Increasingly, industry leaders wish to leverage data-driven models (AI and ML) and well-calibrated simulations to lower R&D costs by replacing expensive laboratory measurements, particularly in the materials design and manufacturing business. The vision is good, but the reasoning is often flawed. Building data-driven models and eliminating experiments are often mutually exclusive ideas in scientific discovery. The strength of data-driven models is directly proportional to the amount and quality of data that they are trained on, and experiments are where the data are produced.
Thus, this vision of simulation and AI driven exploration and design relies on a significant change to the status-quo in many laboratories. Of course there are shining exceptions, but in our exposure to many industrial chemistry and bio-sciences labs, we have consistently seen opportunities to increase the lab throughput by 10, 100, or even 1000x with similarly lower costs per sample. However, when suggested to an industry scientist, these ideas are often perceived as too expensive or impractical. Some of this is a skills gap. Some of it is an imagination gap. And, much of it is an examples gap. Every lab that these scientists have ever seen is designed with significant manual work to formulate samples that are then shipped out to multiple buildings/ laboratories for manual characterization.
The goal of this project is to develop tools, methodologies, and examples that can bridge this gap for scientists so that they embrace instead of be wary of these efforts. There are numerous mental and design shifts needed to achieve these goals, and so the primary goal of the First Congress is to identify and bring together key participants from academia, industry, and government to discuss experimental needs across these sectors, share information on the state of the art, and determine a path for how this community can collaborate, be governed, and produce meaningful results towards shared goals going forward.
First Congress Schedule
February 20, 2020
7:30 - Breakfast
8:30 - Opening Welcome by Roger Bonnecaze, University of Texas at Austin
8:35 - Keynote Speech by Eric Jones, Enthought
1000x Lab: The Rationale, Enablers, and Direction
9:15 - Industry Problem 1 by William Hartt, Proctor & Gamble
10:00 - Break
10:30 - Industry Problem 2 by Mark Simon, Saint-Gobain
11:15 - Industry Problem 3 by Peter Soler, Bristol-Myers Squibb
12:00 - Lunch break
13:00 - Michael Heiber, Enthought
13:20 - Eric Furst, University of Delaware
13:40 - Zachary Trautt, NIST
14:00 - Filippos Tourmomousis, MIT
14:20 - Michael J. McCarthy, University of California, Davis
14:40 - Martin Green, NIST
15:00 - Break
15:30 - Panel Session and Q&A: Industry Problems
16:15 - Panel Session and Q&A: Example Solutions
17:00 - Measurement Accelerator Poster Session and Happy Hour
18:00 - Dinner at El Mercado
February 21, 2020
8:00 - Breakfast
9:00 - Discussion Session Framing by Roger Bonnecaze, University of Texas at Austin
9:15 - Discussion Session: Building The 1000x Lab
10:30 - Break
10:45 - Discussion Session: Building The 1000x Lab Continued
12:00 - Lunch break
13:00 - Summary of discussion session
14:00 - Closing Remarks and Governance
15:00 - End of Congress
Meet the Community
Founder/CEO of Enthought
Eric holds both a Ph.D. and M.S. in Electrical Engineering from Duke University and a B.S.E. in Mechanical Engineering from Baylor University. Widely known as one of the founding fathers of Python’s scientific community, Eric drives business growth through digital transformation.
For more than 15 years, he has been an innovator in applying machine learning, image processing, 3D visualization, and parallel computing to elegantly solve the most complex business problems.
Prior to founding Enthought in 2001, Eric focused on numerical electromagnetics and genetic optimization in the Department of Electrical Engineering at Duke University.