Title: Machine Learning and How Physicists Can Think About It
Abstract:
In the last eight years there has been an explosion of progress in Machine Learning. In this colloquium I'll explain the (very simple) ideas underlying Neural Networks, and give a few examples of their structure and current capabilities. Then I'll survey the increasing scales of data and computation in this field, and make some comparisons and projections to see where it could be headed. If time permits, I'll also discuss my recent work on scaling laws for machine learning, and its connection to language models and GPT-3.
Speaker: Jared Kaplan(Johns Hopkins U.)
Time and Date: 22 June 2020, 03:30 PM - 05:30 PM CEST (Berlin Time zone)
17th seminar in the Quantum Aspects of Space-Time and Matter (QASTM) Seminar series
Contact: Choudhury, Sayantan (sayantan.choudhury@aei.mpg.de)
- Talk by Professor Jared Kaplan, Dept. of Physics, Johns Hopkins University, USA at QASTM seminar ( Download)
- Talk by Professor Marc Kamionkowski, Dept. of Physics Johns Hopkins University, USA at QASTM seminar ( Download)
- Jared Kaplan - The Temperature of a Pure State in CFT ( Download)
- Machine Learning II (Jared Kaplan) ( Download)
- Neural Scaling Laws and GPT-3 - Jared Kaplan ( Download)
- Machine Learning III (Jared Kaplan) ( Download)
- Johns Hopkins Physics Professor Shares In $3 Million Award ( Download)
- Neural Scaling Laws and GPT-3 ( Download)
- Strings 2017 - Jared Kaplan - AdS 3 CFT 2 and the information paradox ( Download)
- 2017 Bootstrap School - Jared Kaplan, Lecture 1 ( Download)
- Jared Kaplan - CEO Interview - OppFi ( Download)
- Jared Kaplan - CEO - OppLoans ( Download)
- Lecture 25 Sparse Subspace Clustering (Hopkins) ( Download)
- HE Seminar - 9/24/21 - Surjeet Rajendran - Johns Hopkins University ( Download)
- Lecture 6 Model Selection for PCA (Hopkins) ( Download)
