Interactive Machine Learning: Leveraging Human Intelligence

Channel:
Microsoft Research
Subscribers:
344,000
Published on ● Video Link: https://www.youtube.com/watch?v=-2ggKevM-_8


Duration: 1:06:14
620 views
6

Advances in processor speed and in machine learning algorithms have brought machine learning within the interactive time scale. It is now possible on many problems for a user to annotate data, train a classifier and receive feedback on the classification in 4-5 seconds. This very tight feedback loop on machine learning opens many possibilities for user interfaces that interactively manage far more information than the user can visualize. This not only changes the way we might interact with information but also changes the way we pose our machine learning problems. Having a user in a tight loop changes the distribution of training data as well as imposes constraints on training algorithms. This talk will explore these issues as well as present preliminary data on how humans and machine learning can interact.



Other Videos By Microsoft Research


2016-09-06Sensor Networks Workshop 05 - Second Generation Sensor Querying
2016-09-06Sensor Networks Workshop 05 - Life Under Your Feet: Using WSN in Soil Ecology
2016-09-06Sensor Networks Workshop 05 - Embedded Networked Sensing Redux
2016-09-06Towards Expressive and Scalable Publish/Subscribe    
2016-09-06Extending the Internet Architecture to Sensor Networks: Some Open Questions
2016-09-06India's Emerging Competitiveness
2016-09-06Towards a Service-Oriented Architecture for Reconfigurable Networked Embedded Sensor Systems
2016-09-06Extracting Product Features and Opinions from Reviews
2016-09-06Automatically proving the termination of C programs
2016-09-06From Dust to Doctors: Wireless Sensor Networks for Medical Applications     
2016-09-05Interactive Machine Learning: Leveraging Human Intelligence
2016-09-05Sound Transaction-based Reduction without Cycle Detection
2016-09-05SSCLI RFP II Capstone Workshop ΓÇô Aspect.NET - An Aspect-Oriented Programming Tool for .NET
2016-09-05SSCLI RFP II Capstone Workshop ΓÇô Typed Compilation of .NET Common Intermediate Language
2016-09-05SSCLI RFP II Capstone Workshop - A Hardware-Based CIL-Machine
2016-09-05Static Analysis for Identifying and Allocating Clusters of Immortal Objects
2016-09-05SSCLI RFP II Capstone Workshop - Parallel, Real-Time Garbage Collection in Rotor
2016-09-05SSCLI RFP II Capstone Workshop - Xtatic: Native XML Processing for C#
2016-09-05SSCLI RFP II Capstone Workshop - The Grid-Occam Project
2016-09-05SSCLI RFP II Capstone Workshop - MSIL USER MANUAL
2016-09-05SSCLI RFP II Capstone Workshop ΓÇô GCspy for Rotor


Tags:
microsoft research