# Research Papers — 2025-12 · bundle page 153 of 175

> Full-text markdown of papers #7601–7650 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/153
[← Previous bundle](https://aipod.dev/papers/archive/2025-12/page/152.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-12/page/154.md)

### CRAFT-E: A Neuro-Symbolic Framework for Embodied Affordance Grounding

Authors: Zhou Chen, Joe Lin, Carson Bulgin, Sathyanarayanan N. Aakur
Published: 2025-12-03
Categories: cs.RO, cs.AI
arXiv: https://arxiv.org/abs/2512.04231v1

Assistive robots operating in unstructured environments must understand not only what objects are, but what they can be used for. This requires grounding language-based action queries to objects that both afford the requested function and can be physically retrieved. Existing approaches often rely on black-box models or fixed affordance labels, limiting transparency, controllability, and reliability for human-facing applications. We introduce CRAFT-E, a modular neuro-symbolic framework that composes a structured verb-property-object knowledge graph with visual-language alignment and energy-based grasp reasoning. The system generates interpretable grounding paths that expose the factors influencing object selection and incorporates grasp feasibility as an integral part of affordance inference. We further construct a benchmark dataset with unified annotations for verb-object compatibility, segmentation, and grasp candidates, and deploy the full pipeline on a physical robot. CRAFT-E achieves competitive performance in static scenes, ImageNet-based functional retrieval, and real-world trials involving 20 verbs and 39 objects. The framework remains robust under perceptual noise and provides transparent, component-level diagnostics. By coupling symbolic reasoning with embodied perception, CRAFT-E offers an interpretable and customizable alternative to end-to-end models for affordance-grounded object selection, supporting trustworthy decision-making in assistive robotic systems.

---

### Addressing Logical Fallacies In Scientific Reasoning From Large Language Models: Towards a Dual-Inference Training Framework

Authors: Peter B. Walker, Hannah Davidson, Aiden Foster, Matthew Lienert, Thomas Pardue, Dale Russell
Published: 2025-12-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2512.04228v1

Large Language Models (LLMs) have transformed natural language processing and hold growing promise for advancing science, healthcare, and decision-making. Yet their training paradigms remain dominated by affirmation-based inference, akin to \textit{modus ponens}, where accepted premises yield predicted consequents. While effective for generative fluency, this one-directional approach leaves models vulnerable to logical fallacies, adversarial manipulation, and failures in causal reasoning. This paper makes two contributions. First, it demonstrates how existing LLMs from major platforms exhibit systematic weaknesses when reasoning in scientific domains with negation, counterexamples, or faulty premises \footnote{Code to recreate these experiments are at https://github.com/hannahdavidsoncollege-maker/ScientificReasoningForEnvironment-MedicineWithLLMs. Second, it introduces a dual-reasoning training framework that integrates affirmative generation with structured counterfactual denial. Grounded in formal logic, cognitive science, and adversarial training, this training paradigm formalizes a computational analogue of ``denying the antecedent'' as a mechanism for disconfirmation and robustness. By coupling generative synthesis with explicit negation-aware objectives, the framework enables models that not only affirm valid inferences but also reject invalid ones, yielding systems that are more resilient, interpretable, and aligned with human reasoning.

---

### Educational Cone Model in Embedding Vector Spaces

Authors: Yo Ehara
Published: 2025-12-03
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.04227v1

Human-annotated datasets with explicit difficulty ratings are essential in intelligent educational systems. Although embedding vector spaces are widely used to represent semantic closeness and are promising for analyzing text difficulty, the abundance of embedding methods creates a challenge in selecting the most suitable method. This study proposes the Educational Cone Model, which is a geometric framework based on the assumption that easier texts are less diverse (focusing on fundamental concepts), whereas harder texts are more diverse. This assumption leads to a cone-shaped distribution in the embedding space regardless of the embedding method used. The model frames the evaluation of embeddings as an optimization problem with the aim of detecting structured difficulty-based patterns. By designing specific loss functions, efficient closed-form solutions are derived that avoid costly computation. Empirical tests on real-world datasets validated the model's effectiveness and speed in identifying the embedding spaces that are best aligned with difficulty-annotated educational texts.

---

### Black box behavioural modelling: Predicting human activity schedules with a deep conditional generative approach

Authors: Fred Shone, Tim Hillel
Published: 2025-12-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04223v2

Modelling the complexity and diversity of human activity scheduling behaviour is inherently challenging. We demonstrate ActVAE, a deep conditional-generative machine learning approach for the modelling of activity schedules. Suitable for application in activity-based demand modelling frameworks, schedules are modelled as conditional on individual, household and schedule information, such as age, income, and access to public transit. We demonstrate the rapid generation of precise, realistic and diverse schedules dependent on input labels. We extensively evaluate and compare model capabilities against baseline models using a joint-density estimation framework. In addition to providing a novel alternative to existing scheduling approaches, our work highlights the value of explicitly modelling the randomness of complex and diverse human behaviours.

---

### ReasonX: MLLM-Guided Intrinsic Image Decomposition

Authors: Alara Dirik, Tuanfeng Wang, Duygu Ceylan, Stefanos Zafeiriou, Anna Frühstück
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04222v1

Intrinsic image decomposition aims to separate images into physical components such as albedo, depth, normals, and illumination. While recent diffusion- and transformer-based models benefit from paired supervision from synthetic datasets, their generalization to diverse, real-world scenarios remains challenging. We propose ReasonX, a novel framework that leverages a multimodal large language model (MLLM) as a perceptual judge providing relative intrinsic comparisons, and uses these comparisons as GRPO rewards for fine-tuning intrinsic decomposition models on unlabeled, in-the-wild images. Unlike RL methods for generative models, our framework aligns conditional intrinsic predictors by rewarding agreement between the judge's relational assessments and analytically derived relations from the model's outputs. ReasonX is model-agnostic and can be applied to different intrinsic predictors. Across multiple base architectures and modalities, ReasonX yields significant improvements, including 9-25% WHDR reduction on IIW albedo and up to 46% depth accuracy gains on ETH3D, highlighting the promise of MLLM-guided comparative supervision to bridge low- and high-level vision reasoning.

---

### MoReGen: Multi-Agent Motion-Reasoning Engine for Code-based Text-to-Video Synthesis

Authors: Xiangyu Bai, He Liang, Bishoy Galoaa, Utsav Nandi, Shayda Moezzi, Yuhang He, Sarah Ostadabbas
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04221v2

While text-to-video (T2V) generation has achieved remarkable progress in photorealism, generating intent-aligned videos that faithfully obey physics principles remains a core challenge. In this work, we systematically study Newtonian motion-controlled text-to-video generation and evaluation, emphasizing physical precision and motion coherence. We introduce MoReGen, a motion-aware, physics-grounded T2V framework that integrates multi-agent LLMs, physics simulators, and renderers to generate reproducible, physically accurate videos from text prompts in the code domain. To quantitatively assess physical validity, we propose object-trajectory correspondence as a direct evaluation metric and present MoReSet, a benchmark of 1,275 human-annotated videos spanning nine classes of Newtonian phenomena with scene descriptions, spatiotemporal relations, and ground-truth trajectories. Using MoReSet, we conduct experiments on existing T2V models, evaluating their physical validity through both our MoRe metrics and existing physics-based evaluators. Our results reveal that state-of-the-art models struggle to maintain physical validity, while MoReGen establishes a principled direction toward physically coherent video synthesis.

