Task Tracking with AI
Introduction
CALO was an artificial intelligence project that attempted to integrate numerous AI technologies into a cognitive assistant. CALO is an acronym for "Cognitive Assistant that Learns and Organizes". The name was inspired by the Latin word "Calo" which means "servant of the soldier." The project began in May 2003 and lasted five years, ending in 2008. The CALO effort has had many important side effects, most notably the intelligent software assistant Siri now part of Apple iOS since iOS 5, shipping on several phones and tablets; Social Kinetics, a social app that learned personalized intervention and treatment strategies for patients with chronic illnesses, sold to RedBrick Health; the Trapit project, which is a web scraper and news aggregator that makes intelligent selections of web content based on user preferences; Tempo AI, a smart calendar; Desti, a personalized travel guide; and Kuato Studios, a game development startup.
CALO was funded by the Defense Advanced Research Projects Agency (DARPA) under its Personalized Assistant that Learns (PAL) program.[1][2] DARPA's five-year contract brought together more than 300 researchers from 25 leading commercial and university research institutions, with the goal of creating a new generation of cognitive assistants that can reason, learn from experience, know what to do, explain what they are doing, reflect on their experience, and respond robustly to the surprise SRI International was the lead integrator responsible for coordinating the effort to produce an assistant that can live with and learn from its users, provide them with value, and then pass an annual evaluation that measures how well the system has learned to do its job.
Features
CALO assists its user with six high-level functions:
Assessment
Each year, the CALO system, after living with its user for a period of time, is put through an achievement-style test of 153 "administration assistant" questions, focused primarily on what it has learned about the user's life. Raters measure how well CALO's performance improves on these questions year over year, and how much of CALO's performance is due to "learning in the wild" (new knowledge, tasks, and inferences that it has been able to acquire on its own, as opposed to function or knowledge built into the system by a developer).