifies the phase-space coherence of periodic sampling as the physical source of aliasing, recovering classical sampling-theorem consequences. Empirically, across masked image classification, accelerated MRI, and massive MIMO (including over-the-air measurements), $|ΔS_{\mathcal B}|$ consistently ranks sampling geometries and predicts downstream reconstruction/recognition difficulty \emph{without training}. In particular, minimizing $|ΔS_{\mathcal B}|$ enables zero-training selection of variable-density MRI mask parameters that matches designs tuned by conventional pre-reconstruction criteria. These results suggest that phase-space entropy at acquisition reflects downstream learnability, enabling pre-training selection of candidate sampling policies and as a shared notion of information preservation across modalities. --- ### From Pixels to Predicates Structuring urban perception with scene graphs Authors: Yunlong Liu, Shuyang Li, Pengyuan Liu, Yu Zhang, Rudi Stouffs Published: 2025-12-22 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.19221v1 Perception research is increasingly modelled using streetscapes, yet many approaches still rely on pixel features or object co-occurrence statistics, overlooking the explicit relations that shape human perception. This study proposes a three stage pipeline that transforms street view imagery (SVI) into structured representations for predicting six perceptual indicators. In the first stage, each image is parsed using an open-set Panoptic Scene Graph model (OpenPSG) to extract object predicate object triplets. In the second stage, compact scene-level embeddings are learned through a heterogeneous graph autoencoder (GraphMAE). In the third stage, a neural network predicts perception scores from these embeddings. We evaluate the proposed approach against image-only baselines in terms of accuracy, precision, and cross-city generalization. Results indicate that (i) our approach improves perception prediction accuracy by an average of 26% over baseline models, and (ii) maintains strong generalization performance in cross-city prediction tasks. Additionally, the structured representation clarifies which relational patterns contribute to lower perception scores in urban scenes, such as graffiti on wall and car parked on sidewalk. Overall, this study demonstrates that graph-based structure provides expressive, generalizable, and interpretable signals for modelling urban perception, advancing human-centric and context-aware urban analytics. --- ### Selective LoRA for Visual Tokens and Attention Heads Authors: Tiange Luo, Lajanugen Logeswaran, Jaekyeom Kim, Justin Johnson, Honglak Lee Published: 2025-12-22 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.19219v2 Low-rank adaptation (LoRA) is widely used for parameter-efficient fine-tuning, but its standard all-token, all-head design ignores the heterogeneous structure of vision language model (VLM) inputs. We introduce \emph{Image-LoRA}, a vision-oriented PEFT recipe that views LoRA as a token-level residual update and applies this update only to visual tokens. Image-LoRA further restricts adaptation to the value path of a compact subset of attention heads, selected using a one-pass influence estimate from a rank-1 visual-token-only probe. This token-, head-, and value-selective design reduces trainable parameters and adapter-only training FLOPs while leaving the pure-text forward pass of the frozen backbone unchanged when no visual tokens are present. Across visual localization benchmarks with controlled text:image token ratios, Image-LoRA matches or closely approaches standard LoRA, while showing especially favorable trade-offs in image-token-heavy regimes. We further validate its generality on TextVQA and VideoQA, verify pure-text preservation on GSM8K, and show on ViLP that a stronger information bottleneck can yield gains over standard LoRA. --- ### HippMetric: A skeletal-representation-based framework for cross-sectional and longitudinal hippocampal substructural morphometry Authors: Na Gao, Chenfei Ye, Yanwu Yang, Anqi Li, Zhengbo He, Li Liang, Zhiyuan Liu, Xingyu Hao, Ting Ma, Tengfei Guo Published: 2025-12-22 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.19214v1 Accurate characterization of hippocampal substructure is crucial for detecting subtle structural changes and identifying early neurodegenerative biomarkers. However, high inter-subject variability and complex folding pattern of human hippocampus hinder consistent cross-subject and longitudinal analysis. Most existing approaches rely on subject-specific modelling and lack a stable intrinsic coordinate system to accommodate anatomical variability, which limits their ability to establish reliable inter- and intra-individual correspondence. To address this, we propose HippMetric, a skeletal representation (s-rep)-based framework for hippocampal substructural morphometry and point-wise correspondence across individuals and scans. HippMetric builds on the Axis-Referenced Morphometric Model (ARMM) and employs a deformable skeletal coordinate system aligned with hippocampal anatomy and function, providing a biologically grounded reference for correspondence. Our framework comprises two core modules: a skeletal-based coordinate system that respects the hippocampus' conserved longitudinal lamellar architecture, in which functional units (lamellae) are stacked perpendicular to the long-axis, enabling anatomically consistent localization across subjects and time; and individualized s-reps generated through surface reconstruction, deformation, and geometrically constrained spoke refinement, enforcing boundary adherence, orthogonality and non-intersection to produce mathematically valid skeletal geometry. Extensive experiments on two international cohorts demonstrate that HippMetric achieves higher accuracy, reliability, and correspondence stability compared to existing shape models. --- ### InvCoSS: Inversion-driven Continual Self-supervised Learning in Medical Multi-modal Image Pre-training Authors: Zihao Luo, Shaohao Rui, Zhenyu Tang, Guotai Wang, Xiaosong Wang Published: 2025-12-22 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.19213v1 