We're excited to be at Microsoft Japan for our next ML TOKYO TALKS event and welcome two fantastic speakers, sharing insights related to Machine Learning on edge devices: Michael Lanzetta, Microsoft (Applied ML Lead for CSE Asia) and Sakyasingha Dasgupta, CEO EdgeCortix Inc.
● 6:30 pm Doors open
● 7:00 pm Talk 1 + Q&A: "AI on the Edge", Michael Lanzetta, Microsoft
● 7:45 pm Talk 2 + Q&A: "Designing AI accelerators with Software Defined Hardware & Co-exploration", Sakyasingha Dasgupta, EdgeCortix Inc.
● 8:30 pm Networking
● 9:15 pm Closing
● TALK 1: Michael Lanzetta's team at Microsoft works with some of the largest companies in the world, building bespoke AI solutions in some of the most complicated Edge and IoT environments. His talk will focus on a few engagements and walk through solutions, with advice on problems they encountered and how the team addressed them that will help you in your own AI on the Edge journey.
● TALK 2: The method of neural architecture search (NAS) has achieved great success in liberating human labor in the design of neural architectures for image classification and segmentation. However, the conventional NAS with the singular objective on better accuracy, without the consideration in hardware typically leads to poor performance on latency, throughput and memory, rendering the resulting architectures unsuitable for widescale real-world deployment. Such a problem becomes more serious in applications with real-time performance e.g., autonomous driving, surgery robotics, etc. Sakya will present a framework to co-explore and co-design neural architectures and hardware IP to guarantee timing specifications such as latency and throughput. Given a dataset, a machine learning task, a target hardware platform as baseline, and a real-time constraint, our automation tool can find the best neural architecture and hardware design pair to maximize prediction accuracy while satisfying the required real-time performance on chip.
● MICHAEL LANZETTA is a Principal Applied ML Engineer Lead at Microsoft with over 20 years of experience on projects involving Operations Research, Optimization, Forecasting, and Machine and Deep Learning. His background in multiple industries makes him well-suited for his current role - helping Fortune 100 companies realize business transformations using the power of Machine Learning and data at scale. https://www.microsoft.com/developerblog/
● SAKYASHINGHA DASGUPTA Sakya is the CEO & Founder of Edgecortix Inc. a deep tech startup based in Tokyo and Singapore, automating machine learning driven AI hardware & software co-design for an intelligent distributed edge ecosystem. He holds a PhD from the Max Planck Institute, and has held senior R&D positions at organizations like RIKEN, Microsoft and IBM Research, with more than a decade of experience in AI, robotics and embedded hardware. Prior to founding EdgeCortix he was a senior research scientist and lead at IBM Research and more recently the CTO of a successful Singapore based AI in Fintech venture Neuri.
💙 THANK YOU
Huge thank you goes out to Microsoft Japan for having us and for sponsoring drinks! https://www.microsoft.com/ja-jp/mscorp
Become a MLT Patron and help us to keep MLT meetups like this inclusive and for free. https://www.patreon.com/MLTOKYO
Subscribe to our monthly newsletter: https://mltokyo.ai/membership-join
FIND MLT RESOURCES
MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event.
CODE OF CONDUCT
MLT promotes an inclusive environment that values integrity, openness and respect. https://github.com/Machine-Learning-Tokyo/MLT_starterkit