NVIDIA

At the January Chicago IEEE Computer Society meeting (1/20/21 Online) learn how those of you who look at algorithms and see opportunities for massive parallelism can apply that skill to problem solving. Financial modelers can now use advanced computer architectures known as GPUs and a Python-based framework known as RAPIDS to speed up problems using big data and machine learning. Rather than taking minutes to run, these problems will be completed in seconds. We show a detailed analysis of several million loans where the contribution of each feature is discovered via their Shapley values along the lines of this article:

https://developer.nvidia.com/blog/explaining-and-accelerating-machine-learning-for-loan-delinquencies/

Biography:

Mark Bennett is from the Chicago area and has been working in large datasets and real-time high performance computing for over three decades. Mark is currently senior data scientist at Nvidia Corporation where he focuses on acceleration for financial machine learning.

He has taught financial analytics at the University of Iowa and the University of Chicago. Mark holds a Ph.D. from UCLA, an M.S. from the University of Southern California, and a B.S. with Distinction from the University of Iowa, all in computer science.

His early work experience was in applied mathematics at Argonne National Laboratory and as a research scientist at Unisys Corporation. Later he was a member of technical staff at AT&T Bell Laboratories, senior technical advisor and engineering manager at Northrop Grumman aerospace, senior technology specialist at XR Trading Securities and senior quantitative finance analyst at Bank of America Securities.

He is also the co-author of Financial Analytics with R published by Cambridge University Press (ISBN: 1107150752)