Deep learning needs cognitive memory and vice versa. Intelligence, whether natural or artificial, must be defined by more than one approach. In complementary learning, both forms work together to build a more complete AI system. Bruce Horn explores Intel’s Saffron’s cognitive approach, which provides one-shot learning using associative and episodic memories and is more appropriate for individual and dynamic patterns.
This session is sponsored by Intel Saffron.
Bruce Horn is an Intel fellow and chief technical officer for the Intel Saffron cognitive solutions group, where he is responsible for driving new applications and uses for Intel Saffron’s memory-based reasoning system, a fundamentally new approach in the development of intelligent devices and systems. Previously at Intel, Bruce built a team to develop advanced conversational interfaces; that team provided the spoken language technology and mobile application for the Oakley Radar Pace running and cycling coach. Prior to joining Intel, Bruce was a principal research software development engineer at Microsoft, where he worked on the creation and deployment of natural language systems for Bing; worked at Powerset, where he was responsible for the computational infrastructure of the Powerset Natural Language Search System; worked at Apple, where he created and developed the Macintosh Finder, the first widely used desktop graphical user interface, among other components of macOS; and served as a member of the learning research group at the Xerox Palo Alto Research Center, where he contributed to several implementations of the Smalltalk virtual machine. Bruce holds a BS in mathematical sciences from Stanford University and an MS and PhD in computer science from Carnegie Mellon University.
©2017, O'Reilly Media, Inc. • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • firstname.lastname@example.org