---

### On Group Relative Policy Optimization Collapse in Agent Search: The Lazy Likelihood-Displacement

Authors: Wenlong Deng, Yushu Li, Boying Gong, Yi Ren, Christos Thrampoulidis, Xiaoxiao Li
Published: 2025-12-03
Categories: cs.CL
arXiv: https://arxiv.org/abs/2512.04220v2

Tool-integrated (TI) reinforcement learning (RL) enables large language models (LLMs) to perform multi-step reasoning by interacting with external tools such as search engines and retrievers. Group Relative Policy Optimization (GRPO), exemplified by the recent Search-R1, offers fast convergence and a value-free formulation that makes it appealing for this setting, yet consistently suffers from training collapse. We identify Lazy Likelihood Displacement (LLD), a systematic reduction or stagnation in the likelihood of both correct and incorrect responses, as the core mechanism driving this failure. LLD emerges early and triggers a self-reinforcing LLD Death Spiral, where declining likelihood leads to low-confidence responses, inflating gradients, and ultimately causing collapse. We empirically characterize this process across models on a Search-R1-style, search-integrated question answering task, revealing a consistent three-phase trajectory: early stagnation, steady decay, and accelerated collapse. To address this, we propose a likelihood-preserving regularization LLDS that activates only when a response action's likelihood decreases, and regularizes only the tokens responsible. This fine-grained structure mitigates LLD with minimal interference. Our method stabilizes training, prevents gradient explosion, and yields substantial performance improvements across seven benchmarks, including relative improvements of +45.2% on Qwen2.5-3B and +37.1% on Qwen2.5-7B over vanilla GRPO training. Our results establish LLD as a previously overlooked bottleneck in GRPO-based TIRL and provide a practical path toward stable, scalable training of tool-integrated RL.

---

### Generalized Event Partonomy Inference with Structured Hierarchical Predictive Learning

Authors: Zhou Chen, Joe Lin, Sathyanarayanan N. Aakur\\
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04219v1

Humans naturally perceive continuous experience as a hierarchy of temporally nested events, fine-grained actions embedded within coarser routines. Replicating this structure in computer vision requires models that can segment video not just retrospectively, but predictively and hierarchically. We introduce PARSE, a unified framework that learns multiscale event structure directly from streaming video without supervision. PARSE organizes perception into a hierarchy of recurrent predictors, each operating at its own temporal granularity: lower layers model short-term dynamics while higher layers integrate longer-term context through attention-based feedback. Event boundaries emerge naturally as transient peaks in prediction error, yielding temporally coherent, nested partonomies that mirror the containment relations observed in human event perception. Evaluated across three benchmarks, Breakfast Actions, 50 Salads, and Assembly 101, PARSE achieves state-of-the-art performance among streaming methods and rivals offline baselines in both temporal alignment (H-GEBD) and structural consistency (TED, hF1). The results demonstrate that predictive learning under uncertainty provides a scalable path toward human-like temporal abstraction and compositional event understanding.

---

### Look Around and Pay Attention: Multi-camera Point Tracking Reimagined with Transformers

Authors: Bishoy Galoaa, Xiangyu Bai, Shayda Moezzi, Utsav Nandi, Sai Siddhartha Vivek Dhir Rangoju, Somaieh Amraee, Sarah Ostadabbas
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04213v1

This paper presents LAPA (Look Around and Pay Attention), a novel end-to-end transformer-based architecture for multi-camera point tracking that integrates appearance-based matching with geometric constraints. Traditional pipelines decouple detection, association, and tracking, leading to error propagation and temporal inconsistency in challenging scenarios. LAPA addresses these limitations by leveraging attention mechanisms to jointly reason across views and time, establishing soft correspondences through a cross-view attention mechanism enhanced with geometric priors. Instead of relying on classical triangulation, we construct 3D point representations via attention-weighted aggregation, inherently accommodating uncertainty and partial observations. Temporal consistency is further maintained through a transformer decoder that models long-range dependencies, preserving identities through extended occlusions. Extensive experiments on challenging datasets, including our newly created multi-camera (MC) versions of TAPVid-3D panoptic and PointOdyssey, demonstrate that our unified approach significantly outperforms existing methods, achieving 37.5% APD on TAPVid-3D-MC and 90.3% APD on PointOdyssey-MC, particularly excelling in scenarios with complex motions and occlusions. Code is available at https://github.com/ostadabbas/Look-Around-and-Pay-Attention-LAPA-

---

### Balancing Safety and Helpfulness in Healthcare AI Assistants through Iterative Preference Alignment

Authors: Huy Nghiem, Swetasudha Panda, Devashish Khatwani, Huy V. Nguyen, Krishnaram Kenthapadi, Hal Daumé
Published: 2025-12-03
Categories: cs.AI, cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2512.04210v1

Large Language Models (LLMs) are increasingly used in healthcare, yet ensuring their safety and trustworthiness remains a barrier to deployment. Conversational medical assistants must avoid unsafe compliance without over-refusing benign queries. We present an iterative post-deployment alignment framework that applies Kahneman-Tversky Optimization (KTO) and Direct Preference Optimization (DPO) to refine models against domain-specific safety signals. Using the CARES-18K benchmark for adversarial robustness, we evaluate four LLMs (Llama-3B/8B, Meditron-8B, Mistral-7B) across multiple cycles. Our results show up to 42% improvement in safety-related metrics for harmful query detection, alongside interesting trade-offs against erroneous refusals, thereby exposing architecture-dependent calibration biases. We also perform ablation studies to identify when self-evaluation is reliable and when external or finetuned judges are necessary to maximize performance gains. Our findings underscore the importance of adopting best practices that balance patient safety, user trust, and clinical utility in the design of conversational medical assistants.

---

### Orchestrator Multi-Agent Clinical Decision Support System for Secondary Headache Diagnosis in Primary Care

Authors: Xizhi Wu, Nelly Estefanie Garduno-Rapp, Justin F Rousseau, Mounika Thakkallapally, Hang Zhang, Yuelyu Ji, Shyam Visweswaran, Yifan Peng, Yanshan Wang
Published: 2025-12-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2512.04207v2

Unlike most primary headaches, secondary headaches need specialized care and can have devastating consequences if not treated promptly. Clinical guidelines highlight several 'red flag' features, such as thunderclap onset, meningismus, papilledema, focal neurologic deficits, signs of temporal arteritis, systemic illness, and the 'worst headache of their life' presentation. Despite these guidelines, determining which patients require urgent evaluation remains challenging in primary care settings. Clinicians often work with limited time, incomplete information, and diverse symptom presentations, which can lead to under-recognition and inappropriate care. We present a large language model (LLM)-based multi-agent clinical decision support system built on an orchestrator-specialist architecture, designed to perform explicit and interpretable secondary headache diagnosis from free-text clinical vignettes. The multi-agent system decomposes diagnosis into seven domain-specialized agents, each producing a structured and evidence-grounded rationale, while a central orchestrator performs task decomposition and coordinates agent routing. We evaluated the multi-agent system using 90 expert-validated secondary headache cases and compared its performance with a single-LLM baseline across two prompting strategies: question-based prompting (QPrompt) and clinical practice guideline-based prompting (GPrompt). We tested five open-source LLMs (Qwen-30B, GPT-OSS-20B, Qwen-14B, Qwen-8B, and Llama-3.1-8B), and found that the orchestrated multi-agent system with GPrompt consistently achieved the highest F1 scores, with larger gains in smaller models. These findings demonstrate that structured multi-agent reasoning improves accuracy beyond prompt engineering alone and offers a transparent, clinically aligned approach for explainable decision support in secondary headache diagnosis.

---

### Machine Phenomenology: A Simple Equation Classifying Fast Radio Bursts

Authors: Yang Liu, Yuhao Lu, Rahim Moradi, Bo Yang, Bing Zhang, Wenbin Lin, Yu Wang
Published: 2025-12-03
Categories: astro-ph.IM, astro-ph.HE, cs.AI
arXiv: https://arxiv.org/abs/2512.04204v1

This work shows how human physical reasoning can guide machine-driven symbolic regression toward discovering empirical laws from observations. As an example, we derive a simple equation that classifies fast radio bursts (FRBs) into two distinct Gaussian distributions, indicating the existence of two physical classes. This human-AI workflow integrates feature selection, dimensional analysis, and symbolic regression: deep learning first analyzes CHIME Catalog 1 and identifies six independent parameters that collectively provide a complete description of FRBs; guided by Buckingham-$π$ analysis and correlation analysis, humans then construct dimensionless groups; finally, symbolic regression performed by the machine discovers the governing equation. When applied to the newer CHIME Catalog, the equation produces consistent results, demonstrating that it captures the underlying physics. This framework is applicable to a broad range of scientific domains.

