Choose timezone
Your profile timezone:
Artificial Intelligence for the Electron Ion Collider (AI4EIC) 2025
→
America/New_York
Description
The 4th AI4EIC Workshop will take place at MIT, Wong Auditorium (Tang Center), in Boston. Organized in collaboration with the AI Institute for Artificial Intelligence and Fundamental Interactions (IAIFI), the event will feature in-person participation, with live streaming available for remote attendees.
Join on Zoom: https://mit.zoom.us/j/99550594447
To join the wifi at MIT, select "MIT Guest" and then provide a phone number or email address for verification. Eduroam also works.
Live Notes of the workshop can be found at this link.
The previous three AI4EIC workshops fostered meaningful discussions on the full range of AI/ML applications for the EIC—including accelerator and detector design, theory, and analysis—resulting in published proceedings and a community paper (https://eic.ai).
Proceedings will be published in the Journal of Instrumentation. The abstract submission (1 page max, figures can be included) is now open (deadline August 17, 2025).
In this MIT workshop, we will delve deeper into the active and emerging applications of AI/ML within the EIC community, with a focus on ongoing efforts related to the ePIC experiment and beyond.
Scientific Organizing Committee
Abhay Deshpande, Stony Brook, BNL
Ben Nachman, Stanford
Cris Fanelli, W&M
David Lawrence, JLab
Malachi Schram, JLab
Marco Battaglieri, INFN
Mike Williams, MIT
Or Hen, MIT
Phiala Shanahan, MIT
Tanja Horn, CUA
Torre Wenaus, BNL
Contact
Registration
AI4EIC Virtual Registration
AI4EIC Workshop In-Person Registration Form
Participants
160
View full list
-
-
Introduction: Welcome and Introduction
-
1
Welcome and IntroductionSpeakers: Cristiano Fanelli (William & Mary), Marisa LaFleur (MIT)
-
1
-
AI/ML for Accelerators
-
2
Explainable and Differential Reinforcement Learning for Multi Objective Optimization in Particle AcceleratorsSpeaker: Kishansingh Rajput (Jefferson Lab)
-
3
Differentiable beam dynamics codes, their use in AI-ML for accelerators and potential applications to the EICSpeaker: Chenran Xu (Argonne National Laboratory)
-
4
Symplectic machine learning model for fast simulation of space-charge effectsSpeaker: Jinyu Wan (Facility for Rare Isotopes, Michigan State University)
-
11:15
Coffee Break
-
5
Use of AI/ML for higher brightness and higher polarization of hadron beams
We report on the use of AI/ML techniques to advance the pursuit of higher brightness and polarization of hadron beams in the RHIC/EIC injector chain. Bayesian inference applied to individual magnet strengths reduced quadrupole field uncertainty by a factor of two while shifting mean values away from prior expectations, thereby improving the reliability of accelerator models. Bayesian optimization has enabled automated, high-performance tuning of injection alignment and matching, reaching results comparable to expert operators but at faster timescales. Reinforcement learning agents have achieved one-shot optimization of RF voltages for bunch merging and are now being trained for the stabilization of Booster-to-AGS beam transfer under drifting machine conditions. Collectively, these developments demonstrate the capacity of AI/ML methods to deliver adaptive and precise control strategies in support of next-generation polarized hadron beams.
Speaker: Eiad Hamwi (Cornell University) -
6
Framework for the Development of Virtual Accelerator Models for Machine Learning ApplicationsSpeaker: Adwaith Ravichandran (Argonne National Laboratory)
-
7
Machine-Learning–Accelerated Bayesian Uncertainty Quantification for Digital Twin Modeling and Control of the AGS BoosterSpeaker: Christopher Kelly (Brookhaven National Laboratory)
-
8
Machine Learning Approaches to Improved Ion Profile Monitor MeasurementsSpeaker: Christopher Hall (RadiaSoft LLC)
-
9
Use of Generative AI and LLMs for Accelerator DesignSpeaker: Onur Gilanliogullari (member@anl.gov)
-
2
-
Break: Lunch
-
AI/ML for Calibration, Monitoring, and Experimental Control with Data Streams
-
10
Tree-distilled autoencoders on FPGA for anomaly detection and data compressionSpeakers: Tae Min Hong (University of Pittsburgh (US)), Tae Min Hong (University of Pittsburgh)
-
11
Compression by Importance and More
In this talk, we present our work on compressing Time Projection Chamber data by signal importance. We also discuss the challenge of achieving more flexible compression after the neural network is trained.
