We tend to think that the line of work on multi-modal learning, i.e., combining various kinds of inputs (text, audio, image, video) during learning, seems like the next potential breakthrough in …
( read original story …)