AIR Distinguished Speaker Series: Some Recent Insights on Transfer and Multitask Learning

  • Starts: 10:45 am on Monday, November 1, 2021
  • Ends: 12:00 pm on Monday, November 1, 2021
Samory Kpotufe, an Associate Professor of Statistics at Columbia University, will discuss advancements in transfer and multi-task learning as part of our Artificial Intelligence Research (AIR) Distinguished Speaker Series on Monday November 1, 10:45am-12:00pm ET in Room 101 at the Rajen Kilachand Center for Integrated Life Sciences and Engineering, located at 610 Commonwealth Avenue, Boston, MA. A common situation in machine learning is one where training data is not fully representative of a target population due to bias in the sampling mechanism or due to prohibitive target sampling costs. In such situations, researchers aim to ’transfer’ relevant information from the training data (also known as source data) to the target application. Kpotufe argues that much of the traditional notions of 'distance' between source and target domains can yield an over-pessimistic picture of transferability. Instead, new notions of 'relative dimension' between source and target uncover a rich set of situations where transfer is possible even at fast rates. Please register in advance on Eventbrite at the link below.
Rajen Kilachand Center for Integrated Life Sciences and Engineering, 610 Commonwealth Avenue, Room 101, Boston, MA 02215

Back to Calendar