The Moments in Time at MIT-IBM Laboratory

The Moments In Time is an ongoing project at the MIT-IBM laboratory by Dan Gutfreund - a principal investigator at MIT-IBM Watson AI laboratory and Aude Olivia - a principal research scientist at MIT computer science and AI laboratory.

Understanding The Moments In Time dataset, it is built on a compilation of 1 million annotated video streams of rapidly transforming events unrolling within three short seconds. The fact backing up this research is a well-known trait of humans.

For humans, understanding an event depicted in various forms is a pretty straightforward capability. Machines are literally a failure in this case. The final outcome of this research initiative is to teach machines to recognize different actions. Teaching the deep-learning algorithms to recognize actions by providing them with a vast coverage of dynamic visual and auditory moments.

For example, cooking, opening, running actions are fed to the deep-learning algorithm with the minimum of 2000 videos to understand those actions without having to be compulsorily taught under supervision.

IBM #ArtificialIntelligence #MachineLearning #AI #ML