Continual self-supervised learning (CSSL) in medical imaging trains a foundation model sequentially, alleviating the need for collecting multi-modal images for joint training and offering promising improvements in downstream performance while preserving data privacy. However, most existing methods still rely on replaying data from previous stages to prevent catastrophic forgetting, which compromises privacy and limits their applicability in real-world scenarios where data transfer across sites is often restricted. In this work, we propose InvCoSS, an inversion-driven continual self-supervised learning framework for medical multi-modal image pre-training. Specifically, after training on a previous task, InvCoSS inverts the pre-trained self-supervised model to generate synthetic images that approximate the original training distribution. These synthetic images are then combined with data from the new task for joint optimization, which effectively mitigates catastrophic forgetting while strictly adhering to the constraint of no access to previous real data. Furthermore, to improve the fidelity of synthetic images, we introduce a novel InvUNet with a multi-scale fusion architecture to restore both high- and low-frequency components of the inverted images. To enhance diversity and prevent mode collapse, we design a repulsive representation-learning mechanism that encourages a diverse feature space for synthetic images without class guidance. Extensive experiments across nine downstream tasks validate the effectiveness of InvCoSS, achieving performance comparable to or even superior to prior data-replay methods while significantly reducing storage requirements and eliminating data privacy constraints. --- ### Observer, Not Player: Simulating Theory of Mind in LLMs through Game Observation Authors: Jerry Wang, Ting Yiu Liu Published: 2025-12-22 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.19210v1 We present an interactive framework for evaluating whether large language models (LLMs) exhibit genuine "understanding" in a simple yet strategic environment. As a running example, we focus on Rock-Paper-Scissors (RPS), which, despite its apparent simplicity, requires sequential reasoning, adaptation, and strategy recognition. Our system positions the LLM as an Observer whose task is to identify which strategies are being played and to articulate the reasoning behind this judgment. The purpose is not to test knowledge of Rock-Paper-Scissors itself, but to probe whether the model can exhibit mind-like reasoning about sequential behavior. To support systematic evaluation, we provide a benchmark consisting of both static strategies and lightweight dynamic strategies specified by well-prompted rules. We quantify alignment between the Observer's predictions and the ground-truth distributions induced by actual strategy pairs using three complementary signals: Cross-Entropy, Brier score, and Expected Value (EV) discrepancy. These metrics are further integrated into a unified score, the Union Loss, which balances calibration, sensitivity, and payoff alignment. Together with a Strategy Identification Rate (SIR) metric, our framework captures not only predictive accuracy but also whether the model can stably identify the latent strategies in play. The demo emphasizes interactivity, transparency, and reproducibility. Users can adjust LLM distributions in real time, visualize losses as they evolve, and directly inspect reasoning snippets to identify where and why failures occur. In doing so, our system provides a practical and interpretable proxy for mind-like inference in sequential games, offering insights into both the strengths and limitations of current LLM reasoning. --- ### MixKVQ: Query-Aware Mixed-Precision KV Cache Quantization for Long-Context Reasoning Authors: Tao Zhang, Ziqian Zeng, Hao Peng, Huiping Zhuang, Cen Chen Published: 2025-12-22 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2512.19206v1 Long Chain-of-Thought (CoT) reasoning has significantly advanced the capabilities of Large Language Models (LLMs), but this progress is accompanied by substantial memory and latency overhead from the extensive Key-Value (KV) cache. Although KV cache quantization is a promising compression technique, existing low-bit quantization methods often exhibit severe performance degradation on complex reasoning tasks. Fixed-precision quantization struggles to handle outlier channels in the key cache, while current mixed-precision strategies fail to accurately identify components requiring high-precision representation. We find that an effective low-bit KV cache quantization strategy must consider two factors: a key channel's intrinsic quantization difficulty and its relevance to the query. Based on this insight, we propose MixKVQ, a novel plug-and-play method that introduces a lightweight, query-aware algorithm to identify and preserve critical key channels that need higher precision, while applying per-token quantization for value cache. Experiments on complex reasoning datasets demonstrate that our approach significantly outperforms existing low-bit methods, achieving performance comparable to a full-precision baseline at a substantially reduced memory footprint. --- ### On the Koopman-Based Generalization Bounds for Multi-Task Deep Learning Authors: Mahdi Mohammadigohari, Giuseppe Di Fatta, Giuseppe Nicosia, Panos M. Pardalos Published: 2025-12-22 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2512.19199v2 The paper establishes generalization bounds for multitask deep neural networks using operator-theoretic techniques. The authors propose a tighter bound than those derived from conventional norm based methods by leveraging small condition numbers in the weight matrices and introducing a tailored Sobolev space as an expanded hypothesis space. This enhanced bound remains valid even in single output settings, outperforming existing Koopman based bounds. The resulting framework maintains key advantages such as flexibility and independence from network width, offering a more precise theoretical understanding of multitask deep learning in the context of kernel methods. --- ### Adaptive