---

### Network of Theseus (like the ship)

Authors: Vighnesh Subramaniam, Colin Conwell, Boris Katz, Andrei Barbu, Brian Cheung
Published: 2025-12-03
Categories: cs.LG, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2512.04198v1

A standard assumption in deep learning is that the inductive bias introduced by a neural network architecture must persist from training through inference. The architecture you train with is the architecture you deploy. This assumption constrains the community from selecting architectures that may have desirable efficiency or design properties due to difficulties with optimization. We challenge this assumption with Network of Theseus (NoT), a method for progressively converting a trained, or even untrained, guide network architecture part-by-part into an entirely different target network architecture while preserving the performance of the guide network. At each stage, components in the guide network architecture are incrementally replaced with target architecture modules and aligned via representational similarity metrics. This procedure largely preserves the functionality of the guide network even under substantial architectural changes-for example, converting a convolutional network into a multilayer perceptron, or GPT-2 into a recurrent neural network. By decoupling optimization from deployment, NoT expands the space of viable inference-time architectures, opening opportunities for better accuracy-efficiency tradeoffs and enabling more directed exploration of the architectural design space.

---

### BEP: A Binary Error Propagation Algorithm for Binary Neural Networks Training

Authors: Luca Colombo, Fabrizio Pittorino, Daniele Zambon, Carlo Baldassi, Manuel Roveri, Cesare Alippi
Published: 2025-12-03
Categories: cs.LG, cond-mat.dis-nn, cs.AI
arXiv: https://arxiv.org/abs/2512.04189v2

Binary Neural Networks (BNNs), which constrain both weights and activations to binary values, offer substantial reductions in computational complexity, memory footprint, and energy consumption. These advantages make them particularly well suited for deployment on resource-constrained devices. However, training BNNs via gradient-based optimization remains challenging due to the discrete nature of their variables. The dominant approach, quantization-aware training, circumvents this issue by employing surrogate gradients. Yet, this method requires maintaining latent full-precision parameters and performing the backward pass with floating-point arithmetic, thereby forfeiting the efficiency of binary operations during training. While alternative approaches based on local learning rules exist, they are unsuitable for global credit assignment and for back-propagating errors in multi-layer architectures. This paper introduces Binary Error Propagation (BEP), the first learning algorithm to establish a principled, discrete analog of the backpropagation chain rule. This mechanism enables error signals, represented as binary vectors, to be propagated backward through multiple layers of a neural network. BEP operates entirely on binary variables, with all forward and backward computations performed using only bitwise operations. Crucially, this makes BEP the first solution to enable end-to-end binary training for recurrent neural network architectures. We validate the effectiveness of BEP on both multi-layer perceptrons and recurrent neural networks, demonstrating gains of up to +6.89% and +10.57% in test accuracy, respectively. The proposed algorithm is released as an open-source repository.

---

### OnSight Pathology: A real-time platform-agnostic computational pathology companion for histopathology

Authors: Jinzhen Hu, Kevin Faust, Parsa Babaei Zadeh, Adrienn Bourkas, Shane Eaton, Andrew Young, Anzar Alvi, Dimitrios George Oreopoulos, Ameesha Paliwal, Assem Saleh Alrumeh, Evelyn Rose Kamski-Hennekam, Phedias Diamandis
Published: 2025-12-03
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2512.04187v1

The microscopic examination of surgical tissue remains a cornerstone of disease classification but relies on subjective interpretations and access to highly specialized experts, which can compromise accuracy and clinical care. While emerging breakthroughs in artificial intelligence (AI) offer promise for automated histological analysis, the growing number of proprietary digital pathology solutions has created barriers to real-world deployment. To address these challenges, we introduce OnSight Pathology, a platform-agnostic computer vision software that uses continuous custom screen captures to provide real-time AI inferences to users as they review digital slide images. Accessible as a single, self-contained executable file (https://onsightpathology.github.io/ ), OnSight Pathology operates locally on consumer-grade personal computers without complex software integration, enabling cost-effective and secure deployment in research and clinical workflows. Here we demonstrate the utility of OnSight Pathology using over 2,500 publicly available whole slide images across different slide viewers, as well as cases from our clinical digital pathology setup. The software's robustness is highlighted across routine histopathological tasks, including the classification of common brain tumor types, mitosis detection, and the quantification of immunohistochemical stains. A built-in multi-modal chat assistant provides verifiable descriptions of images, free of rigid class labels, for added quality control. Lastly, we show compatibility with live microscope camera feeds, including from personal smartphones, offering potential for deployment in more analog, inter-operative, and telepathology settings. Together, we highlight how OnSight Pathology can deliver real-time AI inferences across a broad range of pathology pipelines, removing key barriers to the adoption of AI tools in histopathology.

---

### Beyond Flicker: Detecting Kinematic Inconsistencies for Generalizable Deepfake Video Detection

Authors: Alejandro Cobo, Roberto Valle, José Miguel Buenaposada, Luis Baumela
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04175v2

Generalizing deepfake detection to unseen manipulations remains a key challenge. A recent approach to tackle this issue is to train a network with pristine face images that have been manipulated with hand-crafted artifacts to extract more generalizable clues. While effective for static images, extending this to the video domain is an open issue. Existing methods model temporal artifacts as frame-to-frame instabilities, overlooking a key vulnerability: the violation of natural motion dependencies between different facial regions. In this paper, we propose a synthetic video generation method that creates training data with subtle kinematic inconsistencies. We train an autoencoder to decompose facial landmark configurations into motion bases. By manipulating these bases, we selectively break the natural correlations in facial movements and introduce these artifacts into pristine videos via face morphing. A network trained on our data learns to spot these sophisticated biomechanical flaws, achieving state-of-the-art generalization results on several popular benchmarks.

---

### Mitigating the Curse of Detail: Scaling Arguments for Feature Learning and Sample Complexity

Authors: Noa Rubin, Orit Davidovich, Zohar Ringel
Published: 2025-12-03
Categories: cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2512.04165v4

Two pressing topics in the theory of deep learning are the interpretation of feature learning (FL) mechanisms and the determination of implicit bias of networks in the rich regime. Current theories of rich FL often appear in the form of high-dimensional non-linear equations, which require computationally intensive numerical solutions. Given the many details that go into defining a deep learning problem, this analytical complexity is a significant and often unavoidable challenge. Here, we propose a powerful heuristic route for predicting the data and width scales at which various patterns of FL emerge. This form of scale analysis is considerably simpler than such exact theories and reproduces the scaling exponents of various known results. In addition, we make novel predictions on complex toy architectures, such as three-layer non-linear networks and attention heads, thus extending the scope of first-principle theories of deep learning.

---

### Enhancing next token prediction based pre-training for jet foundation models

Authors: Joschka Birk, Anna Hallin, Gregor Kasieczka, Nikol Madzharova, Ian Pang, David Shih
Published: 2025-12-03
Categories: hep-ph, cs.LG, hep-ex, physics.data-an
arXiv: https://arxiv.org/abs/2512.04149v2

Next token prediction is an attractive pre-training task for jet foundation models, in that it is simulation free and enables excellent generative capabilities that can transfer across datasets. Here we study multiple improvements to next token prediction, building on the initial work of OmniJet-$α$. Instead of tokenizing particles and subsequently only using the token-ID as the model input for both the generative and the classification task, we adopt a hybrid setup, which allows us to use continuous feature vectors as model input while only using token-IDs in the next token prediction target. Secondly, we explore a combined pre-training strategy that combines masked particle modeling and generative learning objectives. Taken together, these changes greatly improve the performance in downstream classification tasks without any loss in generative performance.

