Duration: 14:49

534 views

2

Adam Kalai of Microsoft Research New EnglandΓÇöin association with Serge Belongie of the University of California, San Diego; Ce Liu of Microsoft Research New England; George Pierrakos of University of California, Berkeley; Ohad Shamir of Microsoft Research New England; and Omer Tamuz of the Weizman Institute of ScienceΓÇöexplains the research focus and issues with machine learning and crowdsourcing. Much of machine learning works with similarity: how similar two separate items are to one another. Having computers create associations or similarities requires context at times, as in the use of numbers. People might look for shape similarities, while the values of numbers are vastly different to the computing system.