Probability Flow Residual Minimization for High-Dimensional Fokker-Planck Equations Authors: Xiaolong Wu, Qifeng Liao Published: 2025-12-22 Categories: physics.comp-ph, cs.LG, math.NA arXiv: https://arxiv.org/abs/2512.19196v4 Solving high-dimensional Fokker-Planck (FP) equations remains a challenging problem in computational physics and stochastic dynamics, due to the curse of dimensionality, unbounded domains, and complex probability landscapes. In this work, we propose an adaptive probability flow residual minimization (A-PFRM) method for this problem. The second-order FP equation is reformulated as an equivalent first-order continuity equation associated with the probability flow ordinary differential equation, based on which a loss function is constructed to train neural network approximations without Hessian computation. To further improve the computational efficiency, the Hutchinson trace estimator is applied to compute the divergence in the corresponding score function, such that the training time can be dimension-independent on GPUs. Adaptive sampling strategies are employed to generate the collocation points, and our analysis shows that the Kullback-Leibler divergence between our A-PFRM approximation and the exact solution is bounded by the residual loss weighted by the estimated density function. Numerical experiments are presented to demonstrate the performance of A-PFRM, which include Ornstein-Uhlenbeck (OU) processes problems, Brownian motions with time-varying diffusion, and Geometric OU processes featuring non-Gaussian solutions up to one hundred dimensions. --- ### Causal Heterogeneous Graph Learning Method for Chronic Obstructive Pulmonary Disease Prediction Authors: Leming Zhou, Zuo Wang, Zhigang Liu Published: 2025-12-22 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.19194v1 Due to the insufficient diagnosis and treatment capabilities at the grassroots level, there are still deficiencies in the early identification and early warning of acute exacerbation of Chronic obstructive pulmonary disease (COPD), often resulting in a high prevalence rate and high burden, but the screening rate is relatively low. In order to gradually improve this situation. In this paper, this study develop a Causal Heterogeneous Graph Representation Learning (CHGRL) method for COPD comorbidity risk prediction method that: a) constructing a heterogeneous Our dataset includes the interaction between patients and diseases; b) A cause-aware heterogeneous graph learning architecture has been constructed, combining causal inference mechanisms with heterogeneous graph learning, which can support heterogeneous graph causal learning for different types of relationships; and c) Incorporate the causal loss function in the model design, and add counterfactual reasoning learning loss and causal regularization loss on the basis of the cross-entropy classification loss. We evaluate our method and compare its performance with strong GNN baselines. Following experimental evaluation, the proposed model demonstrates high detection accuracy. --- ### PEDESTRIAN: An Egocentric Vision Dataset for Obstacle Detection on Pavements Authors: Marios Thoma, Zenonas Theodosiou, Harris Partaourides, Vassilis Vassiliades, Loizos Michael, Andreas Lanitis Published: 2025-12-22 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2512.19190v1 Walking has always been a primary mode of transportation and is recognized as an essential activity for maintaining good health. Despite the need for safe walking conditions in urban environments, sidewalks are frequently obstructed by various obstacles that hinder free pedestrian movement. Any object obstructing a pedestrian's path can pose a safety hazard. The advancement of pervasive computing and egocentric vision techniques offers the potential to design systems that can automatically detect such obstacles in real time, thereby enhancing pedestrian safety. The development of effective and efficient identification algorithms relies on the availability of comprehensive and well-balanced datasets of egocentric data. In this work, we introduce the PEDESTRIAN dataset, comprising egocentric data for 29 different obstacles commonly found on urban sidewalks. A total of 340 videos were collected using mobile phone cameras, capturing a pedestrian's point of view. Additionally, we present the results of a series of experiments that involved training several state-of-the-art deep learning algorithms using the proposed dataset, which can be used as a benchmark for obstacle detection and recognition tasks. The dataset can be used for training pavement obstacle detectors to enhance the safety of pedestrians in urban areas. --- ### Operator-Based Generalization Bound for Deep Learning: Insights on Multi-Task Learning Authors: Mahdi Mohammadigohari, Giuseppe Di Fatta, Giuseppe Nicosia, Panos M. Pardalos Published: 2025-12-22 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2512.19184v2 This paper presents novel generalization bounds for vector-valued neural networks and deep kernel methods, focusing on multi-task learning through an operator-theoretic framework. Our key development lies in strategically combining a Koopman based approach with existing techniques, achieving tighter generalization guarantees compared to traditional norm-based bounds. To mitigate computational challenges associated with Koopman-based methods, we introduce sketching techniques applicable to vector valued neural networks. These techniques yield excess risk bounds under generic Lipschitz losses, providing performance guarantees for applications including robust and multiple quantile regression. Furthermore, we propose a novel deep learning framework, deep vector-valued reproducing kernel Hilbert spaces (vvRKHS), leveraging Perron Frobenius (PF) operators to enhance deep kernel methods. We derive a new Rademacher generalization bound for this framework, explicitly addressing underfitting and overfitting through kernel refinement strategies. This work offers novel insights into the generalization properties of multitask learning with deep learning architectures, an area that has been relatively unexplored until