- Dennis was featured in the LBNL daily newsletter Elements (link to his presentation is here):
10/9/2024
- Many papers accepted at the Machine Learning and Physics Workshop at NeurIPS 2024! (Berkeley names only - we will post links to the full papers when they are public)
Neural Posterior Unfolding: Jing, Krish, Jay, Vinicius, Fernando
Generation of Air Shower Images for Imaging Air Cherenkov Telescopes using Diffusion Models: Jonas, Vinicius, Lark
Multidimensional Deconvolution with Profiling: Richard, Krish, Vinicius
WOTAN: Weakly-supervised Optimal Transport Attention-based Noise Mitigation: Nathan, Vinicius
A Platform, Dataset, and Challenge for Uncertainty-Aware Machine Learning: Wahid, Sascha, Jordan
Point cloud diffusion models for the Electron-Ion Collider: Fernando, Vinicius
10/3/2024
- Our white paper FAIR Universe HiggsML Uncertainty Challenge Competition is now on arXiv!
- Another great summer school on ML for Fundamental Physics! Was great to broaden the scope this year - fantastic leadership by Aishik, Elham, Sascha, and others!
Five days of intensive lectures hands-on tutorials at the ML for Fundamental Physics School at LBL comes to and end. Thanks to the nearly 400 registered participants from ATLAS, CMS, LZ, Dune, Belle 2, phenomenology and many other backgrounds for your participation over the week! pic.twitter.com/U5caFdHSj4
- Our paper Accelerating template generation in resonant anomaly detection searches with optimal transport led by collaborators at the University of Geneva is now on arXiv!
- Our paper Measurement of Track Functions in ATLAS Run 2 Data led by Jing was released as a Conference Note for BOOST!
- The paper we were involved in to compare ML method on realistic simulations, Accuracy versus precision in boosted top tagging with the ATLAS detector, led by Kevin Greif (UC Irvine), is now on arXiv!
- Our paper Parnassus: An Automated Approach to Accurate, Precise, and Fast Detector Simulation and Reconstruction is now on arXiv!
5/30/2024
- Our ATLAS OmniFold paper led by Mariel (and with foundational contributions from Adi) is now on arXiv! Check out the public data and software that go with it! See also this article in the CERN EP Newsletter entitled A high-dimensional jet-powered measurement of the strong force.
5/24/2024
- Radha's paper Constraining the Higgs Potential with Neural Simulation-based Inference for Di-Higgs Production is now on arXiv!
- Our paper Advancing Set-Conditional Set Generation: Diffusion Models for Fast Simulation of Reconstructed Particles is now on arXiv!
5/14/2024
- Alkaid's and Gup's paper Incorporating Physical Priors into Weakly-Supervised Anomaly Detection is now on arXiv!
5/11/2024
- Our paper Design of a SiPM-on-Tile ZDC for the future EIC and its Performance with Graph Neural Networks is now on arXiv!
5/8/2024
- Ben represented all of AI/ML at Berkeley Lab at the AI Expo for National Competitiveness, talking about Reimaging the exploration of fundamental interactions with AI:
- Many papers accepted at the Machine Learning and Physical Sciences Workshop at NeurIPS! Congratulations to Elham, Fernando, Mariel, Shahzar, and Vinicius! Camera-ready versions will be posted soon.
- Mariel is part of a team (called Polymath) to study large language models for science and they have made their debut today! Here is a blog post about it written by Mariel:
10/2/2023
- Our paper on The Optimal use of Segmentation for Sampling Calorimeters led by Fernando is now on arXiv!
9/29/2023
- Ben gave the UC Berkeley BIDMaP seminar on Re-imagining the search for fundamental interactions with machine learning.
- Our paper on CaloScore v2: Single-shot Calorimeter Shower Simulation with Diffusion Models led by Vinicius is now on arXiv!
8/4/2023
- The BOOST conference was a big success! The conference photo below is a link to the agenda with many interesting presentations! See also this summary from ATLAS.
8/1/2023
-The ml4jets conference poster is now online!
7/28/2023
-The US ATLAS ML Training Event was a big success! There were a number of talks by group members (indico linked below).
7/28/2023
- We went for a hike at The Dish to celebrate Ben's promotion to career staff scientist!
- Our paper on Resonant Anomaly Detection with Multiple Reference Datasets with Mayee Chen (Stanford) and Fred Sala (U. Wisconsin) is has been published in JHEP!
