The First Workshop on Unorthodox Computing
September 24, 2011
Location: University of Southern California, Ming Hsieh Department of Electrical Engineering, Room 248
As you well know, most digital CMOS computation is carried out in what is informally thought to be a deterministic manner, in that we expect to get the same answer running the same program on computers A and B. But, new computational fabrics are being created that contain billions of individual elements that are adversely affected by factors such as noise, process variations, material defects and chemical impurities. These fabrics, though imperfect, are attractive since they often provide for enormous gains in such areas as decreasing power and increasing performance. But these gains come at a price, such as lower yield and/or nondeterministic results. To exploit the benefits of these new fabrics, while accepting their inherent low yield and abnormalities, new models of computational are being developed along with the identification of problems that are amenable to stochastic rather than deterministic outcomes.
As a consequence of these and other factors, the organizers of this workshop are attempting to bring together a diverse yet synergistic group of researchers whose work may have great benefits to those working in related domains. Our focus is on computation and information with particular emphasis on: fabrics including CMOS and non-CMOS; circuits including nano, bio and molecular; architectures and systems; and applications such as multi-media and drug delivery.
- Mel Breuer
- Sandeep Gupta
- Antonio Ortega
- All professors of Electrical Engineering at the Viterbi School of Engineering, USC
- Fabrizio Lombardi (North Eastern U.)
- Keshab Parhi (U. of Minnesota)
- Gene Frantz (Texas Instruments)
- Shih-lien Lu (Intel Corp)
- Bernd Becker (Albert-Ludwigs-University)
- John Hayes (U. Michigan)
- Naresh Shanbhag (U. Illinois)
- Abhijit Chatterjee (Geogia Tech)
Viterbi School of Engineering
- Hotel 1
- Hotel 2
- Hotel 3
Venue & Location
University of Southern California Ming Hsieh Department of Electrical Engineering 3740 McClintock Avenue, EEB 248 Los Angeles, CA 90089
PSX, Gate 3?
Revised Agenda 7/6/11
The First Workshop on Unorthodox Computing
September 23-24, 2011
Held at the University of Southern California Los Angeles, CA 90089
Sponsored in part by the Viterbi School of Engineering and the NSF (?)
Friday, September 23, 2011
6:00-8:30pm Optional Dinner for those who arrive early. Please RSVP by _______.
8:30pm -- … Socializing
Saturday, September 24, 2011
7:30 - 8:00am Breakfast
8:00 - 10:00am Opening Remarks, Keynote and Invited Talks
a) Welcome and opening remarks, Melvin A. Breuer, USC (5 minutes) b) Keynote: (55 minutes) Theme: Unorthodox Computing: From Fabrics to Applications Exact title and speaker TBA Abstract: The steering committee expressed a desire for one major keynote that addresses most of the themes to covered by this workshop. Since no one individual may feel comfortable covering such a wide range of topics,possibly steering committee members could consult with this speaker and suggest material to be included in this talk. Assignment: Program committee c) Invited talk A: (25 minutes) Theme: This talk would address one of the key themes in more detail, such as unorthodox fabrics, circuits, architectures or related mathematics. Exact title and speaker TBA Assignment: ???? to select speaker d) Invited talk B: (25 minutes) Theme: This talk would address another of the key themes in more detail, such as unorthodox fabrics, circuits, architectures or related mathematics. Exact title and speaker TBA Assignment: ???? to select speaker
10:00 - 10:15am Break
10:15am - 12:15pm Who are we and what do we do?
At this time we will have about 20 to 25 five-minute presentations by some of the attendees. We will organize these short presentations into four or five groups, each group having about five presentations. Each group will have a common theme, such as fabrics, or architectures. One can think of the speakers in each group as being panelist. Their presentations will be primarily focused on their own research, including: (slide 1) who they are - who are their colleagues and -where are they from; (slide 2) what are the main issues that they are addressing, why are these important and why is their work unorthodox; (slide 3) what are the main technical issues they have solved to date and are currently trying to solve; (slide 4) what important breakthroughs are needed in their or other domains that will aid in the success of their work; (slide 5) 2 key references related to their work and 2 reference related to the general field they are addressing. (We hope to have these short presentations available about one week prior to the workshop so they can be distributed and attendees will have a chance to familiarize themselves with some of this material.)
12:15 - 1:30pm Lunch.
Invited talk C. 12:45-1:15pm. Exact title and speaker TBA. (Or maybe we will have other attendees give 5-minute presentations so we all can get to know each other.
Breakout Sessions (working groups)
1:30 - 2:15pm
Fabrics Computation Discussion Leader TBA TBA Scribe/Reporter TBA TBA
2:15 - 3:00pm
Circuits Architectures, algorithms and applications Discussion Leader TBA TBA Scribe/Reporter TBA TBA
Here we plan to have a pair of two parallel open-mike sessions where the discussion leaders attempt to have many people agree on common or important themes. The panelists will attend their respective session for part of the time, while others are free to go to any session they please.
3:00 - 3:30pm Working group summaries: Each discussion leader along with their panelist and any other volunteers attempt to summarize the results from their session.
3:30 - 3:45pm Break in parallel with summary development
3:45 - 5:00pm Working Group Presentations and Discussion Presentations (say 4 reports of 5 minutes each) to attendees of summaries, conclusions, key themes and recommendations emerging from the working groups. This will be followed by an open-mike discussion with the focus on elaborating on these findings as well as integrating ideas into a more cohesive document.
5:00 - 5:30pm Session leaders revise their reports w.r.t. the previous discussion.
5:30 - 6:00pm Closing Session
• Summarize emerging themes • Identify next steps and assign tasks • Finalize report (via on-line discussions) and dissemination • Outreach to funding agencies and relevant companies • Outreach to research community • Follow-on workshop: If and where/when?
6:00 - 7:00pm Light dinner
Do we ask for 8 slides to be sent in advance?