---

### RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories

Authors: Roy Rinberg, Usha Bhalla, Igor Shilov, Flavio P. Calmon, Rohit Gandikota
Published: 2025-12-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2512.04144v3

Targeted interventions on language models, such as unlearning or model editing, aim to modify specific information, but their effects often propagate to related, unintended areas (e.g., removing virology content may degrade performance on allergies); these side-effects are commonly referred to as the ripple effect. We introduce RippleBench-Maker, an automatic pipeline that retrieves semantic neighbors of any source concept from a knowledge repository and generates multiple-choice questions at varying semantic distances. We instantiate this framework using WikiRAG, an open-source RAG system over English Wikipedia, to construct RippleBench-WMDP-Bio (584 seed topics, 352,961 questions), and evaluate eight unlearning methods on Llama3-8B-Instruct. All eight exhibit accuracy drops that are largest near the unlearned target and decay with semantic distance, each with a distinct propagation profile. We replicate these findings across Mistral-7B, Zephyr-7B, and Yi-34B; cross-model delta curves are nearly identical, suggesting ripple effects are a property of the unlearning method rather than the base model. We validate all major pipeline stages using a four-experiment Mechanical Turk study (5,200+ responses, 61 workers). We release all code, data, and infrastructure.

---

### From FLOPs to Footprints: The Resource Cost of Artificial Intelligence

Authors: Sophia Falk, Nicholas Kluge Corrêa, Sasha Luccioni, Lisa Biber-Freudenberger, Aimee van Wynsberghe
Published: 2025-12-03
Categories: cs.CY, cs.AI, econ.GN
arXiv: https://arxiv.org/abs/2512.04142v1

As computational demands continue to rise, assessing the environmental footprint of AI requires moving beyond energy and water consumption to include the material demands of specialized hardware. This study quantifies the material footprint of AI training by linking computational workloads to physical hardware needs. The elemental composition of the Nvidia A100 SXM 40 GB graphics processing unit (GPU) was analyzed using inductively coupled plasma optical emission spectroscopy, which identified 32 elements. The results show that AI hardware consists of about 90% heavy metals and only trace amounts of precious metals. The elements copper, iron, tin, silicon, and nickel dominate the GPU composition by mass. In a multi-step methodology, we integrate these measurements with computational throughput per GPU across varying lifespans, accounting for the computational requirements of training specific AI models at different training efficiency regimes. Scenario-based analyses reveal that, depending on Model FLOPs Utilization (MFU) and hardware lifespan, training GPT-4 requires between 1,174 and 8,800 A100 GPUs, corresponding to the extraction and eventual disposal of up to 7 tons of toxic elements. Combined software and hardware optimization strategies can reduce material demands: increasing MFU from 20% to 60% lowers GPU requirements by 67%, while extending lifespan from 1 to 3 years yields comparable savings; implementing both measures together reduces GPU needs by up to 93%. Our findings highlight that incremental performance gains, such as those observed between GPT-3.5 and GPT-4, come at disproportionately high material costs. The study underscores the necessity of incorporating material resource considerations into discussions of AI scalability, emphasizing that future progress in AI must align with principles of resource efficiency and environmental responsibility.

---

### Solving N-Queen Problem using Las Vegas Algorithm with State Pruning

Authors: Susmita Sharma, Aayush Shrestha, Sitasma Thapa, Prashant Timalsina, Prakash Poudyal
Published: 2025-12-03
Categories: cs.AI
arXiv: https://arxiv.org/abs/2512.04139v1

The N-Queens problem, placing all N queens in a N x N chessboard where none attack the other, is a classic problem for constraint satisfaction algorithms. While complete methods like backtracking guarantee a solution, their exponential time complexity makes them impractical for large-scale instances thus, stochastic approaches, such as Las Vegas algorithm, are preferred. While it offers faster approximate solutions, it suffers from significant performance variance due to random placement of queens on the board. This research introduces a hybrid algorithm built on top of the standard Las Vegas framework through iterative pruning, dynamically eliminating invalid placements during the random assignment phase, thus this method effectively reduces the search space. The analysis results that traditional backtracking scales poorly with increasing N. In contrast, the proposed technique consistently generates valid solutions more rapidly, establishing it as a superior alternative to use where a single, timely solution is preferred over completeness. Although large N causes some performance variability, the algorithm demonstrates a highly effective trade-off between computational cost and solution fidelity, making it particularly suited for resource-constrained computing environments.

---

### MechDetect: Detecting Data-Dependent Errors

Authors: Philipp Jung, Nicholas Chandler, Sebastian Jäger, Felix Biessmann
Published: 2025-12-03
Categories: cs.LG, cs.DB, cs.IR
arXiv: https://arxiv.org/abs/2512.04138v1

Data quality monitoring is a core challenge in modern information processing systems. While many approaches to detect data errors or shifts have been proposed, few studies investigate the mechanisms governing error generation. We argue that knowing how errors were generated can be key to tracing and fixing them. In this study, we build on existing work in the statistics literature on missing values and propose MechDetect, a simple algorithm to investigate error generation mechanisms. Given a tabular data set and a corresponding error mask, the algorithm estimates whether or not the errors depend on the data using machine learning models. Our work extends established approaches to detect mechanisms underlying missing values and can be readily applied to other error types, provided that an error mask is available. We demonstrate the effectiveness of MechDetect in experiments on established benchmark datasets.

---

### Decoding Large Language Diffusion Models with Foreseeing Movement

Authors: Yichuan Mo, Quan Chen, Mingjie Li, Zeming Wei, Yisen Wang
Published: 2025-12-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04135v1

Large Language Diffusion Models (LLDMs) benefit from a flexible decoding mechanism that enables parallelized inference and controllable generations over autoregressive models. Yet such flexibility introduces a critical challenge: inference performance becomes highly sensitive to the decoding order of tokens. Existing heuristic methods, however, focus mainly on local effects while overlooking long-term impacts. To address this limitation, we propose the Foreseeing Decoding Method (FDM), a novel approach that integrates both local and global considerations to unlock the full potential, employing a search-based strategy to enable effective optimization in discrete spaces. Furthermore, by analyzing the consistency of chosen tokens in the full decoding process, we develop a variant, FDM with Acceleration (FDM-A), which restricts deep exploration to critical steps identified as the exploration and balance circumantences. Extensive experiments across diverse benchmarks and model architectures validate the scalability of FDM and demonstrate the superior efficiency-performance trade-off achieved by FDM-A. Our work might potentially provide a principled step toward more powerful decoding methods for LLDMs.

---

### Artificial Intelligence / Human Intelligence: Who Controls Whom?

Authors: Charlotte Jacquemot
Published: 2025-12-03
Categories: cs.CY, cs.AI
arXiv: https://arxiv.org/abs/2512.04131v1

Using the example of the film 2001: A Space Odyssey, this chapter illustrates the challenges posed by an AI capable of making decisions that go against human interests. But are human decisions always rational and ethical? In reality, the cognitive decision-making process is influenced by cognitive biases that affect our behavior and choices. AI not only reproduces these biases, but can also exploit them, with the potential to shape our decisions and judgments. Behind IA algorithms, there are sometimes individuals who show little concern for fundamental rights and impose their own rules. To address the ethical and societal challenges raised by AI and its governance, the regulation of digital platforms and education are keys levers. Regulation must reflect ethical, legal, and political choices, while education must strengthen digital literacy and teach people to make informed and critical choices when facing digital technologies.

---

### Unique Lives, Shared World: Learning from Single-Life Videos

Authors: Tengda Han, Sayna Ebrahimi, Dilara Gokay, Li Yang Ku, Maks Ovsjanikov, Iva Babukova, Daniel Zoran, Viorica Patraucean, Joao Carreira, Andrew Zisserman, Dima Damen
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04085v2

We introduce the "single-life" learning paradigm, where we train a distinct vision model exclusively on egocentric videos captured by one individual. We leverage the multiple viewpoints naturally captured within a single life to learn a visual encoder in a self-supervised manner. Our experiments demonstrate three key findings. First, models trained independently on different lives develop a highly aligned geometric understanding. We demonstrate this by training visual encoders on distinct datasets each capturing a different life, both indoors and outdoors, as well as introducing a novel cross-attention-based metric to quantify the functional alignment of the internal representations developed by different models. Second, we show that single-life models learn generalizable geometric representations that effectively transfer to downstream tasks, such as depth estimation, in unseen environments. Third, we demonstrate that training on up to 30 hours from one week of the same person's life leads to comparable performance to training on 30 hours of diverse web data, highlighting the strength of single-life representation learning. Overall, our results establish that the shared structure of the world, both leads to consistency in models trained on individual lives, and provides a powerful signal for visual representation learning.

