Last week MassTLC brought together data scientists, analysts, software engineers, and others who focus on machine learning and artificial intelligence in their companies and organizations for our second annual MassIntelligence conference.
The morning began with a keynote from Rohit Prasad, Vice President and Head Scientist at Amazon Alexa. Rohit provided an in-depth look into the complexities of how Amazon developed seamless ambient computing. Using pervasive deep learning, they took advantage of the infinite quantities of data in AWS data sets. With this they were able to process so quickly that the latency in which Alexa responds is low.
Additionally, Amazon provided two open APIs to developers. The first is the Alexa Skillset, in which organizations build a skill to connect specifically with their end customer. The second API is Voice Service, in which organizations can build Alexa into their own devices, such as Sonos. These APIs have led to Amazon Echo having 325K developers in over 190 countries, 50K skills, 100 different devices types, and 87K personal skills.
Our second keynote of the day, Lisa Amini, Director of IBM Research Cambridge, moved away from the consumer-focused uses towards the work happening in the research labs with the next generation of technologies geared towards the enterprise. Lisa led us through a number of different ways they have been teaching the machine to learn, including: Human inspired visual AI (ImageNet); Neuro-symbolic Program Induction (teaching the machine to recognize objects and orientations); and Neuro-symbolic Modular Network (teaching the machine to recognize objects and words).
MassIntelligence also featured three “tech talks” that gave a high level overview of machine learning and AI in specific industries. Heather Ames, of Neurala provided a glimpse into working with neural networks;
Jeff Fried, of InterSystems and Maciej Pacula of Mass General Hospital spoke about using machine learning as a more effective and efficient way to diagnose and treat cancer; and Ed Anthes-Washburn from the New Bedford Port Authority gave us insights into how one of the most traditional industries in the world is turning to data for new ways of monetization.
Rounding out the day were four breakout sessions covering: adoption and utilization of machine learning tools and processes in the enterprise; connected intelligent cities; mobility and transportation; and the ethics around AI.