Speaker: Yi Huang (Brookhaven national lab) - 12
-
15:45
Coffee break
-
13
SMOCS – JLab’s Streaming Monitoring Optimization Control SystemSpeaker: Armen Kasparian (Jefferson Lab)
-
14
Real-time AI-based dead hot map in the ePIC detector: a self-adaptive alternative to traditional big data calibration pipelinesSpeaker: Balazs Ujvari (University of Debrecen)
- 15
-
17:15
Discussion
-
10
-
Reception & Networking Event
-
-
-
Tutorial: RAG4EICConveners: Cristiano Fanelli (W&M), Karthik Suresh (member@wm.edu;employee@wm.edu;faculty@wm.edu;staff@wm.edu)
-
16
RAG4EIC TutorialSpeaker: Karthik Suresh (member@wm.edu;employee@wm.edu;faculty@wm.edu;staff@wm.edu)
-
16
-
AI/ML for ePIC and Beyond
-
17
ePIC AI/ML OverviewSpeaker: Dmitrii Kalinkin (Brookhaven National Laboratory)
-
18
Toward Unified Deep Learning Models for Simulation and PID with Cherenkov Detectors: the hpDIRC caseSpeaker: Cristiano Fanelli (W&M)
- 19
-
20
Multi-FPGA distributed MLP NN model for data reduction in ePIC dRICH readout systemSpeaker: Cristian Rossi (INFN Sezione di Roma)
-
11:05
Coffee Break
-
21
Scalable AI-assisted Workflow Management for EIC Detector Design Across Distributed Heterogeneous Resources with PanDA-iDDSSpeaker: Wen Guan (BNL)
-
22
Machine Learning for the pfRICH Particle Identification subsystem
We present an overview of the proximity-focusing Ring Imaging Cherenkov (pfRICH) detector developed for the ePIC experiment at the Electron-Ion Collider (EIC). Designed for the backward pseudorapidity region (−3.5 ≲ η ≲ −1.5), the pfRICH enables at least 3σ separation of pions, kaons, and protons up to 7 GeV/c, which is crucial for Semi-Inclusive Deep Inelastic Scattering (SIDIS) studies. In this talk, we explore the use of AI/ML techniques for pattern recognition of Cherenkov photon rings on photosensors in order to improve the PID capabilities of the pfRICH as a function of particle momentum. We use data from simulations of optical photon transport in Geant4, accelerated with NVIDIA OptiX and GDML-based detector geometries.
