REDMOND, WA, MARCH 14, 2022 – In a major reveal at the Future In Review event last week, Pattern Computer®, Inc. (PCI) disclosed that it has made a major breakthrough in explainable AI (XAI). Pattern Computer’s proprietary Pattern Discovery Engine™ can derive previously hidden local manifolds from neural net models (including deep neural nets).

While methods for a partial picture of what deep neural networks learn have been around for years1 they don’t actually reveal how the neural network made its decision—the other methods only illustrate which features may have been involved in the neural network making a decision.

According to Dr. J. Ben Brown, Computational Biologist and Staff Scientist in Environmental Genomics and Systems Biology at Lawrence Berkely National Lab, “The major advance here for PCI is that we now actually see – through the eyes of the neural network – precisely the decision function it has learned. That’s a sea change.”

Referring at FiRe 22 to PCI’s solving the “black box” problem shared by all neural nets, Brown noted “That’s what Pattern Computer has been able to do – to turn any deep learning AI algorithm into a ‘glass box.’ XAI has the potential to be as impactful as the arrival of calculus. We are coming to an inflection point, and that is XAI, and that is Pattern Computer. This is completely new and novel. Nothing that is out there today is explainable AI.”

Pattern Computer’s breakthrough enables many scientific approaches, and will impact society in interesting ways. Mark Anderson, Pattern Computer’s CEO, points out, “There is significant direct interest in the application of our breakthroughs to the EU’s GDPR law, which requires firms to explain to their customers how their AI-mediated decisions were decided. This law has been unenforceable until today, because it has not been possible to understand or explain how deep neural networks make decisions. Now, we can: Pattern Computer’s new XAI system can extract the actual decision functions from neural networks, and those decisions can be explained and studied.”

About Pattern Computer

Pattern Computer, Inc. uses its proprietary Pattern Discovery Engine™ to solve the most important and intractable problems in business and medicine. Its proprietary mathematical techniques can find complex patterns in very-high-order data that have eluded detection by others.

While the company is currently applying its computational platform to the challenging field of drug discovery, it is also making Pattern Discoveries for partners in other sectors, including additional biomedical research, materials science, aerospace manufacturing, veterinary medicine, air traffic operations, resource extraction and finance.

CONTACT: Brad Holtz – 301.529.9944 – bh@patterncomputer.com

The foregoing contains statements about the Pattern Computer’s future that are not statements of historical fact. These statements are “forward looking statements” for purposes of applicable securities laws and are based on current information and/or management’s good faith belief as to future events. The words “believe,” “expect,” “anticipate,” “project,” “should,” “could,” “will,” and similar expressions signify forward-looking statements. Forward-looking statements should not be read as a guarantee of future performance. By their nature, forward-looking statements involve inherent risk and uncertainties, which change over time, and actual performance could differ materially from that anticipated by any forward-looking statements. Pattern Computer undertakes no obligation to update or revise any forward-looking statement.

Copyright © 2022 Pattern Computer Inc. All Rights Reserved. Pattern Computer, Inc., Pattern Discovery Engine, and ProSpectral are trademarks of Pattern Computer Inc. or its subsidiaries. Other trademarks may be trademarks of their respective owners.

1Shapley values, feature importance measures, relevance attribution, saliency maps, partial dependence plots