Series explores Deep Machine Learning
By Sharon Henry
Padhraic Smyth, professor and Director of UCI Data Science
Initiative, speaking at Igniting Technology: Deep Machine Learning,
A new frontier in artificial networks
Irvine, June 3, 2016 — Deep Machine Learning, the modern buzz term to describe artificial neural network models inspired by the real-lfe biology of the brain, was the featured topic at the recent Igniting Technology event held Thursday, June 2 at Calit2 Irvine.
G.P. Li, director, Calit2 Irvine, welcomed the standing-room-only audience to the evening’s event. The program included five presentations and an audience Q&A moderated by Michael Guiliana, partner at Knobbe Martens the intellectual property law firm that supports the series.
(Speakers’ presentations can be downloaded below)
DEEP MACHINE LEARNING: A NEW FRONTIER IN ARTIFICIAL NETWORKS
UCI professor and Director of UCI Data Science Initiative, Padhraic Smyth, opened the program with examples of machine learning, describing the applications as “mapping raw inputs to semantic output.” He also noted the dark side of deep learning. Not only do the algorithms require massive amounts of training data, but true experts are hard to find,” he said.
Ali Arsanjani, IBM Distinguished Engineer, CTO, Analytics and Enterprise Content Management offered an overview of IBM Watson and the abilities of cognitive systems to analyze text. Watson is a cognitive system that can tease apart the human language to identify inferences between text passages. It can demonstrate human-like accuracy, perform at speeds far faster, and at a scale far bigger than a human, he explained.
Larry Williams, Director of Product Management, ANSYS, Inc, provided examples of how a digital twin – a virtual or digital representation of a physical machine – is being used by industries to enable troubleshooting and predictive maintenance.This approach promises cost savings over routine or time-based preventive maintenance, because tasks are performed only when warranted.
(Left) Mike Guiliana, moderates Q&A portion of Igniting Technology event.
(Right) L to R, Ron Schoenbaum, Mark S. McNally, Larry Williams, Ali Arsanjani
Ron Schoenbaum, a partner with Knobbe Martens, contributed to the discussion by explaining the types of machine learning ideas can be patented, and factors to consider when seeking patent protection.
The final speaker, Mark S. McNally, CEO, UBIX, shared his experiences heading a startup in the big data space. UBIX makes predictive analytics software, known as Auto-Curious. McNally explained how UBIX software is being used to search for connections between genetic mutations and cancer causes. This involved limitless possibilities to analyze across large data sets, using exploration that needed automation. “The results of this type of technology will empower the next generation of cancer cures which will be individually tailored, more impactful, and far less risky to the patient,” he said.
McNally advises startups to “carve out their defensible niche” and avoid “deep cave development. “It’s an incredibly exciting time for startups in this space, commercial and exit opportunities are broader than people realize,” he said.
___
Igniting Technology programs are held twice a year, and are sponsored by Knobbe Martens in partnership with Calit2 Irvine. Since the series launched in April 2006 the two organizations have sought to bring together experts from academia, business, science and the law to deliver unique insights and experiences to the technology community at UC Irvine.
IGNITING TECHNOLOGY PRESENTATIONS
(Not all presentations may be available)
Mike Guiliana
Padhraic Smyth
Larry Williams
Ron Schoenbaum
Mark S. McNally