---

### SimFlow: Simplified and End-to-End Training of Latent Normalizing Flows

Authors: Qinyu Zhao, Guangting Zheng, Tao Yang, Rui Zhu, Xingjian Leng, Stephen Gould, Liang Zheng
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04084v1

Normalizing Flows (NFs) learn invertible mappings between the data and a Gaussian distribution. Prior works usually suffer from two limitations. First, they add random noise to training samples or VAE latents as data augmentation, introducing complex pipelines including extra noising and denoising steps. Second, they use a pretrained and frozen VAE encoder, resulting in suboptimal reconstruction and generation quality. In this paper, we find that the two issues can be solved in a very simple way: just fixing the variance (which would otherwise be predicted by the VAE encoder) to a constant (e.g., 0.5). On the one hand, this method allows the encoder to output a broader distribution of tokens and the decoder to learn to reconstruct clean images from the augmented token distribution, avoiding additional noise or denoising design. On the other hand, fixed variance simplifies the VAE evidence lower bound, making it stable to train an NF with a VAE jointly. On the ImageNet $256 \times 256$ generation task, our model SimFlow obtains a gFID score of 2.15, outperforming the state-of-the-art method STARFlow (gFID 2.40). Moreover, SimFlow can be seamlessly integrated with the end-to-end representation alignment (REPA-E) method and achieves an improved gFID of 1.91, setting a new state of the art among NFs.

---

### PosterCopilot: Toward Layout Reasoning and Controllable Editing for Professional Graphic Design

Authors: Jiazhe Wei, Ken Li, Tianyu Lao, Haofan Wang, Liang Wang, Caifeng Shan, Chenyang Si
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04082v1

Graphic design forms the cornerstone of modern visual communication, serving as a vital medium for promoting cultural and commercial events. Recent advances have explored automating this process using Large Multimodal Models (LMMs), yet existing methods often produce geometrically inaccurate layouts and lack the iterative, layer-specific editing required in professional workflows. To address these limitations, we present PosterCopilot, a framework that advances layout reasoning and controllable editing for professional graphic design. Specifically, we introduce a progressive three-stage training strategy that equips LMMs with geometric understanding and aesthetic reasoning for layout design, consisting of Perturbed Supervised Fine-Tuning, Reinforcement Learning for Visual-Reality Alignment, and Reinforcement Learning from Aesthetic Feedback. Furthermore, we develop a complete workflow that couples the trained LMM-based design model with generative models, enabling layer-controllable, iterative editing for precise element refinement while maintaining global visual consistency. Extensive experiments demonstrate that PosterCopilot achieves geometrically accurate and aesthetically superior layouts, offering unprecedented controllability for professional iterative design.

---

### Radiance Meshes for Volumetric Reconstruction

Authors: Alexander Mai, Trevor Hedstrom, George Kopanas, Janne Kontkanen, Falko Kuester, Jonathan T. Barron
Published: 2025-12-03
Categories: cs.GR, cs.CV
arXiv: https://arxiv.org/abs/2512.04076v1

We introduce radiance meshes, a technique for representing radiance fields with constant density tetrahedral cells produced with a Delaunay tetrahedralization. Unlike a Voronoi diagram, a Delaunay tetrahedralization yields simple triangles that are natively supported by existing hardware. As such, our model is able to perform exact and fast volume rendering using both rasterization and ray-tracing. We introduce a new rasterization method that achieves faster rendering speeds than all prior radiance field representations (assuming an equivalent number of primitives and resolution) across a variety of platforms. Optimizing the positions of Delaunay vertices introduces topological discontinuities (edge flips). To solve this, we use a Zip-NeRF-style backbone which allows us to express a smoothly varying field even when the topology changes. Our rendering method exactly evaluates the volume rendering equation and enables high quality, real-time view synthesis on standard consumer hardware. Our tetrahedral meshes also lend themselves to a variety of exciting applications including fisheye lens distortion, physics-based simulation, editing, and mesh extraction.

---

### SkillFactory: Self-Distillation For Learning Cognitive Behaviors

Authors: Zayne Sprague, Jack Lu, Manya Wadhwa, Sedrick Keh, Mengye Ren, Greg Durrett
Published: 2025-12-03
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2512.04072v2

Reasoning models leveraging long chains of thought employ various cognitive skills, such as verification of their answers, backtracking, retrying by an alternate method, and more. Previous work has shown that when a base language model exhibits these skills, training that model further with reinforcement learning (RL) can learn to leverage them. How can we get models to leverage skills that aren't exhibited by base models? Our work, SkillFactory, is a method for fine-tuning models to roughly learn these skills during a supervised fine-tuning (SFT) stage prior to RL. Our approach does not rely on distillation from a stronger model, but instead uses samples from the model itself, rearranged to provide training data in the format of those skills. These "silver" SFT traces may be imperfect, but are nevertheless effective for priming a model to acquire skills during RL. Our evaluation shows that (1) starting from SkillFactory SFT initialization helps a model to generalize to harder variants of a task post-RL, despite lower performance pre-RL;(2) cognitive skills are indeed used by the model; (3) RLed SkillFactory models are more robust to regression on out-of-domain tasks than RLed base models. Our work suggests that inductive biases learned prior to RL help models learn robust cognitive skill use.

---

### SkillFactory: Self-Distillation For Learning Cognitive Behaviors

Authors: Zayne Sprague, Jack Lu, Manya Wadhwa, Sedrick Keh, Mengye Ren, Greg Durrett
Published: 2025-12-03
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2512.04072v1

Reasoning models leveraging long chains of thought employ various cognitive skills, such as verification of their answers, backtracking, retrying by an alternate method, and more. Previous work has shown that when a base language model exhibits these skills, training that model further with reinforcement learning (RL) can learn to leverage them. How can we get models to leverage skills that aren't exhibited by base models? Our work, SkillFactory, is a method for fine-tuning models to roughly learn these skills during a supervised fine-tuning (SFT) stage prior to RL. Our approach does not rely on distillation from a stronger model, but instead uses samples from the model itself, rearranged to provide training data in the format of those skills. These "silver" SFT traces may be imperfect, but are nevertheless effective for priming a model to acquire skills during RL. Our evaluation shows that (1) starting from SkillFactory SFT initialization helps a model to generalize to harder variants of a task post-RL, despite lower performance pre-RL; (2) cognitive skills are indeed used by the model; (3) RLed SkillFactory models are more robust to regression on out-of-domain tasks than RLed base models. Our work suggests that inductive biases learned prior to RL help models learn robust cognitive skill use.

---

### SpaceTools: Tool-Augmented Spatial Reasoning via Double Interactive RL

Authors: Siyi Chen, Mikaela Angelina Uy, Chan Hee Song, Faisal Ladhak, Adithyavairavan Murali, Qing Qu, Stan Birchfield, Valts Blukis, Jonathan Tremblay
Published: 2025-12-03
Categories: cs.CV, cs.RO
arXiv: https://arxiv.org/abs/2512.04069v2

Vision Language Models (VLMs) demonstrate strong qualitative visual understanding, but struggle with metrically precise spatial reasoning required for embodied applications. The agentic paradigm promises that VLMs can use a wide variety of tools that could augment these capabilities, such as depth estimators, segmentation models, and pose estimators. Yet it remains an open challenge how to realize this vision without solely relying on handcrafted prompting strategies or enforcing fixed, predefined tool pipelines that limit VLMs' ability to discover optimal tool-use patterns. Reinforcement Learning could overcome this gap, but has so far been limited to reasoning with a single visual tool due to the large search space in multi-tool reasoning. We introduce Double Interactive Reinforcement Learning (DIRL), a two-phase training framework where VLMs learn to coordinate multiple tools through interactive exploration and feedback. In the teaching phase, we combine demonstrations from a single tool specialist trained via interactive RL with traces from a frontier model using all tools. In the exploration phase, the model further refines multi-tool coordination through continued RL. Our model, SpaceTools, with tool-augmented spatial reasoning ability, achieves state-of-the-art performance on spatial understanding benchmarks (RoboSpatial-Home, BLINK, BOP-ASK) and demonstrates reliable real-world manipulation using a 7-DOF robot as a tool. DIRL provides substantial improvements over the vanilla SFT (+12% on RoboSpatial) and RL (+16% on RoboSpatial) baselines. Project page: https://spacetools.github.io/.

