Google AI Senior Researcher: Towards Machines that Perceive and Communicate
Details
** WORLD CUP 2018 = WE WILL BE FINISHED IN TIME FOR THE SEMI_FINALS **
For our July talk, we have a very special guest coming all the way from Mountain View, California.
Kevin is a contender for the most illustrious Machine Learning research scientist in the world. He graduated from Cambridge with a Bachelors in Computer Science, completed a PhD at Berkeley on "Dynamic Bayesian Networks: Representation, Inference and Learning". (of significant interest to those wishing to process sequential data).
He worked on probabilistic models for computer vision and mobile robotics at MIT before becoming Associate Professor at the University of British Columbia where he wrote the (best selling) book on Machine Learning, A Probabilistic Perspective. He joined Google in 2011 as a Senior Research Scientist in Mountain View, where he was instrumental in the acquisition and integration of Deep Mind. He has spoken at top universities across the world and has over 30,000 citations in the scientific literature (https://ai.google/research/people/KevinMurphy)
Kevin will summarize some recent work from his group related to visual scene understanding and "grounded" language understanding:
- His labs DeepLab system for semantic segmentation (PAMI'17) [1]
- His labs object detection system (CVPR ‘17 and 1st place in COCO'16) [2].
- His labs instance segmentation system (2nd place in COCO'16)
- His labs person detection/pose estimation system [3] (CVPR'17 and 2nd place in COCO'16)
- Visually grounded referring expressions (CVPR'16) [4].
- Discriminative image captioning (CVPR'17) [5].
- Optimizing semantic metrics for image captioning using RL (ICCV'17) [6]
- Generative models of visual imagination (ICLR'18). [7]