Caspar.AI - AI for Physically Embodied Systems & Intro to Apache Quickstep

SVBDS Makers,

Join us this evening with Dr. Ashutosh Saxena, Co-Founder, and CEO of Caspar.AI and Harshad Deshmukh, PhD candidate in the Database Systems Group at the University of Wisconsin - Madison.

See you then!

Agenda

6:00 - 6:30 PM - Doors open & food

6:30 - 7:15 PM - Ashutosh's talk

7:15 - 8:00 PM - Harshad's talk

8:00 - 8:30 PM - Q&A & networking

Abstract:

People spend over 60% of their lives in their home. However, the technology found in homes today is all from the last century! In this talk, I'll describe how technologies such as Artificial Intelligence are changing the landscape in this area. Systems such as homes, cars and robots are installed with 100+ sensors and full control of devices in each home (shades, lighting, climate, ...). This gives an AI very rich data to play with.

In this talk, I'll describe the space of smart homes at Caspar.AI, some technical details of new deep learning algorithms we have developed, and future directions.

Bio:

Dr. Ashutosh Saxena is the Co-founder and CEO of Caspar.AI. Caspar.AI is building smart homes of the future which interact with you, learns from your preferences and work with you. In 2015 Ashutosh was selected among top 8 innovators to watch by the Smithsonian institution. He is also the recipient of the World Technology Award in 2015. Before Caspar.AI, Ashutosh spent four years as an assistant professor in Computer Science Department at Cornell University, where he founded Robot Learning Lab and co-founded Zibby. His vision is to build artificial intelligence for embodied systems such as robots, cars, and homes.

Dr. Ashutosh Saxena received his PhD in machine learning from Stanford University with Andrew Ng, and his B.Tech. in 2004 from Indian Institute of Technology (IIT) Kanpur. He has also won best paper awards in 3DRR, RSS, IEEE ACE and IROS. He was named a co-chair of IEEE technical committee on robot learning and is an associate editor of the IEEE Transactions on Robotics. He was a recipient of Google Faculty award in 2011. He was named a Alfred P. Sloan Fellow in 2011, named a Microsoft Faculty Fellow in 2012, received a NSF Career award in 2013, and received an Early Career Award at RSS 2014. His work has received substantial amount of attention in popular press, including the front-page of New York Times, BBC, ABC, New Scientist Discovery Science, and Wired Magazine. He was also a Chief Scientist at Holopad that built the 3D experience for Steven Spielberg's movie TinTin.

Abstract:

Modern hardware offers tremendous computing and storage capabilities, which are largely untapped by most traditional database software. Apache Quickstep is a new, open source in-memory data platform designed to leverage the high degree of processing parallelism that is present in modern processing hardware. The system uses a novel “scaling in” approach, and specialized query optimization, scheduling, and data storage and organization techniques to achieve high performance. The system also has a “no-tuning-knobs” approach, making it easy to deploy in cloud settings. Our first prototype targets single node deployment environments, and performance evaluation reveals that we are often an order of magnitude faster than other open source systems.

In this talk, I will describe the system architecture, the novel query processing methods in Quickstep, and shed light on some of the use cases for deploying the system.

Quickstep is currently under Apache incubation.

Speaker name: Harshad Deshmukh

Speaker bio: I am a PhD candidate in the Database Systems Group at the University of Wisconsin - Madison where I work with Prof. Jignesh Patel on the Quickstep project. I work on resource management in Quickstep, particularly its query scheduler. Currently I am doing an internship with Microsoft's CISL group where I am working on stream processing systems. In the past I have done internships with Pivotal and Samsung research. I received B.E. in Computer Science and M.Sc in Mathematics from BITS Pilani, Pilani campus in India.