---

### SpaceTools: Tool-Augmented Spatial Reasoning via Double Interactive RL

Authors: Siyi Chen, Mikaela Angelina Uy, Chan Hee Song, Faisal Ladhak, Adithyavairavan Murali, Qing Qu, Stan Birchfield, Valts Blukis, Jonathan Tremblay
Published: 2025-12-03
Categories: cs.CV, cs.RO
arXiv: https://arxiv.org/abs/2512.04069v1

Vision Language Models (VLMs) demonstrate strong qualitative visual understanding, but struggle with metrically precise spatial reasoning required for embodied applications. The agentic paradigm promises that VLMs can use a wide variety of tools that could augment these capabilities, such as depth estimators, segmentation models, and pose estimators. Yet it remains an open challenge how to realize this vision without solely relying on handcrafted prompting strategies or enforcing fixed, predefined tool pipelines that limit VLMs' ability to discover optimal tool-use patterns. Reinforcement Learning could overcome this gap, but has so far been limited to reasoning with a single visual tool due to the large search space in multi-tool reasoning. We introduce Double Interactive Reinforcement Learning (DIRL), a two-phase training framework where VLMs learn to coordinate multiple tools through interactive exploration and feedback. In the teaching phase, we combine demonstrations from a single tool specialist trained via interactive RL with traces from a frontier model using all tools. In the exploration phase, the model further refines multi-tool coordination through continued RL. Our model, SpaceTools, with tool-augmented spatial reasoning ability, achieves state-of-the-art performance on spatial understanding benchmarks (RoboSpatial-Home, BLINK, BOP-ASK) and demonstrates reliable real-world manipulation using a 7-DOF robot as a tool. DIRL provides substantial improvements over the vanilla SFT (+12% on RoboSpatial) and RL (+16% on RoboSpatial) baselines. Project page: https://spacetools.github.io/.

---

### Learning Steerable Clarification Policies with Collaborative Self-play

Authors: Jonathan Berant, Maximillian Chen, Adam Fisch, Reza Aghajani, Fantine Huot, Mirella Lapata, Jacob Eisenstein
Published: 2025-12-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04068v3

To handle underspecified or ambiguous queries, AI assistants need a policy for managing their uncertainty to determine (a) when to guess the user intent and answer directly, (b) when to enumerate and answer multiple possible intents, and (c) when to ask a clarifying question. However, such policies are contextually dependent on factors such as user preferences or modality. For example, enumerating multiple possible user intentions is cumbersome on small screens or in a voice setting. In this work, we propose to train steerable policies for managing this uncertainty using self-play. Given two agents, one simulating a user and the other an AI assistant, we generate conversations where the user issues a potentially ambiguous query, and the assistant needs to determine how to respond. Importantly, the model takes as input the numerical cost of each clarification question, and each generated word, and is asked to take the action that will maximize its final reward, which is the cost-penalized accuracy. We use Reinforced Self-Training (ReST) to train our model to achieve high reward and show this leads to a steerable policy that changes its behavior predictably conditioned on the provided costs, leading to higher reward and accuracy. Moreover, our procedure also generalizes to numerical cost values that were unobserved at training time.

---

### Learning Steerable Clarification Policies with Collaborative Self-play

Authors: Jonathan Berant, Maximillian Chen, Adam Fisch, Reza Aghajani, Fantine Huot, Mirella Lapata, Jacob Eisenstein
Published: 2025-12-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04068v1

To handle underspecified or ambiguous queries, AI assistants need a policy for managing their uncertainty to determine (a) when to guess the user intent and answer directly, (b) when to enumerate and answer multiple possible intents, and (c) when to ask a clarifying question. However, such policies are contextually dependent on factors such as user preferences or modality. For example, enumerating multiple possible user intentions is cumbersome on small screens or in a voice setting. In this work, we propose to train steerable policies for managing this uncertainty using self-play. Given two agents, one simulating a user and the other an AI assistant, we generate conversations where the user issues a potentially ambiguous query, and the assistant needs to determine how to respond. Importantly, the model takes as input the numerical cost of each clarification question, and each generated word, and is asked to take the action that will maximize its final reward, which is the cost-penalized accuracy. We use Reinforced Self-Training (ReST) to train our model to achieve high reward and show this leads to a steerable policy that changes its behavior predictably conditioned on the provided costs, leading to higher reward and accuracy. Moreover, our procedure also generalizes to numerical cost values that were unobserved at training time.

---

### Fare Comparison App of Uber, Ola and Rapido

Authors: Ashlesha Gopinath Sawant, Sahil S. Jadhav, Vidhan R. Jain, Shriraj S. Jagtap, Prachi Jadhav, Soham Jadhav, Ichha Raina
Published: 2025-12-03
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2512.04065v1

In todays increasing world, it is very important to have good hailing services like Ola, Uber, and Rapido as it is very essential for our daily transportation. Users often face difficulties in choosing the most appropriate and efficient ride that would lead to both cost-effective and would take us to our destination in less time. This project provides you with the web application that helps you to select the most beneficial ride for you by providing users with the fare comparison between Ola, Uber, Rapido for the destination entered by the user. The backend is use to fetch the data, providing users with the fare comparison for the ride and finally providing with the best option using Python. This research paper also addresses the problem and challenges faced in accessing the data using APIs, Android Studios emulator, Appium and location comparison. Thus, the aim of the project is to provide transparency to the users in ride-hailing services and increase efficiency and provide users with better experience.

---

### Eval Factsheets: A Structured Framework for Documenting AI Evaluations

Authors: Florian Bordes, Candace Ross, Justine T Kao, Evangelia Spiliopoulou, Adina Williams
Published: 2025-12-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04062v1

The rapid proliferation of benchmarks has created significant challenges in reproducibility, transparency, and informed decision-making. However, unlike datasets and models -- which benefit from structured documentation frameworks like Datasheets and Model Cards -- evaluation methodologies lack systematic documentation standards. We introduce Eval Factsheets, a structured, descriptive framework for documenting AI system evaluations through a comprehensive taxonomy and questionnaire-based approach. Our framework organizes evaluation characteristics across five fundamental dimensions: Context (Who made the evaluation and when?), Scope (What does it evaluate?), Structure (With what the evaluation is built?), Method (How does it work?) and Alignment (In what ways is it reliable/valid/robust?). We implement this taxonomy as a practical questionnaire spanning five sections with mandatory and recommended documentation elements. Through case studies on multiple benchmarks, we demonstrate that Eval Factsheets effectively captures diverse evaluation paradigms -- from traditional benchmarks to LLM-as-judge methodologies -- while maintaining consistency and comparability. We hope Eval Factsheets are incorporated into both existing and newly released evaluation frameworks and lead to more transparency and reproducibility.

---

### Closing the problem of which causal structures of up to six total nodes have a classical-quantum gap

Authors: Shashaank Khanna, Matthew Pusey, Roger Colbeck
Published: 2025-12-03
Categories: quant-ph, cs.LG, math.ST
arXiv: https://arxiv.org/abs/2512.04058v1

The discovery of Bell that there exist quantum correlations that cannot be reproduced classically is one of the most important in the foundations of quantum mechanics, as well as having practical implications. Bell's result was originally proven in a simple bipartite causal structure, but analogous results have also been shown in further causal structures. Here we study the only causal structure with six or fewer nodes in which the question of whether or not there exist quantum correlations that cannot be achieved classically was open. In this causal structure we show that such quantum correlations exist using a method that involves imposing additional restrictions on the correlations. This hence completes the picture of which causal structures of up to six nodes support non-classical quantum correlations. We also provide further illustrations of our method using other causal structures.