recent developments. --- ### Practical Quantum-Classical Feature Fusion for complex data Classification Authors: Azadeh Alavi, Fatemeh Kouchmeshki, Abdolrahman Alavi Published: 2025-12-22 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2512.19180v1 Hybrid quantum and classical learning aims to couple quantum feature maps with the robustness of classical neural networks, yet most architectures treat the quantum circuit as an isolated feature extractor and merge its measurements with classical representations by direct concatenation. This neglects that the quantum and classical branches constitute distinct computational modalities and limits reliable performance on complex, high dimensional tabular and semi structured data, including remote sensing, environmental monitoring, and medical diagnostics. We present a multimodal formulation of hybrid learning and propose a cross attention mid fusion architecture in which a classical representation queries quantum derived feature tokens through an attention block with residual connectivity. The quantum branch is kept within practical NISQ budgets and uses up to nine qubits. We evaluate on Wine, Breast Cancer, Forest CoverType, FashionMNIST, and SteelPlatesFaults, comparing a quantum only model, a classical baseline, residual hybrid models, and the proposed mid fusion model under a consistent protocol. Pure quantum and standard hybrid designs underperform due to measurement induced information loss, while cross attention mid fusion is consistently competitive and improves performance on the more complex datasets in most cases. These findings suggest that quantum derived information becomes most valuable when integrated through principled multimodal fusion rather than used in isolation or loosely appended to classical features. --- ### Vision-Language-Policy Model for Dynamic Robot Task Planning Authors: Jin Wang, Kim Tien Ly, Jacques Cloete, Jin Jin, Nikos Tsagarakis, Ioannis Havoutis Published: 2025-12-22 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2512.19178v2 Bridging the gap between natural language commands and autonomous execution in unstructured environments remains an open challenge for robotics. This requires robots to perceive and reason over the current task scene through multiple modalities, and to plan their behaviors to achieve their intended goals. Traditional robotic task-planning approaches often struggle to bridge low-level execution with high-level task reasoning, and cannot dynamically update task strategies when instructions change during execution, which ultimately limits their versatility and adaptability to new tasks. In this work, we propose a novel language model-based framework for dynamic robot task planning. Our Vision-Language-Policy (VLP) model, based on a vision-language model fine-tuned on real-world data, can interpret semantic instructions and integrate reasoning over the current task scene to generate behavior policies that control the robot to accomplish the task. Moreover, it can dynamically adjust the task strategy in response to changes in the task, enabling flexible adaptation to evolving task requirements. Experiments conducted with different robots and a variety of real-world tasks show that the trained model can efficiently adapt to novel scenarios and dynamically update its policy, demonstrating strong planning autonomy and cross-embodiment generalization. Videos: https://robovlp.github.io/ --- ### CycleChart: A Unified Consistency-Based Learning Framework for Bidirectional Chart Understanding and Generation Authors: Dazhen Deng, Sen Yang, Yuchen He, Yuan Tian, Yingcai Wu Published: 2025-12-22 Categories: cs.CL, cs.CV arXiv: https://arxiv.org/abs/2512.19173v2 Current chart-related tasks, such as chart generation (NL2Chart), chart schema parsing, chart data parsing, and chart question answering (ChartQA), are typically studied in isolation, preventing models from learning the shared semantics that link chart creation and interpretation. We introduce CycleChart, a consistency-based learning framework for bidirectional chart understanding and generation. Unlike conventional multi-task approaches that draw training samples independently across tasks, CycleChart organizes all tasks around each single data instance. From a source table and natural-language query, the model generates a chart specification, renders and executes it, then learns to recover the schema and underlying data from the resulting chart image. This per-instance lifecycle design lets the model capture the full chain of transformations, from raw data through visual encoding to structured recovery, and a generate--parse consistency objective enforces semantic alignment between the forward generation and reverse parsing directions. To support this framework, we construct CycleChart-Bench, a lifecycle-aligned benchmark where every chart sample carries aligned annotations for generation, schema parsing, data parsing, and question answering. CycleChart achieves strong results across all four tasks and transfers effectively to unseen external benchmarks, demonstrating improved cross-task generalization and marking a step toward more general chart understanding models. --- ### Finite-sample guarantees for data-driven forward-backward operator methods Authors: Filippo Fabiani, Barbara Franci Published: 2025-12-22 Categories: math.OC, cs.LG, eess.SY arXiv: https://arxiv.org/abs/2512.19172v1 We establish finite sample certificates on the quality of solutions produced by data-based forward-backward (FB) operator splitting schemes. As frequently happens in stochastic regimes, we consider the problem of finding a zero of the sum of two operators, where one is either unavailable in closed form or computationally expensive to evaluate, and shall therefore be approximated using a finite number of noisy oracle samples. Under the lens of algorithmic stability, we then derive probabilistic bounds on the distance between a true zero and the FB output without making specific assumptions about the underlying data distribution. We show that under weaker conditions ensuring the convergence of FB schemes, stability bounds grow proportionally