This talk will discuss the different multi-armed bandit-based online learning algorithms and their applications. We will highlight the need to realize these algorithms on edge platforms for wireless radio, Internet of Things (IoT), and robotics applications and the various challenges of mapping these algorithms on SoC. Some of our proposed architectures for three well-known algorithms: Upper Confidence Bound (UCB), Kullback Leibler UCB (KLUCB), and Thompson Sampling will be presented. We will then explore hardware-software co-design and fixed-point analysis for the efficient realization on heterogeneous SoC. To support future richness and flexibility in dynamic environments, intelligent reconfigurable architecture will be discussed.
Schedule
Timezone: PDT
Multi-armed Bandit on System-on-Chip: Go Frequentist or Bayesian?
Sumit DARAK, Associate Professor
IIIT-Delhi
Sumit DARAK, Associate Professor
IIIT-Delhi
Dr. Sumit J Darak received an Engineering degree from Pune University, India, and PhD from NTU, Singapore, in 2007 and 2013, respectively. He is an Associate Professor with IIIT-Delhi, and SoC Consultant with Apexplus Technologies, Hyderabad, India. His current research interests include the design of efficient algorithms for wireless, radar, and artificial intelligence (AI) applications and mapping to reconfigurable and intelligent architectures on SoC. His students have received numerous awards, such as DST Inspire Faculty Award, Second Best Paper Award at IEEE DASC 2017, Best Demo CROWNCOM 2016, Second Best Poster at COMSNETs 2019, Design Contest in VLSID 2022 and 2023, NI academic research grants and Qualcomm Innovation Fellowship 2023.
Schedule subject to change without notice.