The proposed PARA-PV framework integrates physical knowledge with frozen foundation models to improve the accuracy of solar power forecasting.
A new study examines the impact of activation-based persona steering on short-answer generation and automated grading across three language models.
This paper presents a structured pruning method for large language models that adapts unstructured techniques to maintain optimization direction and feature retention.
This study simulates a multi-agent marketplace to identify the formal communication and structural mechanisms required to maintain market stability among self-interested AI agents.
Hugging Face has published an analysis focusing on the role and preparation of training data specifically optimized for artificial intelligence agents.
Researchers propose a new evaluation standard to address long-term risks and engagement-driven biases in healthcare-focused large language models.
This paper argues that current AI evaluation frameworks should expand beyond technical metrics to include psychological competence for models interacting directly with humans.
A new framework aims to improve the explainability of temporal graph networks by analyzing how historical events stored in memory modules influence model predictions.
A new graph-regularized learning framework incorporates psychological interdependencies among emotion classes to improve EEG-based emotion recognition.
A comprehensive survey evaluates the progress of large language models in medical reasoning by mapping clinical competencies to computational methods.
ProsMAE is a multi-source masked autoencoder framework designed to learn robust representations from gigapixel pathology slides for prostate cancer grading.
A new image enhancement framework called NamedCurves+ utilizes color naming concepts to make deep-learning-based photo editing more interpretable and adjustable for users.
The RhyMix network is a lightweight model designed to capture diverse temporal patterns simultaneously for more accurate long-term time series forecasting.
Researchers propose using a live relational data structure called a Context Graph to enable enterprise AI agents to proactively share information before being asked.
The proposed ADORN framework utilizes Q-learning to optimize retraining schedules for machine learning models facing performance drift in Open Radio Access Networks.
This paper explores the analytical potential of utilizing UMAP's internal k-nearest-neighbor graph to analyze high-dimensional data before projection distortion occurs.
This paper argues that causal abstraction theory provides a valuable framework for explaining how neural networks implement specific computational processes.
The paper proposes a robust weighted triangulation method to improve causal effect estimation when dealing with model uncertainty in observational data.
This paper introduces second-order actor-critic methods utilizing policy Hessian decomposition to accelerate convergence in discounted Markov decision processes.
This work localizes the internal subgraph responsible for representing and resolving temporal preferences and trade-offs within a distilled large language model.
This paper demonstrates that optimal generalization in offline reinforcement learning depends on the structure of the pessimistic bias rather than the absolute level of conservatism.
Researchers have introduced MambaLIE, a state space model-based framework designed to enhance low-light images by leveraging scene light intensity.
The study introduces adversarial decoys, a method designed to bypass attention-based defenses in Vision Transformers by misdirecting token suppression mechanisms.
Researchers have introduced the first provably efficient learning algorithms for repeated assistance games to optimize shared reward functions between human and AI agents.