Choose timezone
Your profile timezone:
Powered by Indico
v3.3.2
The Annual RHIC & AGS Users' Meeting will be held online on June 8–11, 2021. The meeting will highlight the latest physics results from the STAR and PHENIX experiments and provide an outlook with the continuation of the STAR beam energy scan, the sPHENIX program, and the EIC.
Workshops that will be held on Tuesday, June 8th and Wednesday June 9th will enable more in-depth discussions of the beam energy scan, measurements of hard probes in heavy ion collisions, studies of small systems, strong field effects, cold QCD, careers in data science, physics opportunities at the EIC, and detector developments at the EIC. There will be a poster session on Thursday, June 10th. We will have plenary sessions on Thursday, June 10th and Friday, June 11th.
Presentations of highlights from all four RHIC experiments, the latest physics results from STAR and PHENIX experiments, an in-depth look at the proposed detector upgrades and future experiments, reports from representatives from the funding agencies, and award ceremonies will also be held during the plenary sessions
Event ID: 0000003671
Note: This meeting falls under Exemption E. Meetings such as Advisory Committee and Federal Advisory Committee meetings. Solicitation/Funding Opportunity Announcement Review Board meetings, peer review/objective review panel meetings, evaluation panel/board meetings, and program kick-off and review meetings (including those for grants and contracts) are open to the public.
Daniel McDonald spent years at STAR/RHIC while a graduate student at Rice University, getting his masters from his work on the TOF detector and his PhD from the higher moments of net-particle distributions in the Beam Energy Scan. He is the former lead of the STAR fluctuations PWG and a member of the UEC. Following a Postdoc at the University of Houston working on the ALICE experiment, he left to work in the financial services industry in 2015.
Daniel is presently the Head of Analytics and Strategy in Home Lending at Citizens Bank. His team has transformed data-based strategic thinking at the bank, driving growth that has seen the Home Mortgage origination group climb from not being a top 50 national player to 6th largest bank originator in the country, helping drive over $45 Billion annually in new home loans.
Javier Orjuela-Koop obtained his doctorate from the University of Colorado in 2018. His thesis comprised an exploration of various aspects of kinetic transport in small-system collectivity, as well as a measurement of separated heavy-flavor production in proton collisions using the PHENIX silicon tracker. Since leaving academia, he has found a rewarding career in the world of self-driving cars. Currently, as Lead ML & AI Engineer at HERE Technologies, he applies many of the skills of a heavy-ion physicist to the problem of building high-precision crowdsourced digital maps to enable autonomous driving in partnership with some of the world's best known car manufacturers.
Dr Frank Laue received his PhD in 1999 from Frankfurt University. He spent 2 years as a Post Doc at The Ohio State University and 6 years at BNL before leaving science for Wall Street in 2007. He worked for various Investment Banks and Asset Managers in NYC and California. He is currently an SVP at loanDepot, the nations 2nd largest non-bank mortgage lender.
Jan Uphoff did his PhD at the Goethe University Frankfurt in 2014 investigating heavy flavor in heavy-ion collisions at RHIC and LHC through numerical Monte Carlo simulations. After graduating he joined a startup offering Data Science as a Service to companies in the retail sector. Since 2017 Jan has been working at Booking.com – the world’s largest Online Travel Agency. He joined as a Data Scientist and is now leading the Data Science and Analytics teams focusing on the supply side of the marketplace.
Tim Schuster graduated from Frankfurt University in 2012 with a thesis on fluctuations in heavy-ion collisions. He analyzed data from the NA49 experiment at CERN, and contributed simulations to the original STAR Beam Energy Scan proposal. Tim continued his research at Yale University, studying correlations through ALICE at CERN.
Since 2015, Tim has worked in data science-related roles at Google. He uses data to detect and prevent spam and abuse, or to directly improve products ranging from Google Ads to YouTube and Stadia.
