News

Successful Tech Transfer

In a pre-print of a 2020 paper for EPEPS, IBM revealed that they use CAEML’s BALDO algorithm to optimize the equalization settings for their Power9 processor.

CAEML Grad Student Joshua Hanson to present at L4DC

11 June 2020 Illinois CAEML Graduate Student Joshua Hanson will be giving an oral presentation of his paper Universal Simulation of Dynamical Systems by Recurrent Neural Nets at the 2nd Conference on Learning for Dynamics and Control (L4DC)  on June 11, 2020. There were 131 submissions for the conference and Hanson’s was one of only 14 were selected for oral presentations.

Raginsky presents video lecture for IAS

19 May 2020  CAEML PI Maxim Raginsky presented a video lecture for the Institute for Advance Study School of Mathematics entitled: Neural SDEs: Deep Generative Models in the Diffusion Limit. The lecture is available on the IAS website.

Bowen Li, Paul Franzon at CAEML NCSU Win Best Paper Award at DesignCon 2020

15 May 2020 Congratulations to CAEML researchers graduate Bowen Li, NCSU student Brandon Jiao and their advisor Paul Franzon, along with Geoff Zhang and Chih-Hsun Chou at CAEML Industry member Xilinx  for receiving a Best Paper Award at DesignCon 2020 for their submission Self-Evolution Cascade Deep Learning for SerDes Adaptive Equalization!

NeurIPS poster presentation by CAEML student

28 Oct 19 CAEML Graduate Student Joshua Hanson will be presenting a poster at NeurIPS 2019 the Thirty-third Conference on Neural Information Processing Systems December 8-13 in Vancouver Canada. NeurIPS is one of the top machine learning conferences with a highly competitive acceptance rate, this year about 21%. Hanson is a a student of CAEML Faculty member Maxim Raginsky.

CAEML Director Rosenbaum Presents Keynote at MWSCAS

4 Sept 19 In August, CAEML director Elyse Rosenbaum presented a keynote lecture “Machine Learning for Design Efficiency” at the Midwest Symposium on Circuits and Systems (MWSCAS 2019) in Dallas, TX. A copy of her presentation slides may be found here.

CAEML Center PI Paul Franzon on Panel at 2019 IEEE International Microwave Symposium

15 July 19 A Panel on “AI Friend or Foe” was presented at the 2019 IEEE International Microwave Symposium on June 4. In addition to CAEML PI and NCSU Site Director Paul Franzon, panelists included:Taylor Hogan, Cadence Design Systems

  • Thomas Rondeau, program manager, Defense Advanced Research Projects Agency (DARPA)
  • Sankalp Modi, MathWorks
  • Ron Rohrer, Southern Methodist University
  • Moderators: Osama Shanaa, MediaTek, and Prof. Francois Rivet, University of Bordeaux

A recap of the panel is available on the EET Asia website https://www.eetasia.com/news/article/AI-Friend-or-Foe

Design Con Interview of Paul Franzon

7 March 19 CAEML NCSU Site Director Paul Franzon was interviewed at Design Con by Sierra Circuits about his work in machine learning in the electrics industry and CAEML research and projects. The interview video is available through YouTube.    https://youtu.be/PRKUKqMJGl8

CAEML Director Rosenbaum on ISQED Syposium 19 Panel

6 Feb 19 CAEML Director Elyse Rosenbaum will be part of a panel at  20th International Symposium on Quality Electronic Design (ISQED) in Santa Clara, CA March 6 and 7th. The panel “Hype or hope: Is machine learning the next generation of design and design automation?” also features Mark Ren, NVIDIA, Noel Menezes, Intel, Pradiptya Ghosh, Mentor Graphics and Sachin Sapatnekar University of Minnesota. A full announcement about the conference is available from The Daily Telescope.

CAEML at DesignCon – a New Machine Learning Track

19 Dec 18  See the article and announcement from CAEML Industry Advisory Board Chair Chris Cheng, HPE here https://www.designnews.com/content/machine-learning-designcon-2019-get-details/45682803959962

Hakki Torun wins Best Student Paper Award at EPEPS 2018

19 Oct 18    Hakki Torun, graduate student at Georgia Tech working on the CAEML project Optimization of Power Delivery Networks for Maximizing Signal Integrity, won the Best Student Paper Award at the IEEE 27th Conference on Electrical Performance of Electronic Packaging and Systems (EPEPS) held October 14-17 in San Jose. The paper “Bayesian Active Learning for Uncertainty Quantification of High-Speed Channel Signaling” was coauthored by CAEML Industry members Jose Hejase, Junyan Tang and Dale Becker from IBM, and GaTech CAEML Site Director Madhavan Swaminaathan.

CAEML in the News:

1/31/18   EE Times    AI Expands Role in Design: EDA’s CAEML grows more humps

Machine Learning Offers Helping Hand To Edit Chips  Electronic Design July 6, 2017

3/23/17 – The most recent CAEML UPDATE newsletter is now available!

AI Tapped to Improve Design:  EE Times  2/3/17