to the number of iterations. Conversely, stronger assumptions yield stability guarantees that are independent of the iteration count. We then specialize our results to a popular FB stochastic Nash equilibrium seeking algorithm and validate our theoretical bounds on a control problem for smart grids, where the energy price uncertainty is approximated by means of historical data. --- ### JEPA-Reasoner: Decoupling Latent Reasoning from Token Generation Authors: Bingyang Kelvin Liu, Ziyu Patrick Chen, David P. Woodruff Published: 2025-12-22 Categories: cs.CL arXiv: https://arxiv.org/abs/2512.19171v3 Current autoregressive language models couple high-level reasoning and low-level token generation into a single sequential process, making the reasoning trajectory vulnerable to compounding expression errors. We propose JEPA-Reasoner, a novel architectural paradigm that decouples these tasks using a Joint-Embedding Predictive Architecture (JEPA) for pure latent-space reasoning and a separate Talker module for linguistic reconstruction. By isolating the reasoning engine from the discrete token-sampling process, our architecture enables: (1) Error Containment, where token-level failures cannot propagate into the latent reasoning chain; (2) Continuous Guidance, providing the generator with access to the entire lossless reasoning trajectory; and (3) Representation of Uncertainty, allowing the model to maintain multiple hypotheses via mixed latent vectors. Controlled experiments on synthetic and natural language tasks demonstrate that this decoupling enables a 0.9B model to achieve a 149.5\% improvement in 8-shot GSM8K accuracy over a coupled Transformer baseline trained on identical data. This work shifts the focus from scaling coupled models to investigating decoupled architectures as a more robust foundation for complex reasoning. --- ### From Speech to Subtitles: Evaluating ASR Models in Subtitling Italian Television Programs Authors: Alessandro Lucca, Francesco Corso, Francesco Pierri Published: 2025-12-22 Categories: cs.CL arXiv: https://arxiv.org/abs/2512.19161v2 Subtitles are essential for video accessibility and audience engagement. Modern Automatic Speech Recognition (ASR) systems, built upon Encoder-Decoder neural network architectures and trained on massive amounts of data, have progressively reduced transcription errors on standard benchmark datasets. However, their performance in real-world production environments, particularly for non-English content like long-form Italian videos, remains largely unexplored. This paper presents a case study on developing a professional subtitling system for an Italian media company. To inform our system design, we evaluated four state-of-the-art ASR models (Whisper Large v2, AssemblyAI Universal, Parakeet TDT v3 0.6b, and WhisperX) on a 50-hour dataset of Italian television programs. The study highlights their strengths and limitations, benchmarking their performance against the work of professional human subtitlers. The findings indicate that, while current models cannot meet the media industry's accuracy needs for full autonomy, they can serve as highly effective tools for enhancing human productivity. We conclude that a human-in-the-loop (HITL) approach is crucial and present the production-grade, cloud-based infrastructure we designed to support this workflow. --- ### OmniMoGen: Unifying Human Motion Generation via Learning from Interleaved Text-Motion Instructions Authors: Wendong Bu, Kaihang Pan, Yuze Lin, Jiacheng Li, Kai Shen, Wenqiao Zhang, Juncheng Li, Jun Xiao, Siliang Tang Published: 2025-12-22 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.19159v1 Large language models (LLMs) have unified diverse linguistic tasks within a single framework, yet such unification remains unexplored in human motion generation. Existing methods are confined to isolated tasks, limiting flexibility for free-form and omni-objective generation. To address this, we propose OmniMoGen, a unified framework that enables versatile motion generation through interleaved text-motion instructions. Built upon a concise RVQ-VAE and transformer architecture, OmniMoGen supports end-to-end instruction-driven motion generation. We construct X2Mo, a large-scale dataset of over 137K interleaved text-motion instructions, and introduce AnyContext, a benchmark for evaluating interleaved motion generation. Experiments show that OmniMoGen achieves state-of-the-art performance on text-to-motion, motion editing, and AnyContext, exhibiting emerging capabilities such as compositional editing, self-reflective generation, and knowledge-informed generation. These results mark a step toward the next intelligent motion generation. Project Page: https://OmniMoGen.github.io/. --- ### Can We Test Consciousness Theories on AI? Ablations, Markers, and Robustness Authors: Yin Jun Phua Published: 2025-12-22 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.19155v1 The search for reliable indicators of consciousness has fragmented into competing theoretical camps (Global Workspace Theory (GWT), Integrated Information Theory (IIT), and Higher-Order Theories (HOT)), each proposing distinct neural signatures. We adopt a synthetic neuro-phenomenology approach: constructing artificial agents that embody these mechanisms to test their functional consequences through precise architectural ablations impossible in biological systems. Across three experiments, we report dissociations suggesting these theories describe complementary functional layers rather than competing accounts. In Experiment 1, a no-rewire Self-Model lesion abolishes metacognitive calibration while preserving first-order task performance, yielding a synthetic blindsight analogue consistent with HOT predictions. In Experiment 2, workspace capacity proves causally necessary for information access: a complete workspace lesion produces qualitative collapse in access-related markers, while partial reductions show graded degradation, consistent with GWT's ignition framework. In Experiment 3, we uncover a broadcast-amplification effect: GWT-style broadcasting amplifies internal noise, creating extreme fragility. The B2 agent family is