Deep Learning

Understanding and using deep learning networks

MathWorks at NeurIPS 2020

Post by Shounak Mitra, Product Manager – Deep Learning and AI
We are very close to the start of the biggest AI conference of the year! Last year when I left Vancouver after attending NeurIPS 2019, I had marked a few restaurants to try out in 2020. Little did we all know the unprecedented times we all will have to face.
MathWorks is proud be a sponsor at the conference and this year we are doing a bunch of activities at the virtual booth, meetups, socials, and workshops. I am writing this blog to share what all MathWorks is doing at NeurIPS 2020, and encourage you to join us if you can!

Virtual Booth

If you have used or are currently using MATLAB or Simulink, or looking for your next job opportunity, come say a ‘hi’ to us at the MathWorks Virtual Booth. We will have a team of developers, hiring managers, and product managers who’ll be more than happy to chat with you.
Talks at the booth: We are arranging 11 Talks at our booth this year. The talks will be on a variety of topics, ranging from deploying AI to embedded devices to how researchers at UMASS Amherst have used deep learning to mimic physics behind finite element modeling. There will also be booth talks on other interesting topics such as LiDAR, how MATLAB works with TensorFlow and PyTorch, AI workflows for Radar, Signal, Time series and Reinforcement learning. For a detailed agenda, please see the full agenda.
 

Tuesday – Dec 8

Wednesday – Dec 9

Thursday – Dec 10

2:10 - 2:30 PM
AI System Design - Simulation and Code Generation Bill Chou & Emmanouil, Product Managers
Application of GANs to Improve Characterization of Porous Media Microstructure and Fabric
Tim Anderson, Stanford
AI for Time-series and Signal Applications   Gabrielle Bunkheila & Esha Shah Product Managers
2:30-2:50 PM
AI System Design - Simulation and Code Generation (Continued)
 
 
On the Effectiveness of Bayesian AutoML methods for Physics Emulators
Peetak Mitra, UMASS
Reinforcement Learning with MATLAB and Simulink
Emmanouil, Product Manager
7:10-7:30 PM
Lidar Perception: Semantic Segmentation and Object Detection on 3D Point Clouds
Avi Nehemiah, Product Manager
AI workflows for Radar and Communications
Rick Gentile, Product Manager
AI Workflows in the Cloud
David Willingham, Product Manager
7:30-7:50 PM
Evolution of Deep Learning with MATLAB David Willingham, Product Manager How MATLAB works with TensorFlow and PyTorch Shounak Mitra, Product Manager Quantizing Deep Learning Networks for GPUs and FPGAs with MATLAB Greg Coppenrath, Product Manager
Hiring at the booth: The booth will also have private chats scheduled with our recruiters and hiring managers. To explore a list of opportunities, please go to Careers at MathWorks.

Paper at NeurIPS

This year we have a paper at NeurIPS: On the Effectiveness of Bayesian AutoML methods for Physics Emulators. The authors of this paper will also give a talk at the conference. In this work, a data driven emulator for turbulence closure terms in the context of Large Eddy Simulation (LES) models is trained using Artificial Neural Networks and an autoML frame- work based on Bayesian Optimization, incorporating priors to jointly optimize the hyper-parameters as well as conduct a full neural network architecture search to converge to a global minima, is proposed.
To watch the recording of the talk, please click here.

 

NeurIPS 2020 Socials

We are organizing three socials at NeurIPS 2020, and two at WiML (Women in Machine Learning) workshop co-located with NeurIPS.
Title When Contact Link
Equity and Ethics in AI from the Perspective of Black Women in AI Tuesday December 8, 5-7pm EST Wednesday December 9, 8-9am EST (WiML) Louvere Walker-Hannon sign-up
Un-Bookclub Race After Technology Wednesday December 9, 8-9am (WiML) Wednesday December 9, 5-7pm EST Anoush Najarian sign-up
Women in AI Ignite Thursday December 10, 3-5pm EST Anoush Najarian sign-up
NeurIPS registration is needed to attend socials. We also plan to also make these available through the Boston Meetup.

NeurIPS 2020 Meetups

The goal of NeurIPS meetups is to open up access to communities worldwide and to connect participants by geographic area and language. With this, NeurIPS hopes to support the growth of AI expertise around the world, including in underrepresented communities in tech, and to fuel innovation responsibly.
We are excited to have two MathWorkers on the NeurIPS organizing committee: Louvere Walker-Hannon and Anoush Najarian are NeurIPS meetup co-chairs, working with Emtiyaz Khan of the Riken Center for AI in Tokyo.
MathWorkers are helping organize NeurIPS meetups on three continents! These are free and open to the public.
 
|
  • print
  • send email

Comments

To leave a comment, please click here to sign in to your MathWorks Account or create a new one.