078 - Machine teaching with Dr. Patrice Simard
Microsoft Research Podcast

If at school you struggle with getting the right answer because you can never agree on what the question is asking, research might be for you. Changing the question and trying to answer it in a slightly different way is what Dr. Patrice Simard does at Microsoft for his job.

Machine learning is where a computer learns to extract knowledge from data, without being told what the knowledge is. Machine teaching is all about a machine being able to learn from a human teacher, but without any other data. When humans teach another human, the process looks very different from when a human teaches a machine, but can we teach a machine like we teach another human?

In order to perform "machine learning", we must first collect a large dataset and teach the machine to distinguish that data, learning what makes for the right answer in the process, based on our guidance. We teach the machine to memorize and to pattern-match. Humans don't learn like that. Instead, we learn to ask the right questions about the data, and use those to make choices.

If we could skip the costly dataset training and instead use the labeling directives to teach the machines to understand, it could save a lot of money and make machine learning more efficient.

powered by SmashNotes