7/20/2023
- Our paper on The Interplay of Machine Learning--based Resonant Anomaly Detection Methods led by Radha is now on arXiv!
7/18/2022
- Ben was quoted in a Symmetry article about AI/ML and Simulation:
7/17/2023
- Fast Point Cloud Generation with Diffusion Models in High Energy Physics led by Vinicius and Mariel is now on arXiv!
- Our paper on Resonant Anomaly Detection with Multiple Reference Datasets with Mayee Chen (Stanford) and Fred Sala (U. Wisconsin) is has been accepted for publication in JHEP!
7/10/2023
- Comparison of Point Cloud and Image-based Models for Calorimeter Fast Simulation led by Fernando and Vinicius is now on arXiv!
- Ben gave a talk about Machine Learning for Precision Physics at the LHC at LoopFest.
- Our paper on Learning to Isolate Muons in Data with Ed Witkowski and Daniel Whiteson now on arXiv!
- Ben wrote an article published in the CERN EP Newsletter
6/22/2023
- Our paper on covariant particle transformers is now published in PRD!:
6/20/2023
- Our paper on modeling theory uncertainties is now published in SciPost!:
6/19/2023
- Our paper on Unbinned Profiled Unfolding led by Jay has also been accepted to the The Synergy of Scientific and Machine Learning Modelling (SynS & ML) Workshop at ICML!
6/16/2023
- Our joint grant with Anja Butter's group at LPNHE in Paris on unfolding has been funded through the France-Berkeley Fund!
- Our paper on High-dimensional and Permutation Invariant Anomaly Detection led by Vinicius is now on arXiv!
6/1/2023
- Ben gave a talk about the Range of Machine Learning methods in Particle Physics at the PhyStat-2sample workshop.
5/26/2023
- Our paper on Fitting a Deep Generative Hadronization Model led by Jay is now on arXiv!
5/22/2023
- Paper on Statistical Patterns of Theory Uncertainties with Aishik Ghosh, Tilman Plehn, Lily Shire, Tim Tait, and Daniel Whiteson has been accepted for publication at SciPost!
5/17/2023
- Our paper on Learning Likelihood Ratios with Neural Network Classifiers led by Mariel and Shahzar is now on arXiv!
5/12/2023
- Our paper on Enhanced latent spaces for improved collider simulations is now on arXiv!
- Our paper on implementing 2+1 Abelian lattice gauge theories on a quantum computer with Chris Kane, Dorota Grabowska, and Christian Bauer is now on arXiv!
11/15/2022
- Our paper on optimizing long-lived particle detectors with Tom Gorordo, Simon Knapen, Dean Robinson, and Adi Suresh is now on arXiv!
11/7/2022
- Our Fair Universe Project led by Wahid Bhimji has been highlighted in a Berkeley Lab press release:
- Our NNSA NA-22 project Deep Learning Based Particle Identification Methods for Silicon CCDs (led by Ren Cooper) was funded! We have a postdoc opening for someone to apply machine learning for particle identification.
9/22/2022
- Ben gave the Physics and Astronomy Department Colloquium at the University of San Francisco.
- Our Calomplification paper has been accepted for publication in JINST!
- Elham and Aishik (with help from many others from our group!) organized a successful ML training event, sponsored by US ATLAS and with help from NERSC! Ben gave an Overview of Machine Learning for HEP.
- Ben gave a talk on silicon sensor simulation in an instrumentation parallel session at the Snowmass Community Summer Study.
7/17/2022
- The Snowmass Community Summer Study began today! We are involved in many parts of this event, mostly in the context of the Computational Frontier. I won't list all of the Computational Frontier events (there are many) - see the website for details!
- Yao's differential electron-jet correlation measurement with H1 data is now public in time for his talk at ICHEP!
6/27/2022
- Many abstracts related to work in our group have been accepted for talks at BOOST 2022! In particular, Matt LeBlanc (CERN) will talk about topic modeling for the strong coupling, Sascha Diefenbacher (Hamburg) will talk about online density estimation, Sebastian Bieringer (Hamburg) will talk about statistical amplification of generative models, Tobias Quadfasel (Hamburg) will talk about anomaly detection with CATHODE, Dilia Quintero (TRIUMF) will talk about ml4pions, and Christine McLean (ANL) will talk about the future of jet substructure.
