Grainger College of Engineering, All Events

View Full Calendar

NCSA Center for AI Fall Seminar Series: Rajaram Krishnamurth, IBM

Event Type
Seminar/Symposium
Sponsor
Center for Artificial Intelligence Innovation (CAII)
Virtual
wifi event
Date
Sep 14, 2020   11:00 am - 12:00 pm  
Registration
Registration is preferred
Originating Calendar
ECE ILLINOIS Alumni Events

CAII kicks off their fall seminar series on Monday, September 14 (11:00 a.m. - 12:00 p.m. CDT) via Zoom with IBM's Rajaram Krishnamurthy (Technical Staff Member from IBM Systems) on a talk about AI adoption, challenges, and opportunities in Enterprise. Attend weekly and learn about how researchers use AI and how collaboration advances research across multiple disciplines to impact society and the world. 

Zoom event: https://illinois.zoom.us/j/99581343946?pwd=ckhnQjRKVFA5L0tZdVp4ckpKQmltUT09#success 

Talk title: "End-to-End Issues in AI: Challenges and Research Opportunities."

Abstract: While significant industry focus for AI has been in the development of systems, accelerators and SW frameworks for training and inferencing, challenges remain in operationalizing AI across the Enterprise. Skills, lack of labeled data, lack of trust in models and decisions, robustness, transparency are often cited as challenges to widespread adoption of AI. We will begin by sharing our experiences with building real enterprise AI solutions in the industry. Enterprise AI solutions must preserve existing trust and reliability in mission-critical systems and cannot become the "weakest link." We will then compare Enterprise AI challenges with funding priorities of DARPA, NSF and recent VC investments. We will list any AI investment gaps and Research opportunities to address these gaps. Our approach will be to address issues surrounding entire AI workflows from data-prep, training, inferencing, retraining, model monitoring and management. We will also take a whole-stack approach with emphasis on compute, accelerators, runtimes, SW frameworks, SW applications and systems design. We will address issues spanning rich modalities of data encountered in the Enterprise.

link for robots only