robust to the same latent perturbation; this robustness persists in a Self-Model-off / workspace-read control, cautioning against attributing the effect solely to $z_{\text{self}}$ compression. We also report an explicit negative result: raw perturbational complexity (PCI-A) decreases under the workspace bottleneck, cautioning against naive transfer of IIT-adjacent proxies to engineered agents. These results suggest a hierarchical design principle: GWT provides broadcast capacity, while HOT provides quality control. We emphasize that our agents are not conscious; they are reference implementations for testing functional predictions of consciousness theories. --- ### Beyond Sliding Windows: Learning to Manage Memory in Non-Markovian Environments Authors: Geraud Nangue Tasse, Matthew Riemer, Benjamin Rosman, Tim Klinger Published: 2025-12-22 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2512.19154v1 Recent success in developing increasingly general purpose agents based on sequence models has led to increased focus on the problem of deploying computationally limited agents within the vastly more complex real-world. A key challenge experienced in these more realistic domains is highly non-Markovian dependencies with respect to the agent's observations, which are less common in small controlled domains. The predominant approach for dealing with this in the literature is to stack together a window of the most recent observations (Frame Stacking), but this window size must grow with the degree of non-Markovian dependencies, which results in prohibitive computational and memory requirements for both action inference and learning. In this paper, we are motivated by the insight that in many environments that are highly non-Markovian with respect to time, the environment only causally depends on a relatively small number of observations over that time-scale. A natural direction would then be to consider meta-algorithms that maintain relatively small adaptive stacks of memories such that it is possible to express highly non-Markovian dependencies with respect to time while considering fewer observations at each step and thus experience substantial savings in both compute and memory requirements. Hence, we propose a meta-algorithm (Adaptive Stacking) for achieving exactly that with convergence guarantees and quantify the reduced computation and memory constraints for MLP, LSTM, and Transformer-based agents. Our experiments utilize popular memory tasks, which give us control over the degree of non-Markovian dependencies. This allows us to demonstrate that an appropriate meta-algorithm can learn the removal of memories not predictive of future rewards without excessive removal of important experiences. Code: https://github.com/geraudnt/adaptive-stacking --- ### AMap: Distilling Future Priors for Ahead-Aware Online HD Map Construction Authors: Ruikai Li, Xinrun Li, Mengwei Xie, Hao Shan, Shoumeng Qiu, Xinyuan Chang, Yizhe Fan, Feng Xiong, Han Jiang, Yilong Ren, Haiyang Yu, Mu Xu et al. Published: 2025-12-22 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.19150v1 Online High-Definition (HD) map construction is pivotal for autonomous driving. While recent approaches leverage historical temporal fusion to improve performance, we identify a critical safety flaw in this paradigm: it is inherently ``spatially backward-looking." These methods predominantly enhance map reconstruction in traversed areas, offering minimal improvement for the unseen road ahead. Crucially, our analysis of downstream planning tasks reveals a severe asymmetry: while rearward perception errors are often tolerable, inaccuracies in the forward region directly precipitate hazardous driving maneuvers. To bridge this safety gap, we propose AMap, a novel framework for Ahead-aware online HD Mapping. We pioneer a ``distill-from-future" paradigm, where a teacher model with privileged access to future temporal contexts guides a lightweight student model restricted to the current frame. This process implicitly compresses prospective knowledge into the student model, endowing it with ``look-ahead" capabilities at zero inference-time cost. Technically, we introduce a Multi-Level BEV Distillation strategy with spatial masking and an Asymmetric Query Adaptation module to effectively transfer future-aware representations to the student's static queries. Extensive experiments on the nuScenes and Argoverse 2 benchmark demonstrate that AMap significantly enhances current-frame perception. Most notably, it outperforms state-of-the-art temporal models in critical forward regions while maintaining the efficiency of single current frame inference. --- ### RP-CATE: Recurrent Perceptron-based Channel Attention Transformer Encoder for Industrial Hybrid Modeling Authors: Haoran Yang, Yinan Zhang, Wenjie Zhang, Dongxia Wang, Peiyu Liu, Yuqi Ye, Kexin Chen, Wenhai Wang Published: 2025-12-22 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.19147v1 Nowadays, industrial hybrid modeling which integrates both mechanistic modeling and machine learning-based modeling techniques has attracted increasing interest from scholars due to its high accuracy, low computational cost, and satisfactory interpretability. Nevertheless, the existing industrial hybrid modeling methods still face two main limitations. First, current research has mainly focused on applying a single machine learning method to one specific task, failing to develop a comprehensive machine learning architecture suitable for modeling tasks, which limits their ability to effectively represent complex industrial scenarios. Second, industrial datasets often contain underlying associations (e.g., monotonicity or periodicity) that are not adequately exploited by current research, which can degrade model's predictive performance. To address these limitations, this paper proposes the Recurrent Perceptron-based Channel Attention Transformer Encoder (RP-CATE), with three distinctive characteristics: 1: We developed a novel architecture by replacing the self-attention mechanism with channel attention and incorporating our proposed Recurrent Perceptron (RP) Module into Transformer, achieving enhanced effectiveness for industrial modeling tasks compared to the original Transformer. 