Intel Labs demonstrated Kapoho Point and Lava v0.5 for the first time.
The recent Intel Innovation 2022 (ION) event featured several neuromorphic innovations with both a Kapoho Point demo booth and a hands-on Lava lab session. At the booth, we announced Intel Labs' first deliveries of the 8-chip Loihi 2 boards to INRC members.
Kapoho Point advances Intel’s goal of bringing neuromorphic technology to commercial applications by providing a compact form-factor suitable for low-latency edge devices and a stackable design suitable for high performance computing in the lab. Built with Loihi 2, Intel’s latest generation neuromorphic computing architecture, Kapoho Point enables developers to accelerate AI models with up to 1 billion parameters or solve optimization problems with up to 8 million variables.
Our ION booth and lab demonstrated a 9-layer PilotNet deep learning sample workload running on a Kapoho Point system in real-time. We showcased how the modular, open-source software framework Lava makes it easy to program neuromorphic models and how Loihi 2’s new graded spike and convolutional compression features, both now supported in the Lava v0.5 release, make deep learning inference on Loihi hardware like Kapoho Point more efficient than ever.
Accenture’s Andreea Danielescu joined our team in the Innovation Zone to share her views on the role that next-generation neuromorphic architectures will play in intelligent edge computing, powering applications from extended-reality headsets for the Metaverse to intelligent mobile robots. As an integral member of the INRC, Accenture’s research on improving the developer experience for neuromorphic computing was right at home in the Dev-focused ION event.
I also had the pleasure of discussing the future of our neuromorphic research program with Intel CTO Greg Lavender and Intel Labs Director Rich Uhlig. Look out for a recording of those conversations to come.
If you’d like to learn more about running your own applications on Kapoho Point, visit https://intel-ncl.atlassian.net/wiki/spaces/INRC/overview
Join us Tuesday, November 15, 2022 @ 9:00-10:30am PST / 18:00-19:30 CET for another great INRC Forum, this time featuring an update on optimization at Intel and George Mason University!
Agenda
Intel Labs introduction and update on lava-optimization.
Research Talk by INRC member Maryam Parsa of George Mason University.
Sumedh Risbud (Deactivated) hosts Lava Open-Door Q&A. Ask questions and get feedback from Lava developers.
Eventifying and Applying Bayesian Optimization in Lava
Abstract: Despite the monumental success of deep learning over the past few decades, researchers have yet to find a solution that overcomes the inherent fragility of machine learning models when facing adversarial attacks and stochastic environments. A variety of online, continual learning methods have been proposed to address these limitations with varying levels of success and cohesion. We introduce the first phase of the Eventified Bayesian Optimization (EBO) project implemented in Lava-Optimization, which aims to provide a generalized framework for eventified hyper-parameter optimization and lifelong continual learning applications to the neuromorphic research community. Here, we present an outline of the project over the next few years and highlight the completion of the first phase that implements a Bayesian optimization on Lava.
Bio: Maryam Parsa is an Assistant Professor in Electrical and Computer Engineering (ECE) department at George Mason University (GMU). Prior to joining GMU, she was a Postdoctoral Researcher at Oak Ridge National Laboratory (ORNL). She received her PhD in ECE from the Center for Brain Inspired Computing (C-BRIC) at Purdue university in 2020. She was the recipient of several prestigious awards including a four-year Intel/SRC PhD fellowship, ORNL ASTRO fellowship, Purdue university Ross fellowship, TECHCON'18 student presenter award, and ICONS'21 best paper award. She also worked at Intel Corporation in 2014 and 2016 as research intern. Dr. Parsa's research interests are in the areas of neuromorphic computing, neural architecture search and Bayesian optimization/learning across the full stack of materials, devices, circuits, algorithms, and applications.
How to join:
INRC Members can find the meeting link on INRC Fall Forum Schedule. If you are interested in becoming a member, join the INRC.
Intel Labs announces the next INRC Workshop on Nov. 2nd and 3rd, 2022
An explosion of Lava! In case you haven’t heard, Lava has made tremendous progress over the last few releases. It’s now possible to develop a wide range of networks for Loihi 2, from feed-forward sparse DNNs to optimization solvers. Join us for our next INRC workshop to get hands-on with Loihi 2 and Lava and connect with other leading researchers. This upcoming event will provide members of the INRC the latest insights, analysis, and outlook on neuromorphic computing.
Key Information
What: Research and technical talks from Intel Labs and members of the INRC.
When: November 2nd and 3rd, 2022 - Time TBD
Save-the-dates:
Registration details will be shared as soon as possible.
Where: Virtual talks on Microsoft Teams
What else:
This workshop will be open to current INRC members. If you’re not a member, please see Join the INRC or email inrc_interest@intel.com to learn more.
Given sufficient demand, Intel Labs may offer in-person “bootcamp” training sessions in northern California, Oregon, and Munich area. Complete the registration survey below to let us know if you’d like to attend in person.
Take the survey and register today
Please register to receive links to attend the workshop and share your opinions on what matters to you:
Questions or feedback? Email inrc_interest@intel.com or leave a comment below.
Are your teammates or colleagues interested in neuromorphic computing? Invite them to attend!