---

### Convergence for Discrete Parameter Update Schemes

Authors: Paul Wilson, Fabio Zanasi, George Constantinides
Published: 2025-12-03
Categories: cs.LG, math.OC
arXiv: https://arxiv.org/abs/2512.04051v2

Modern deep learning models require immense computational resources, motivating research into low-precision training. Quantised training addresses this by representing training components in low-bit integers, but typically relies on discretising real-valued updates. We introduce an alternative approach where the update rule itself is discrete, avoiding the quantisation of continuous updates by design. We establish convergence guarantees for a general class of such discrete schemes, and present a multinomial update rule as a concrete example, supported by empirical evaluation. This perspective opens new avenues for efficient training, particularly for models with inherently discrete structure.

---

### Convergence for Discrete Parameter Updates

Authors: Paul Wilson, Fabio Zanasi, George Constantinides
Published: 2025-12-03
Categories: cs.LG, math.OC
arXiv: https://arxiv.org/abs/2512.04051v1

Modern deep learning models require immense computational resources, motivating research into low-precision training. Quantised training addresses this by representing training components in low-bit integers, but typically relies on discretising real-valued updates. We introduce an alternative approach where the update rule itself is discrete, avoiding the quantisation of continuous updates by design. We establish convergence guarantees for a general class of such discrete schemes, and present a multinomial update rule as a concrete example, supported by empirical evaluation. This perspective opens new avenues for efficient training, particularly for models with inherently discrete structure.

---

### Stable Signer: Hierarchical Sign Language Generative Model

Authors: Sen Fang, Yalin Feng, Hongbin Zhong, Yanxin Zhang, Dimitris N. Metaxas
Published: 2025-12-03
Categories: cs.CV, cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2512.04048v2

Sign Language Production (SLP) is the process of converting the complex input text into a real video. Most previous works focused on the Text2Gloss, Gloss2Pose, Pose2Vid stages, and some concentrated on Prompt2Gloss and Text2Avatar stages. However, this field has made slow progress due to the inaccuracy of text conversion, pose generation, and the rendering of poses into real human videos in these stages, resulting in gradually accumulating errors. Therefore, in this paper, we streamline the traditional redundant structure, simplify and optimize the task objective, and design a new sign language generative model called Stable Signer. It redefines the SLP task as a hierarchical generation end-to-end task that only includes text understanding (Prompt2Gloss, Text2Gloss) and Pose2Vid, and executes text understanding through our proposed new Sign Language Understanding Linker called SLUL, and generates hand gestures through the named SLP-MoE hand gesture rendering expert block to end-to-end generate high-quality and multi-style sign language videos. SLUL is trained using the newly developed Semantic-Aware Gloss Masking Loss (SAGM Loss). Its performance has improved by 48.6% compared to the current SOTA generation methods.

---

### Stable Signer: Hierarchical Sign Language Generative Model

Authors: Sen Fang, Yalin Feng, Hongbin Zhong, Yanxin Zhang, Dimitris N. Metaxas
Published: 2025-12-03
Categories: cs.CV, cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2512.04048v1

Sign Language Production (SLP) is the process of converting the complex input text into a real video. Most previous works focused on the Text2Gloss, Gloss2Pose, Pose2Vid stages, and some concentrated on Prompt2Gloss and Text2Avatar stages. However, this field has made slow progress due to the inaccuracy of text conversion, pose generation, and the rendering of poses into real human videos in these stages, resulting in gradually accumulating errors. Therefore, in this paper, we streamline the traditional redundant structure, simplify and optimize the task objective, and design a new sign language generative model called Stable Signer. It redefines the SLP task as a hierarchical generation end-to-end task that only includes text understanding (Prompt2Gloss, Text2Gloss) and Pose2Vid, and executes text understanding through our proposed new Sign Language Understanding Linker called SLUL, and generates hand gestures through the named SLP-MoE hand gesture rendering expert block to end-to-end generate high-quality and multi-style sign language videos. SLUL is trained using the newly developed Semantic-Aware Gloss Masking Loss (SAGM Loss). Its performance has improved by 48.6% compared to the current SOTA generation methods.

---

### Polarization by Design: How Elites Could Shape Mass Preferences as AI Reduces Persuasion Costs

Authors: Nadav Kunievsky
Published: 2025-12-03
Categories: econ.GN, cs.AI, cs.CY
arXiv: https://arxiv.org/abs/2512.04047v1

In democracies, major policy decisions typically require some form of majority or consensus, so elites must secure mass support to govern. Historically, elites could shape support only through limited instruments like schooling and mass media; advances in AI-driven persuasion sharply reduce the cost and increase the precision of shaping public opinion, making the distribution of preferences itself an object of deliberate design. We develop a dynamic model in which elites choose how much to reshape the distribution of policy preferences, subject to persuasion costs and a majority rule constraint. With a single elite, any optimal intervention tends to push society toward more polarized opinion profiles - a ``polarization pull'' - and improvements in persuasion technology accelerate this drift. When two opposed elites alternate in power, the same technology also creates incentives to park society in ``semi-lock'' regions where opinions are more cohesive and harder for a rival to overturn, so advances in persuasion can either heighten or dampen polarization depending on the environment. Taken together, cheaper persuasion technologies recast polarization as a strategic instrument of governance rather than a purely emergent social byproduct, with important implications for democratic stability as AI capabilities advance.

---

### MarkTune: Improving the Quality-Detectability Trade-off in Open-Weight LLM Watermarking

Authors: Yizhou Zhao, Zhiwei Steven Wu, Adam Block
Published: 2025-12-03
Categories: cs.LG, cs.AI, cs.CR
arXiv: https://arxiv.org/abs/2512.04044v1

Watermarking aims to embed hidden signals in generated text that can be reliably detected when given access to a secret key. Open-weight language models pose acute challenges for such watermarking schemes because the inference-time interventions that dominate contemporary approaches cannot be enforced once model weights are public. Existing watermaking techniques for open-weight models, such as the recently proposed GaussMark, typically rely on small modifications to model weights, which can yield signals detectable to those equipped with a secret key, but achieving detection power comparable to inference-time watermarks generally requires weight perturbations that noticeably reduce generation quality. We introduce MarkTune, a theoretically principled, on-policy fine-tuning framework that treats the GaussMark signal as a reward while simultaneously regularizing against degradation in text quality. We derive MarkTune as an improvement on GaussMark and demonstrate that MarkTune consistently improves the quality-detectability trade-off over GaussMark by steering finer-grained, watermark-aware weight updates within the model's representation space while preserving generation quality. Empirically, we show that MarkTune pushes the quality-detectability frontier of GaussMark close to that of inference-time watermarking, remains robust to paraphrasing and fine-tuning attacks, and exhibits strong generalization: a model fine-tuned on one dataset retains substantial watermark detection power on unseen datasets. Together, these results establish MarkTune as a general strategy for embedding robust, high-quality watermarks into open-weight LMs.

---

### RELIC: Interactive Video World Model with Long-Horizon Memory

Authors: Yicong Hong, Yiqun Mei, Chongjian Ge, Yiran Xu, Yang Zhou, Sai Bi, Yannick Hold-Geoffroy, Mike Roberts, Matthew Fisher, Eli Shechtman, Kalyan Sunkavalli, Feng Liu et al.
Published: 2025-12-03
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04040v1

A truly interactive world model requires three key ingredients: real-time long-horizon streaming, consistent spatial memory, and precise user control. However, most existing approaches address only one of these aspects in isolation, as achieving all three simultaneously is highly challenging-for example, long-term memory mechanisms often degrade real-time performance. In this work, we present RELIC, a unified framework that tackles these three challenges altogether. Given a single image and a text description, RELIC enables memory-aware, long-duration exploration of arbitrary scenes in real time. Built upon recent autoregressive video-diffusion distillation techniques, our model represents long-horizon memory using highly compressed historical latent tokens encoded with both relative actions and absolute camera poses within the KV cache. This compact, camera-aware memory structure supports implicit 3D-consistent content retrieval and enforces long-term coherence with minimal computational overhead. In parallel, we fine-tune a bidirectional teacher video model to generate sequences beyond its original 5-second training horizon, and transform it into a causal student generator using a new memory-efficient self-forcing paradigm that enables full-context distillation over long-duration teacher as well as long student self-rollouts. Implemented as a 14B-parameter model and trained on a curated Unreal Engine-rendered dataset, RELIC achieves real-time generation at 16 FPS while demonstrating more accurate action following, more stable long-horizon streaming, and more robust spatial-memory retrieval compared with prior work. These capabilities establish RELIC as a strong foundation for the next generation of interactive world modeling.