2: We proposed a new data type called Pseudo-Image Data (PID) tailored for channel attention requirements and developed a cyclic sliding window method for generating PID. 3: We introduced the concept of Pseudo-Sequential Data (PSD) and a method for converting industrial datasets into PSD, which enables the RP Module to capture the underlying associations within industrial dataset more effectively. An experiment aimed at hybrid modeling in chemical engineering was conducted by using RP-CATE and the experimental results demonstrate that RP-CATE achieves the best performance compared to other baseline models. --- ### A Convex Loss Function for Set Prediction with Optimal Trade-offs Between Size and Conditional Coverage Authors: Francis Bach Published: 2025-12-22 Categories: cs.LG, math.OC, stat.ML arXiv: https://arxiv.org/abs/2512.19142v1 We consider supervised learning problems in which set predictions provide explicit uncertainty estimates. Using Choquet integrals (a.k.a. Lov{á}sz extensions), we propose a convex loss function for nondecreasing subset-valued functions obtained as level sets of a real-valued function. This loss function allows optimal trade-offs between conditional probabilistic coverage and the ''size'' of the set, measured by a non-decreasing submodular function. We also propose several extensions that mimic loss functions and criteria for binary classification with asymmetric losses, and show how to naturally obtain sets with optimized conditional coverage. We derive efficient optimization algorithms, either based on stochastic gradient descent or reweighted least-squares formulations, and illustrate our findings with a series of experiments on synthetic datasets for classification and regression tasks, showing improvements over approaches that aim for marginal coverage. --- ### Understanding Chain-of-Thought in Large Language Models via Topological Data Analysis Authors: Chenghao Li, Chaoning Zhang, Yi Lu, Shuxu Chen, Xudong Wang, Jiaquan Zhang, Zhicheng Wang, Zhengxun Jin, Kuien Liu, Sung-Ho Bae, Guoqing Wang, Yang Yang et al. Published: 2025-12-22 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.19135v2 With the development of large language models (LLMs), particularly with the introduction of the long reasoning chain technique, the reasoning ability of LLMs in complex problem-solving has been significantly enhanced. While acknowledging the power of long reasoning chains, we cannot help but wonder: Why do different reasoning chains perform differently in reasoning? What components of the reasoning chains play a key role? Existing studies mainly focus on evaluating reasoning chains from a functional perspective, with little attention paid to their structural mechanisms. To address this gap, this work is the first to analyze and evaluate the quality of the reasoning chain from a structural perspective. We apply persistent homology from Topological Data Analysis (TDA) to map reasoning steps into semantic space, extract topological features, and analyze structural changes. These changes reveal semantic coherence, logical redundancy, and identify logical breaks and gaps. By calculating homology groups, we assess connectivity and redundancy at various scales, using barcode and persistence diagrams to quantify stability and consistency. Our results show that the topological structural complexity of reasoning chains correlates positively with accuracy. More complex chains identify correct answers sooner, while successful reasoning exhibits simpler topologies, reducing redundancy and cycles, enhancing efficiency and interpretability. This work provides a new perspective on reasoning chain quality assessment and offers guidance for future optimization. --- ### QuCo-RAG: Quantifying Uncertainty from the Pre-training Corpus for Dynamic Retrieval-Augmented Generation Authors: Dehai Min, Kailin Zhang, Tongtong Wu, Lu Cheng Published: 2025-12-22 Categories: cs.CL, cs.IR arXiv: https://arxiv.org/abs/2512.19134v2 Dynamic Retrieval-Augmented Generation adaptively determines when to retrieve during generation to mitigate hallucinations in large language models (LLMs). However, existing methods rely on model-internal signals (e.g., logits, entropy), which are fundamentally unreliable because LLMs are typically ill-calibrated and often exhibit high confidence in erroneous outputs. We propose QuCo-RAG, which shifts from subjective confidence to objective statistics computed from pre-training data. Our method quantifies uncertainty through two stages: (1) before generation, we identify low-frequency entities indicating long-tail knowledge gaps; (2) during generation, we verify entity co-occurrence in the pre-training corpus, where zero co-occurrence often signals hallucination risk. Both stages leverage Infini-gram for millisecond-latency queries over 4 trillion tokens, triggering retrieval when uncertainty is high. Experiments on multi-hop QA benchmarks show QuCo-RAG achieves EM gains of 5--12 points over state-of-the-art baselines with OLMo-2 models, and transfers effectively to models with undisclosed pre-training data (Llama-3, Qwen2.5, GPT-4.1/5-chat), improving EM by up to 14 points. Generalization to long-form generation and biomedical QA further validates the robustness of our paradigm. These results establish corpus-grounded verification as a principled, practically model-agnostic paradigm for dynamic RAG. Our code is publicly available at https://github.com/ZhishanQ/QuCo-RAG. --- ### WorldRFT: Latent World Model Planning with Reinforcement Fine-Tuning for Autonomous Driving Authors: Pengxuan Yang, Ben Lu, Zhongpu Xia, Chao Han, Yinfeng Gao, Teng Zhang, Kun Zhan, XianPeng Lang, Yupeng Zheng, Qichao Zhang Published: 2025-12-22 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2512.19133v1 Latent World Models enhance scene representation through temporal self-supervised learning, presenting a perception annotation-free paradigm for end-to-end autonomous driving. However, the reconstruction-oriented representation learning tangles perception with planning tasks, leading to suboptimal