Speakers: Bishoy Dongwi, Charles-Joseph Naïm (Stonybrook) -
23
RAG-inspired Open-source based Q&A system for scholarly articles in EICSpeaker: Tapasi Ghosh (Ramaiah University of Applied Sciences)
-
24
ESI-Fastlight: a Conditional Normalizing Flow for Fast pfRICH Hitmap GenerationSpeaker: Gabor Galgoczi (BNL)
-
25
ML for the hKLM at the 2nd DetectorSpeaker: Rowan Kelleher (student@umich.edu;member@umich.edu;employee@umich.edu;student@annarbor.umich.edu;member@annarbor.umich.edu;employee@annarbor.umich.edu)
-
26
SRO AI/ML Models for Meson Structure Function ExtractionSpeaker: Sandeep Shiraskar (Catholic University of America)
-
27
Discussion
-
17
-
Break: Lunch
-
AI/ML for Data Analysis and Theory
-
28
Relativity Wasn’t in the Training SetSpeaker: Miles Cranmer (University of Cambridge)
-
29
Symbolic RegressionSpeaker: Douglas Adams (University of Virginia)
-
30
Neural Net ensembles for Bayesian inference of PDFsSpeaker: Maria Ubiali (University of Cambridge)
-
15:30
Coffee Break
-
31
Artificial Intelligence in the EIC era at the BSM-PDF frontierSpeaker: Tim Hobbs (Argonne National Laboratory)
-
32
ML-accelerated sampling for theorySpeaker: Phiala Shanahan (MIT)
-
33
Generative AI for data analysis and preservationSpeaker: Marco Battaglieri (Jefferson Lab)
-
34
What we talk about when we talk about gluon saturationSpeaker: Peter Jacobs (Lawrence Berkeley National Laboratory)
-
17:05
Coffee Break
-
35
DeepSub: Deep Image Reconstruction for Background Subtraction in Heavy-Ion CollisionsSpeaker: Umar Sohail Qureshi (Stanford University)
-
36
Deep Neural Networks for Extracting the 3D Structure of Nucleon at the EICSpeaker: Dr Ishara Fernando (University of Virginia)
-
37
Extraction of Chiral Odd Compton form factors using Maximum Likelihood Method from Exclusive 𝛑0 production experiment.Speaker: Dr Saraswati Pandey (member@virginia.edu;staff@virginia.edu;employee@virginia.edu)
-
38
Neural Network Generalized Parton DistributionsSpeaker: Zaki Panjsheeri (member@virginia.edu;student@virginia.edu;alum@virginia.edu;employee@virginia.edu;staff@virginia.edu)
-
28
-
-
-
Tutorial: RAG4EICConveners: Cristiano Fanelli (W&M), Karthik Suresh (member@wm.edu;employee@wm.edu;faculty@wm.edu;staff@wm.edu)
-
Trends in Data Science
-
39
AuroraGPT: A Foundation Model for ScienceSpeaker: Rajeev Thakur (Argonne National Laboratory)
-
40
Smart Pixel Project : AI for pixel readoutSpeaker: Lindsey Gray (Fermilab)
- 41
-
42
Geometric GNNs for Charged Particle TrackingSpeaker: Ahmed Mohammed (member@jlab.org;employee@jlab.org)
-
43
Electron-Proton Scattering Event Generation using Structured TokenizationSpeaker: Steven Goldenberg (Jefferson Lab)
-
44
FM4NPP: A Scaling Foundation Model for Nuclear and Particle PhysicsSpeaker: Shuhang Li (Columbia University)
-
11:30
Coffee Break
-
45
Examples of AI for Particle Physics
We give two examples of using the diffusion model for Physics. The first is in the LHC CaloChallenge. The second study is unfolding and surrogates for Jefferson Laboratory physics.
Speaker: Geoffrey Fox (University of Virginia) -
46
SAGIPS: A scalable Framework for SciDAC QuantOmSpeaker: Daniel Lersch (Jefferson Lab)
-
47
LLMs for particle physics analysis: results from studies on a toy modelSpeaker: Deepak Samuel (Central University of Karnataka)
-
48
AI Reasoning for Theoretical PhysicsSpeaker: Yurii Kvasiuk (University of Wisconsin)
-
49
Discussion
-
39
-
Break: Lunch
-
AI/ML in Production and Distributed ML
-
50
A Unified Vision of AI/ML at the Electron-Ion Collider: Infrastructure and CapabilitiesSpeaker: Linh Nguyen (Brookhaven National Laboratory)
-
51
ML in Production at SNS AcceleratorSpeaker: Anant Raj (ORNL)
-
52
Developing Fast ML on FPGA for Particle Identification and TrackingSpeaker: Denis Furletov (Brandeis University)
-
16:00
Coffee Break
- 53
- 54
-
55
Discussion
-
50
-
Introduction: Closing
-
