We present an uncertainty-aware online ensemble method for handling non-stationary data streams in fusion science applications. The approach dynamically weights multiple models …
This paper addresses the challenge of continual learning in particle accelerators, where machine learning models must adapt to changing data distributions and system configurations …
Hydra is an advanced framework designed for training and managing AI models for near real time data quality monitoring at Jefferson Lab. Deployed in all four experimental halls, …
This paper addresses the challenge of efficiently representing compute cluster workloads using deep learning. We develop models that capture the essential characteristics of …
This paper presents results from ongoing efforts to develop continual learning methods for particle accelerators. We address the challenge of model adaptation to changing …
This paper presents results from the AI4EIC Hackathon, specifically focusing on particle identification (PID) with the ePIC dRICH detector. Machine learning techniques were applied …
Deep learning models for particle accelerator control and optimization can suffer from overconfidence when deployed to new operating conditions. This paper presents techniques for …
This paper presents a probabilistic envelope-based visualization technique for monitoring drilling well data logging. The method enables effective identification of anomalies and …