4165 papers
This paper identifies a coherent rank-one mean bias as the primary cause of numerical instability in low-bit LLM training and demonstrates that simply subtracting this mean restores stability and performance in FP4 quantization, offering a hardware-efficient alternative to complex spectral methods.
FAR-Dex is a hierarchical framework that combines few-shot data augmentation via the IsaacLab simulator with an adaptive residual policy refinement module to overcome data scarcity and high-dimensional action space challenges, achieving robust and precise dexterous arm-hand coordination with over 80% success in real-world tasks.
UniPINN is a unified Physics-Informed Neural Network framework that addresses the challenges of multi-task learning for diverse Navier-Stokes equations by integrating a shared-specialized architecture, cross-flow attention, and dynamic weight allocation to prevent negative transfer and stabilize training across heterogeneous flow regimes.
The paper proposes G-STAR, an end-to-end system that integrates a time-aware speaker-tracking module with a Speech-LLM backbone to achieve robust, timestamped speaker-attributed recognition for long-form, overlapping multi-party speech while maintaining global identity consistency.
This paper proposes a unified framework for news recommendation that addresses the time-sensitive nature of user interests by combining global collaborative signals for long-term preferences with a stage-wise temporal subgraph approach, enhanced by LSTM and self-attention mechanisms, to effectively model both stable habits and rapidly evolving short-term dynamics.
This paper introduces SearchLLM, the first large language model designed for open-ended generative search on platforms like RedNote, which utilizes a hierarchical multi-dimensional reward system and Gated Aggregation Strategy with GRPO to balance safety, factual grounding, and user alignment, resulting in measurable improvements in generation quality and user engagement.
This paper introduces a multi-agent negotiation framework that trains large language models to align with Collective Agency and resolve value conflicts through self-play deliberation optimized via RLAIF with GRPO, demonstrating improved conflict-resolution capabilities without compromising general language performance.
The paper introduces JEDI, a hierarchical model that jointly learns shared embeddings over recurrent neural network weights to robustly infer generalizable, task-specific neural dynamics from noisy, high-dimensional experimental recordings, successfully recovering underlying mechanistic structures and providing insights into motor control.
This paper demonstrates that the naive exposure of powerful reasoning and image refinement capabilities in commercial generative AI chatbots fundamentally undermines state-of-the-art deepfake detectors by allowing adversaries to use benign, policy-compliant prompts to generate high-quality, identity-preserving images that evade detection, revealing a critical structural mismatch between current threat models and real-world AI capabilities.
This paper proposes a lightweight hybrid framework for the Game of the Amazons that integrates Graph Attention Autoencoders, Stochastic Graph Genetic Algorithms, and GPT-4o-mini to overcome resource constraints, achieving decision accuracy improvements of 15%–56% over baselines and outperforming its teacher model by effectively denoising LLM outputs through structural graph reasoning.
The paper introduces IH-Challenge, a reinforcement learning dataset designed to enhance instruction hierarchy robustness in frontier LLMs, which significantly improves their ability to prioritize instructions against conflicts and adversarial attacks while maintaining helpfulness and minimizing capability regression.
This paper presents a Multi-Agent Reinforcement Learning framework using Proximal Policy Optimization to coordinate UAV fleets for time-critical medical supply delivery, demonstrating that classical PPO outperforms asynchronous and sequential strategies in dynamically prioritizing tasks and reallocating resources under uncertain conditions using real-world geographic data.
This study evaluates 11 promptable foundation models for musculoskeletal CT segmentation across four anatomical regions, revealing that while specific models like SAM and nnInteractive perform best under ideal conditions, all models exhibit significant sensitivity to human prompting variations, leading to performance drops and highlighting the challenge of selecting robust models for real-world clinical applications.
The paper introduces SCORE, a lightweight deep learning architecture that replaces traditional layer stacking with a weight-shared, ODE-inspired contractive recurrent update mechanism to improve training stability, accelerate convergence, and reduce parameter counts across various model types.
This paper introduces a novel language-guided framework that leverages pretrained vision-language models and a specialized adapter to achieve zero-shot, generative detection and localization of subsurface defects in carbon fiber-reinforced polymers using active infrared thermography, thereby eliminating the need for costly, task-specific training datasets while significantly improving signal-to-noise ratios and detection accuracy.
This paper proposes a novel self-finetuning framework that enables Generative AI agents to autonomously learn continuous control for dynamic Radio Access Network slicing by distilling long-horizon experiences into model parameters via a bi-perspective reflection mechanism, thereby outperforming traditional Reinforcement Learning and standard LLM-based agents in sample efficiency and multi-objective optimization without relying on handcrafted reward signals.
This paper presents CUAAudit, a large-scale meta-evaluation demonstrating that while Vision-Language Models can serve as autonomous auditors for Computer-Use Agents with strong accuracy and calibration, their significant performance degradation in complex environments and notable inter-model disagreement reveal fundamental limitations that necessitate explicit accounting for evaluator reliability and uncertainty in real-world deployments.
This paper empirically demonstrates that contrary to the hypothesis that moral reasoning alignment requires diversity-seeking algorithms, standard reward-maximizing RLVR methods are equally or more effective because high-reward moral responses exhibit a concentrated distribution in semantic space similar to logical reasoning tasks.
This paper establishes a mathematical framework called Gradient Flow Drifting that proves the equivalence between the recently proposed Drifting Model and the Wasserstein gradient flow of the forward KL divergence under KDE approximation, while extending the approach to a mixed-divergence strategy on Riemannian manifolds to simultaneously mitigate mode collapse and blurring.
This paper proposes the Progressive Retrospective Framework (PRF), a plug-and-play method that utilizes a cascade of retrospective units and a rolling-start training strategy to effectively address the challenge of variable-length trajectory prediction in autonomous driving by progressively aligning features from incomplete observations with complete ones.