We invite you to join Georgia Tech and the Center for Research into Novel Computing Hardware (CRNCH) for a workshop on April 27th exploring neuromorphic computing and specifically the Field Programmable Analog Array (FPAA), a low-power neuromorphic computing device designed by ECE/CRNCH professor Dr. Jennifer Hasler and her research group. This workshop will involve hands-on experience with the FPAA device, an exploration of how algorithms and applications can be mapped to the platform, and time for a wider discussion by attendees related to common interests and research issues in the field of neuromorphic computing. Please let us know if you are able to join by emailing email@example.com and/or filling out the registration form at the link below. You are also welcome to share this opportunity with interested colleagues who are interested in either the FPAA platform or neuromorphic work in general.
The workshop was a great success, thanks to all the participants and Dr. Hasler's group! A brief write-up of the event can be found here.
Date/time: April 27th, 2018, 9 AM to 5 PM
Location: Georgia Tech Conference Center and Hotel, 800 Spring Street NW, Atlanta, Georgia, 30308
Meeting Room: Conference Room #4, 2nd floor of the GT Conference Center and Hotel
- We do not have a dedicated hotel block for this workshop, but please contact firstname.lastname@example.org if you are looking for a place to stay nearby.
- Parking will be validated for those who park in the GT conference center parking garage. Please use the entrance to the garage on Spring street (right past 5th street and Spring and at the intersection of Spring and Armstead Place).
- Lunch will be provided in the hotel dining room, providing different options for attendees with dietary restrictions.
9:00-9:10 Welcome and Introduction - Dr. Tom Conte, Director of the CRNCH Center
9:10-10:30 Introduction to the FPAA - Dr. Jennifer Hasler
- Analog computing theory brief overview
- What is the FPAA? At a high-level, what can we design better with analog vs. digital algorithms?
- Examples of FPAA classification
- Introduction to neuromorphic computing in the context of the FPAA
- What structures does the FPAA provide, and how do they work?
10:45-12:00 Getting started with the FPAA tools
- Initial overview on FPAA tools and specifics for compiling physical structures
- Hands-on work by attendees with the tools for simple examples
1:00-2:30 FPAA Further Development and Classification
- Continue working with simple examples
- FPAA classifier overview and measurement
2:45-4:00 Lightning Talks and Discussion
- Catherine Schumann, Oak Ridge National Labratory, "Evolutionary optimization (EO) training for neuromorphic systems"
- Fred Rothganger, Sandia National Labratories, "N2A: A computational tool for modeling from neurons to algorithms"
- Constantine Dovrolis, Georgia Tech, "From the Spatio-Temporal Organization of the Brain to Adaptive and Safe Lifelong Learning Machines"
- Antonino Tumeo, Pacific Northwest National Labratory, "Exploring the intersection of graph analytics and machine learning"
- Discussion and Q&A
- Where to find more details on the FPAA and how to get access to the FPAA hardware
- Further resources and engagement opportunities for attendees
Suggested Prerequisites for Attendees:
We will have FPAA hardware for use by attendees, but we suggest that attendees download the following software in preparation for the workshop:
- Attendees are encouraged to bring a laptop to run the tools VM and connect to an FPAA device via USB-A.
- The FPAA tools are provided as a VM image available from Dr. Hasler's website, and we encourage you to download the following: 1) VirtualBox to run the VM image 2) The Ubuntu 12.04 Tools image (labeled as "OVA file for Virtual Box") and 3) the VirtualBox Extension Pack to support USB usage with the VM.
- We will also plan to have 1-2 remote FPAAs available for use with the tools.
- We also encourage attendees to check out CRNCH's Rogues Gallery, which will soon host some additional FPAA devices for external users!