- Our review on searches with machine learning with Georgia Karagiorgi, Gregor Kasieczka, Scott Kravitz, and David Shih has been published in Nature Reviews Physics!
- Snowmass white paper deadline is today! A few white papers came out early (see below) and some will be late (see above), but two came out right on time:
- Our review on searches with machine learning with Georgia Karagiorgi, Gregor Kasieczka, Scott Kravitz, and David Shih has been accepted for publication in Nature Reviews Physics!
- Our paper on decorrelated autoencoders with Vinicius Mikuni and David Shih is has been accepted by PRD and also selected as a highlight as an Editor's Suggestion!
1/28/2022
- Our paper on decorrelating theory uncertainties is now published in Eur. Phys. J. C!
- Krish, Radha, and Vinny presented excellent posters at a poster session at the Berkeley Institue for Data Science:
12/02/2021
- Our group has been awarded a grant from the DOE, office of Nuclear Physics to optimize the detectors of the Electron Ion Collider with machine learning! The press release from the DOE is here and the article in Elements is below:
- Ben gave a talk about silicon simulations for the silicon instrumentation Snowmass meeting
- Our paper on decorrelated autoencoders with Vinicius Mikuni and David Shih is now on arXiv!
11/4/2021
- The LHC Olympics paper is now published in Reports on Progress in Physics!
11/2/2021
- Ben gave a talk on Deep Learning, Quantum Information, and the LHC as a Gluon Factory at the weekly Harvard LPPC seminar.
10/29/2021
- The LHC Olympics paper has been accepted for publication in Reports on Progress in Physics!
10/27/2021
- Our paper building on the Random Identity Insertion Method for quantum gate error mitigation with Vince Pascuzzi, Andre He, Christian Bauer, and Bert de Jong is now on arXiv!
- The combination of OmniFold with BSM searches paper was published today in Phys. Rev. D!
10/22/2021
- Four papers from our group have been accepted for poster presentations a NeurIPS workshops this year!
A. Hallin, et al., Classifying Anomalies THrough Outer Density Estimation (CATHODE), ML4PS2021
A. Ghosh, B. Nachman, D. Whiteson, Uncertainty Aware Learning for High Energy Physics With A Cautionary Tale, ML4PS2021
K. Desai, B. Nachman, J. Thaler, Symmetry Discovery with Deep Learning, ML4PS2021
R. Winterhalder, M. Bellegente, B. Nachman, Latent Space Refinement for Deep Generative Models, DGMs2021
- Ben was awarded the 2022 Henry Primakoff Award for Early-Career Particle Physics by the American Physical Society:
10/14/2021
- Ben gave a talk about the H1 ML-based unfolding measurement in the Physics Division ML Group Meeting (slides are only visible to the Berkeley community).
- At this week's AI4EIC workshop, Ben will co-organize a session on Accelerator and Detector Control (Thursday) and give talks about simulation (Tuesday) and reconstruction (Wednesday).
- Welcome to Mariel Pettee who is a new Chamberlain Fellow and Vinicius Mikuni who is a new NESAP Fellow!
- A joint venture of many collaborators has now produced CATHODE, a density-based anomaly detection method that outperforms many other approaches on the LHC Olympics. The manuscript is now on arXiv!
- The paper we contributed to on memory confidence in rats has now been published in Current Biology!
- The Mainz Institute for Theoretical Physics 2 week program on Machine Learning for Particle Physics started today with interesting talks about optimal transport!
- The ML4Jets program is now public! We had 100 abstract submissions and for the first time, we have to have parallel sessions to fit in all of the exciting talks.
- Krish gave an excellent talk in today's Physics Division Machine Learning Group meeting on his Symmetry Discovery project.
4/21/2021
- Ben is the Physical Science Area's representative to the lab committe Future Berkeley Lab Computing Working Group. This committee was highlighted in today's lab research update article.
- Congratulations to Kees for accepting an offer for physics graduate school at Harvard and to Luc for accepting an offer for physics graduate school at UC Berkeley!
- Ben was a panelist in the ML Club debate on anomaly detection for discoveries:
4/13/2021
- CERN Yellow Report on radiation damage at the LHC is now published! This has been a multi-year effort, resulting from multple workshops and coordination across all four LHC experiments and including RD50. Ben was the co-editor of the sensor measurements section and contributed to the simulation section.
