Data Science and Learning

Machine learning is proving to be extremely interesting and broadly applicable in many fields of science. From accelerating the design of algorithms themselves, to AI-driven design loops, to efficient methods to handle and interrogate zettabyte datasets, and much more, data science is changing the way science is done. Scientists at PPPL are working on a wide range of machine learning projects, including data fusion; real-time control systems for tokamaks; acceleration of exascale simulations of tokamak turbulence; real-time control systems for microelectronic fabrication facilities; tokamak disruption avoidance using recurrent neural networks; prediction and control of smaller scale, transients bursts of heat from confined thermonuclear plasmas; AI-driven stellarator optimization algorithms; and more. 

Machine learning lunch meetings

The Computational Science Department arranges a weekly seminar every Wednesday at 12:30pm. More information is on the seminar page .

General machine learning resources

ASTRO code repos and software orgs

HEP code repos and software orgs

ASTRO meetings with ML/DL content

HEP meetings with ML/DL content

ASTRO presentations and papers

HEP presentations and papers

Machine Learning in the LIGO-Virgo Era 

Machine learning in particle physics 

MicroBooNE Search for Low-Energy Excess Using Deep Learning Algorithms

Automated proton track identification in MicroBooNE using gradient boosted decision trees 

Deep Learning Applications in the NOvA Experiment 

Exploration of Deep Convolutional and Domain Adversarial Neural Networks in MINERvA 

Advanced machine-learning solutions in LHCb operations and data analysis 

Deep Learning and DUNE 

Applying Deep Learning in MicroBooNE 

Identifying object in ATLAS through machine learning techniques 

Prototype of Machine Learning “as a Service” for CMS Physics in Signal vs Background discrimination 

Deep learning: experimental developments 

DeepFlavour and Tensorflow in CMSSW 

HiggsML challenge with optimized DNN 

HEP Machine Learning on HPCs 

Event Reconstruction with Deep Learning 

Exploring Raw HEP Data using Deep Neural Networks at NERSC 

State-of-the-art Machine Learning in event reconstruction and object identification 

Machine learning techniques for heavy flavour identification

Development of Machine Learning based muon trigger algorithms for the Phase2 upgrade of the CMS detector 

Machine and deep learning techniques in heavy-ion collisions with ALICE