Research Directions in AI Algorithms and Systems

This first chapter is an inspiring overview of research projects from the MIT-IBM Watson AI Lab, which has recently celebrated a two-year anniversary. Lisa Amini describes three tiers of AI research spanning narrow, broad, and general AI. Currently we are practicing narrow AI, meaning performing only a single task in a single domain. Recommending an item from a list, spam filtering tools, and image recognition are examples of narrow AI. These narrow AI approaches are excellent in executing a focused complex task, but their abilities are limited in other domains. On the other hand, general AI involves unsupervised and autonomous systems. Drones guiding each other without any supervision and performing several tasks is an example of general AI. Today, the AI research community is beginning a journey into broad AI, whereas general AI is still a long-term goal for the future. The chapter also provides a sampling of IBM-MIT joint work in the areas of AI algorithms, physics of AI, ethics of AI, and AI for social good.