Betty Abelev graduated from Yale with a doctorate in Experimental Nuclear Physics in 2007, and then continued work on multi-strange baryons at Lawrence Livermore Lab, as a part of the ALICE collaboration. She left academia to pursue data science in the industry in 2013, and since then worked as a data scientist in several startups, one of which grew into a large corporation. That company is Invitae, a leader in clinical genetic testing. Betty worked at Invitae for the past five years, fulfilling various roles, including leading the production data-science group, and supporting clinical regulatory activities. She now leads data science activities in the Process Improvement group. Betty lives in San Francisco with her 12 year old daughter.
The European Commission published the "Ethics Guidelines for Trustworthy AI" in 2019. Currently, also major USA governmental organizations and corporations call for “responsible AI”. I will give you concreate examples of the relevant challenges and technical solutions from the real-life deep learning applications in recruitment industry.
'Matching of candidates to jobs by AI' is no longer science fiction. We understand it as a process to provide a fairer, more predictable and effective candidate pre-selection. However, bias can arise whenever AI is trained with historical data of human choices, which then calls into question the objectivity of AI-supported decisions. Conversely, AI can also be used to identify or prevent just such a bias. The use of algorithmic selection methods can lead to more diversity. As a partner in a Europe-wide research network, we are working on the open question of to what extent and under what conditions 'matching by AI' can represent a technologically feasible and ethically and legally legitimate solution, which we base on the intelligent comparison of competencies and qualifications.
A challenge of robustness is illustrated by another use case: predictive marketing, supporting recruiters in creating successful job ads tailored to the specific target group. The deep learning algorithm is continuously trained on the web tracking data of job posts. However, even when the models are regularly adjusted, AI systems would not respond optimally to disruptive events or to adversarial input that manifest themselves outside the captured stimuli. Therefore, we must create methods that automatically detect if the parameters of the input are significantly outside the range "seen" in the training set.
Newest advancement of AI technology is brought about by the idea of Hybrid AI as a combination of statistics, machine learning and knowledge systems. And these advancements are unavoidable to reach the levels of responsible or trustable AI.
Olena Linnyk is a lecturer for “Artificial Intelligence” at the Justus Liebig University of Giessen and the head of the “AI Lab” at the software company “Milch&zucker”, Giessen, Germany. At the Frankfurt Institute for Advanced Studies (FIAS), she contributes to the research of the “Deep Thinkers” Group, successfully applying Deep Learning techniques to the detector calibration and data analysis withing the NA61/SHINE Collaboration at CERN.
Industry adoption of deep learning computer vision has been driven by easy to use deep learning frameworks such as PyTorch and TensorFlow combined with the widespread availability of GPU compute and cheap cameras. By applying computer vision to images and video-streams, governments are generating societal benefits. Applications range from protecting minors by identifying sexual abuse to enhancing traffic safety of vulnerable bicyclists and pedestrians.
In this step-by-step hands-on contribution Hans will take us through the process of creating a deep learning computer vision solution for traffic safety. The use case requires the computer vision algorithm to not only detect moving objects in video streams with bounding boxes but also to generate finer-shaped segmentations of each object instance. Hans will show how you can create and train such a computer vision model, including creation and labeling of your own dataset from publicly available traffic video data, with just your laptop and no other pre-requisites in a mere two days.
Attendees of this session will leave empowered to create their own deep learning computer vision demo for their favorite use case in a very short time frame. Creating such a demo will help RHIC / AGS physicists to decide whether a Data Science career in industry is attractive to them and will help them in impressing even further potential employers in industry with their unique skillset.
DXC Technology is a 130 000 employee strong consultancy with highest ratings for its Data Science services. Hans coordinates DXC’s Data Science in Europe, Middle-East, and Africa. DXC is actively hiring Sr. Data Scientists with a Heavy-Ion / High-Energy Physics background in the US and globally.
Will be held at this link: https://indico.bnl.gov/e/aum2021-posters