Nahid Emad

Parallel and Distributed Eigenproblems for Large AI models

In this talk, we highlight the importance of high-performance eigenvalue computation for data analysis in large AI models, particularly in areas ranging from dimensionality reduction to understanding the internal dynamics of data and learning-models through the separation of target signals and noises in signal processing and analysis of and stability and convergence analysis of RNNs. We focus on dimensionality reduction problem which is a representative example of the strong interactions between machine learning and linear algebra, and which plays a prominent role in high-performance big data analysis. Using this problem, which is based on large sparse eigenproblem, we show how to take advantage of these interactions and commonalities to propose new approaches to problem solving in both domains. An innovative machine learning approach based on Unite and Conquer methods, used in linear algebra, will be presented. In addition to its efficiency from an accuracy point of view, the important characteristics of this inherently parallel and scalable technique make it well suited to multi-level and heterogeneous parallel and/or distributed architectures. Experimental results, partly on the #1 supercomputer of the HPCG list, Fugaku demonstrating the interest of the approach for efficient data analysis in the case of applications such as clustering, cybersecurity and health will be presented. We will finally emphasize that these methods open up very general perspectives showing their applicability in areas such as high-performance processing of GNNs and LLMs.
 

back to overview
 

Biography

Nahid Emad received the Habilitation to Direct Research in computer science from the University of Paris Saclay/Versailles, the PhD and MS in applied mathematics from Pierre and Marie Currie University (Sorbonne University) and BS in pure mathematics from the University of Arak (Iran). She is a Professor at the University of Paris Saclay/Versailles and affiliated with the Maison de la Simulation and LI-PARAD laboratories where she heads the Intensive Numerical Computing group. She maintains long international collaborations, notably with Japan and Germany through the bi- and trilateral ANR and SPPEXA projects and with the United States where she is a regular visiting professor at the University of California at Berkeley. She has been scientific supervisor of 20 PhDs and HDRs and is the author of more than 150 articles in international journals, conferences, and book chapters. Her main research interests include numerical algorithms, linear algebra, parallel and distributed programming methodology, software engineering for parallel and distributed numerical computing, and big data analysis.
Footer

Imprint / Privacy