- Our first result as a member of the H1 Collaboration with Miguel Arratia is now public: using OmniFold applied to deep inelastic scattering data to probe the strong force in new ways! Here is an event display from the measurement:
- The video and slides from the deep generative models for fundamental physics workshop are now online.
4/2/2021
- Our papers on extending (1) OmniFold to include background and acceptance effects and (2) the statistical power of deep generative models were accepted at the simDL workshop at ICLR 2021!
3/31/2021
- Congratulations to Jason Huang for accepting an offer for physics graduate school at UIUC!
3/30/2021
- Today we heard interesting talks about machine learning for jet physics at the KITP workshop from Jesse Thaler and Frederic Dreyer.
- Welcome Diana Chamaki (Berkeley Undergraduate) to the group!
3/11/2021
- The ninth meeting of the Physics Division ML group in 2021 took place today! Aishik Ghosh spoke about our project that incorporates uncertainty into deep learning inference.
3/4/2021
- The eigth meeting of the Physics Division ML group in 2021 took place today! David Shih spoke about anomaly detection applied to steller streams.
- Ben was elected as the secretary of the American Physical Society's Group on Data Science:
2/5/2021
- The readout rebalancing paper was published today in Phys. Rev. A!
2/4/2021
- The fourth meeting of the Physics Division ML group in 2021 took place today! Daniel Murname spoke about metric learning and graph neural networks for particle tracking.
- The third meeting of the Physics Division ML group in 2021 took place today! Vanessa Boehm spoke about differentiable simulations for weak cosmic lensing.
- Ben is now a member of the H1 Collaboration at DESY, working on a multi-differential cross section measurement with Miguel Arratia probing azimuthal decorrelations.
- The second meeting of the Physics Division ML group in 2021 took place today! Jack Newsom and Ethan Lu spoke about the application of machine learning to neutrino event reconstruction.
- Ben gave the talk A Neural Resampler for Monte Carlo Reweighting with Preserved Uncertainties at the CMS Machine Learning forum (unfortunately, the meeting agenda was private).
- The first meeting of the Physics Division ML group in 2021 took place today! George Stein spoke about Self-Supervised Representation Learning for Astronomical Images.
1/12/2021
- UC IrvineYvonne Ng won an US ATLAS Center grant to visit Berkeley Lab last fall to work on ML-based anomaly detection. That was and is still not possible, but we are happy to welcome her as a "virtual" visitor this spring!
- Welcome Ming Fong (Berkeley Undergraduates) to the group!
- Many people from our group and collaborators are presenting posters at the Machine Learning for Physical Sciences workshop at NeurIPS today. In particular:
- The fourth regular meeting of the Physics Division ML group took place today! Giuseppe Puglisi spoke about deep generative models for the Cosmic Microwave Background (CMS).
12/9/2020
- The SRGN paper has been accepted for publication in PRD!
- The Chamberlain Fellowship interviews have begun! As candidates cannot travel to Berkeley this year, the Physics Division has prepared a virtual tour. Towards the end, you can see a brief highlight of the Division's exciting work in quantum information and machine learning:
- Today we had a first Nachman ML Group meeting! Was great to see everyone in the same (virtual) room!
12/3/2020
- The third regular meeting of the Physics Division ML group took place today! Our very own Adi Suresh gave a very intersting and engaging talk about Parameter Estimation using Neural Networks in the Presence of Detector Effects
- The second regular meeting of the Physics Division ML group took place today! We had interesting discussions about machine learning prospects for LZ and DESI.
- The first regular meeting of the Physics Division ML group took place today! We had about 45 people in attendence from all areas of fundamental physics and across the lab and wider Berkelely community.
- BIDS advertises the Berkeley Lab Physics Division ML group meetings.
- Our unfolding paper for quantum computer readout error mitigation was featured in a press release by Berkeley Lab.
Ouail Kitouni on Enhancing searches for resonances with machine learning and moment decomposition at 7:30 am PDT on Friday. Here are the slides and also the recording.
- The DCTRGAN paper was accepted for publication in JINST!
9/17/2020
- The Neural Resampler paper was accepted for publication in PRD!
- Ben gave a talk at the IF03 Solid State Detectors and Tracking Snowmass topical group meeting about the challenges for radiation damage simulation tools.
- The agenda for the Snowmass Computational Frontier workshop next week is finalized: August 10-11. There are many interesting talks, especially in the machine learning and quantum computing parallel sessions!
