Abstract:
Dr. Agarwal has spent the last 17 years applying computer and data science research to science challenges. This work has led to many insights into gaps in the available computing research. It has also led to many insights into how to engage scientists and have impact. This talk will discuss some of these challenges and insights as well as lessons learned along the way.

Dr. Deborah Agarwal is a Senior Scientist and Division Director for the Scientific Data Division, at Lawrence Berkeley National Laboratory (LBNL). The Scientific Data Division (SciData) transforms data-driven discovery and understanding through the development and application of novel data science methods, technologies, and infrastructures with scientific partners. Dr. Agarwal’s current research focuses on developing computational tools to enable scientists to organize and use Earth Science data to address challenges more effectively. She has worked on projects involving carbon flux, watershed understanding, tropical forests, soil carbon, carbon capture, cosmology, particle accelerators, data repositories, and satellite data. She is also active in efforts to broaden diversity in computing research and a member of the Computing Research Association Committee on Widening Participation.

Abstract:
We begin by describing the rapid strides made in artificial intelligence by the growing availability of computational power, and the democratization of these capabilities via Cloud. In particular, we describe the dramatic recent progress of generative AI models. We then describe several open challenges as we seek to advance AI in a manner that benefits billions of human beings who have been relatively untouched by technology, starting with people in India. We present our work on multilingual models to democratize information access in a diverse set of Indian languages, dealing with challenges like code mixing, non-unique transliterations, idiomatic misspellings and mispronunciation of English words influenced by the native language of the user. We describe a couple of problems in healthcare in the Global South where we lack data in digital form to begin with, but where AI can help. Finally, we describe our efforts to build foundational layers of an agricultural stack, based on analysis of satellite imagery, to help transform agriculture and the lives of farmers. Through these examples, we hope to convey the excitement of being an AI researcher at this time when there are fascinating open problems to tackle and an unprecedented opportunity to make a difference to the world.

Abstract:
Next generation applications such as augmented and virtual reality, autonomous driving, and Industry 4.0, have tight latency constraints and produce large amounts of data. To address the real-time nature and high bandwidth usage of these applications, hierarchical cloud computing provides an extension to the cloud infrastructure through a hierarchy of datacenters located between the edge devices and the cloud. This proximity to the location of data generation enables new systems that provide low latency and reduce network load.  Simultaneously, the network and the application workload outside of the wide-area cloud datacenter and closer to the edge, becomes more dynamic requiring system software to adapt more frequently.  In this talk, I will describe new challenges and opportunities for the design of system software for hierarchical cloud networks.  I will provide examples from my group’s recent efforts to create distributed systems to simplify the deployment of novel applications on a hierarchy of data centers that span from the edge of the network for the wide-area cloud.

Prof. Eyal De Lara was awarded his PhD in Electrical and Computer Engineering from Rice University in 2002. Upon graduation, he joined the University of Toronto as an Assistant Professor in the Department of Computer Science. In 2007, he was awarded tenure and promoted to Associate Professor. In 2015, he was promoted to the rank of Professor. He currently serves as the Chair of the department. Professor de Lara’s research lies in the areas of cloud and mobile computing where he has contributed novel algorithms for system virtualization, edge computing, application scaling, indoor localization, mobile security, and continuous mobile sensing. Professor de Lara’s research has had a significant impact in both academia and industry. His research on VM fork, a new cloud computing abstraction, was commercialized by GridCentric, a Toronto based start-up that was acquired by Google. Professor de Lara’s work has been recognized with the EuroSys Test of Time Award, the CACS/AIC Outstanding Young Computer Science Researcher Prize, an NSERC Discovery Accelerator Award, Faculty Awards from IBM and VMware, as well as 3 best paper awards and 2 best paper honorable mentions. Professor de Lara has served as the editor in chief of GetMobile, the flagship publication of ACM SIGMOBILE, and has co-chaired the technical program committees of several conferences including ACM MobiSys and ACM/IEEE SEC, which are respectively the top venues in mobile systems and edge computing. His contributions to the mobile research community have been recognized with the 2020 SIGMOBILE Distinguished Service Award.