---

### Fast & Efficient Normalizing Flows and Applications of Image Generative Models

Authors: Sandeep Nagar
Published: 2025-12-03
Categories: cs.CV, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.04039v1

This thesis presents novel contributions in two primary areas: advancing the efficiency of generative models, particularly normalizing flows, and applying generative models to solve real-world computer vision challenges. The first part introduce significant improvements to normalizing flow architectures through six key innovations: 1) Development of invertible 3x3 Convolution layers with mathematically proven necessary and sufficient conditions for invertibility, (2) introduction of a more efficient Quad-coupling layer, 3) Design of a fast and efficient parallel inversion algorithm for kxk convolutional layers, 4) Fast & efficient backpropagation algorithm for inverse of convolution, 5) Using inverse of convolution, in Inverse-Flow, for the forward pass and training it using proposed backpropagation algorithm, and 6) Affine-StableSR, a compact and efficient super-resolution model that leverages pre-trained weights and Normalizing Flow layers to reduce parameter count while maintaining performance. The second part: 1) An automated quality assessment system for agricultural produce using Conditional GANs to address class imbalance, data scarcity and annotation challenges, achieving good accuracy in seed purity testing; 2) An unsupervised geological mapping framework utilizing stacked autoencoders for dimensionality reduction, showing improved feature extraction compared to conventional methods; 3) We proposed a privacy preserving method for autonomous driving datasets using on face detection and image inpainting; 4) Utilizing Stable Diffusion based image inpainting for replacing the detected face and license plate to advancing privacy-preserving techniques and ethical considerations in the field.; and 5) An adapted diffusion model for art restoration that effectively handles multiple types of degradation through unified fine-tuning.

---

### Domain Feature Collapse: Implications for Out-of-Distribution Detection and Solutions

Authors: Hong Yang, Devroop Kar, Qi Yu, Alex Ororbia, Travis Desell
Published: 2025-12-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04034v2

Why do state-of-the-art OOD detection methods exhibit catastrophic failure when models are trained on single-domain datasets? We provide the first theoretical explanation for this phenomenon through the lens of information theory. We prove that supervised learning on single-domain data inevitably produces domain feature collapse -- representations where I(x_d; z) = 0, meaning domain-specific information is completely discarded. This is a fundamental consequence of information bottleneck optimization: models trained on single domains (e.g., medical images) learn to rely solely on class-specific features while discarding domain features, leading to catastrophic failure when detecting out-of-domain samples (e.g., achieving only 53% FPR@95 on MNIST). We extend our analysis using Fano's inequality to quantify partial collapse in practical scenarios. To validate our theory, we introduce Domain Bench, a benchmark of single-domain datasets, and demonstrate that preserving I(x_d; z) > 0 through domain filtering (using pretrained representations) resolves the failure mode. While domain filtering itself is conceptually straightforward, its effectiveness provides strong empirical evidence for our information-theoretic framework. Our work explains a puzzling empirical phenomenon, reveals fundamental limitations of supervised learning in narrow domains, and has broader implications for transfer learning and when to fine-tune versus freeze pretrained models.

---

### Domain Feature Collapse: Implications for Out-of-Distribution Detection and Solutions

Authors: Hong Yang, Devroop Kar, Qi Yu, Alex Ororbia, Travis Desell
Published: 2025-12-03
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04034v1

Why do state-of-the-art OOD detection methods exhibit catastrophic failure when models are trained on single-domain datasets? We provide the first theoretical explanation for this phenomenon through the lens of information theory. We prove that supervised learning on single-domain data inevitably produces domain feature collapse -- representations where I(x_d; z) = 0, meaning domain-specific information is completely discarded. This is a fundamental consequence of information bottleneck optimization: models trained on single domains (e.g., medical images) learn to rely solely on class-specific features while discarding domain features, leading to catastrophic failure when detecting out-of-domain samples (e.g., achieving only 53% FPR@95 on MNIST). We extend our analysis using Fano's inequality to quantify partial collapse in practical scenarios. To validate our theory, we introduce Domain Bench, a benchmark of single-domain datasets, and demonstrate that preserving I(x_d; z) > 0 through domain filtering (using pretrained representations) resolves the failure mode. While domain filtering itself is conceptually straightforward, its effectiveness provides strong empirical evidence for our information-theoretic framework. Our work explains a puzzling empirical phenomenon, reveals fundamental limitations of supervised learning in narrow domains, and has broader implications for transfer learning and when to fine-tune versus freeze pretrained models.

---

### jina-vlm: Small Multilingual Vision Language Model

Authors: Andreas Koukounas, Georgios Mastrapas, Florian Hönicke, Sedigheh Eslami, Guillaume Roncari, Han Xiao
Published: 2025-12-03
Categories: cs.CL, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2512.04032v4

We present jina-vlm, a token-efficient 2.4B parameter vision-language model that achieves state-of-the-art multilingual VQA performance among open 2B-scale VLMs. The model couples a SigLIP2 vision encoder with a Qwen3 language decoder and makes use of image tiling and attention-pooling for token-efficient processing of arbitrary-resolution images. To understand the contribution of different training data categories, we conduct a leave-one-out data mixture ablation study-systematically removing task, domain, modality, and language categories-to diagnose which data types are necessary versus redundant and whether task benefits transfer across domains. Model weights and code are publicly released at https://huggingface.co/jinaai/jina-vlm.

---

### Jina-VLM: Small Multilingual Vision Language Model

Authors: Andreas Koukounas, Georgios Mastrapas, Florian Hönicke, Sedigheh Eslami, Guillaume Roncari, Scott Martens, Han Xiao
Published: 2025-12-03
Categories: cs.CL, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2512.04032v1

We present Jina-VLM, a 2.4B parameter vision-language model that achieves state-of-the-art multilingual visual question answering among open 2B-scale VLMs. The model couples a SigLIP2 vision encoder with a Qwen3 language backbone through an attention-pooling connector that enables token-efficient processing of arbitrary-resolution images. Across standard VQA benchmarks and multilingual evaluations, Jina-VLM outperforms comparable models while preserving competitive text-only performance.

---

### Large Language Models for Limited Noisy Data: A Gravitational Wave Identification Study

Authors: Yixuan Li, Yuhao Lu, Yang Liu, Liang Li, R. Ruffini, Di Li, Rong-Gen Cai, Xiaoyan Zhu, Wenbin Lin, Yu Wang
Published: 2025-12-03
Categories: astro-ph.IM, astro-ph.HE, cs.AI
arXiv: https://arxiv.org/abs/2512.04031v2

This work investigates whether large language models (LLMs) offer advantages over traditional neural networks for astronomical data processing, in regimes with non-Gaussian, non-stationary noise and limited labeled samples. Gravitational wave observations provide an suitable test case, using only 90 LIGO events, finetuned LLMs achieve 97.4\% accuracy for identifying signals. Further experiments show that, in contrast to traditional networks that rely on large simulated datasets, additional simulated samples do not improve LLM performance, while scaling studies reveal predictable gains with increasing model size and dataset size. These results indicate that LLMs can extract discriminative structure directly from observational data and provide an efficient assessment for gravitational wave identification. The same strategy may extend to other astronomical domains with similar noise properties, such as radio or pulsar observations.

---