optimization for planning. To address this challenge, we propose WorldRFT, a planning-oriented latent world model framework that aligns scene representation learning with planning via a hierarchical planning decomposition and local-aware interactive refinement mechanism, augmented by reinforcement learning fine-tuning (RFT) to enhance safety-critical policy performance. Specifically, WorldRFT integrates a vision-geometry foundation model to improve 3D spatial awareness, employs hierarchical planning task decomposition to guide representation optimization, and utilizes local-aware iterative refinement to derive a planning-oriented driving policy. Furthermore, we introduce Group Relative Policy Optimization (GRPO), which applies trajectory Gaussianization and collision-aware rewards to fine-tune the driving policy, yielding systematic improvements in safety. WorldRFT achieves state-of-the-art (SOTA) performance on both open-loop nuScenes and closed-loop NavSim benchmarks. On nuScenes, it reduces collision rates by 83% (0.30% -> 0.05%). On NavSim, using camera-only sensors input, it attains competitive performance with the LiDAR-based SOTA method DiffusionDrive (87.8 vs. 88.1 PDMS). --- ### Evidential Trust-Aware Model Personalization in Decentralized Federated Learning for Wearable IoT Authors: Murtaza Rangwala, Richard O. Sinnott, Rajkumar Buyya Published: 2025-12-22 Categories: cs.DC, cs.LG arXiv: https://arxiv.org/abs/2512.19131v2 Decentralized federated learning (DFL) enables collaborative model training across edge devices without centralized coordination, offering resilience against single points of failure. However, statistical heterogeneity arising from non-identically distributed local data creates a fundamental challenge: nodes must learn personalized models adapted to their local distributions while selectively collaborating with compatible peers. Existing approaches either enforce a single global model that fits no one well, or rely on heuristic peer selection mechanisms that cannot distinguish between peers with genuinely incompatible data distributions and those with valuable complementary knowledge. We present Murmura, a framework that leverages evidential deep learning to enable trust-aware model personalization in DFL. Our key insight is that epistemic uncertainty from Dirichlet-based evidential models directly indicates peer compatibility: high epistemic uncertainty when a peer's model evaluates local data reveals distributional mismatch, enabling nodes to exclude incompatible influence while maintaining personalized models through selective collaboration. Murmura introduces a trust-aware aggregation mechanism that computes peer compatibility scores through cross-evaluation on local validation samples and personalizes model aggregation based on evidential trust with adaptive thresholds. Evaluation on three wearable IoT datasets (UCI HAR, PAMAP2, PPG-DaLiA) demonstrates that Murmura reduces performance degradation from IID to non-IID conditions compared to baseline (0.9% vs. 19.3%), achieves 7.4$\times$ faster convergence, and maintains stable accuracy across hyperparameter choices. These results establish evidential uncertainty as a principled foundation for compatibility-aware personalization in decentralized heterogeneous environments. --- ### AWPO: Enhancing Tool-Use of Large Language Models through Adaptive Integration of Reasoning Rewards Authors: Zihan Lin, Xiaohan Wang, Hexiong Yang, Jiajun Chai, Jie Cao, Guojun Yin, Wei Lin, Ran He Published: 2025-12-22 Categories: cs.CL arXiv: https://arxiv.org/abs/2512.19126v3 While Reinforcement Learning (RL) shows promise in training tool-use Large Language Models (LLMs) using verifiable outcome rewards, existing methods largely overlook the potential of reasoning rewards based on chain-of-thought quality for better tool utilization. Furthermore, naïvely combining reasoning and outcome rewards may yield suboptimal performance or conflict with the primary optimization objective. To address this, we propose Advantage-Weighted Policy Optimization (AWPO), a principled RL framework that adaptively integrates reasoning rewards into advantage estimation to improve tool-use performance. AWPO incorporates variance-aware gating and difficulty-aware weighting to adaptively modulate advantages from reasoning signals based on group-relative statistics, alongside a tailored clipping mechanism for stable optimization. Extensive experiments demonstrate that AWPO achieves state-of-the-art performance across standard tool-use benchmarks, significantly outperforming strong baselines and leading closed-source models in challenging multi-turn scenarios. Notably, with exceptional parameter efficiency, our 4B model surpasses Grok-4 by $16.0\%$ in multi-turn accuracy while preserving generalization capability on the out-of-distribution MMLU-Pro benchmark. --- ### SAP: Syntactic Attention Pruning for Transformer-based Language Models Authors: Tzu-Yun Lee, Ding-Yong Hong, Jan-Jan Wu Published: 2025-12-22 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2512.19125v1 This paper introduces Syntactic Attention Pruning (SAP), a novel method for effectively pruning attention heads in Transformer models. Unlike conventional approaches that rely solely on mathematical analysis of model weights and activations, SAP incorporates both the syntactic structure and attention patterns of sentences to guide the pruning process. By leveraging these linguistic features, SAP not only achieves performance comparable to state-of-the-art methods but also enhances the interpretability of model behavior. To further improve robustness, we propose Candidate Filtering (CF), a mechanism that prioritizes heads based on their contribution to model performance, mitigating degradation during pruning. Experimental results indicate that SAP effectively preserves critical heads of a high density of strong attention values, outperforming existing head pruning strategies in retrain-free settings. These findings position SAP as a promising foundation for a new direction in model compression research, offering high flexibility for pruning across all transformer-based language models. ---