Shared publicly  - 
Live and Learn: How Big Data and Machine Learning Power the Internet

This week, Stanford University hosted the NASA Innovative Advanced Concepts (NAIC) 2014 Symposium. As part of NASA's Space Technology Mission Directorate (, the NAIC supports early studies of visionary concepts that may enable new missions or advances for use in NASA's future missions.

+Peter Norvig, a Director of Research at Google and former Head of NASA Ames's Computational Sciences Division, was on hand to present Live and Learn: How Big Data and Machine Learning Power the Internet

In the talk, Peter provides a qualitative overview of the different ways in which data enables machine algorithms to classify, learn, and discover information, allowing advances in fields such as Natural Language Understanding (NLU), machine translation, computer vision, and more.

After the ~30 minute talk, Peter stayed for a Q&A session with the audience, fielding questions covering topics such as online education (MOOCS), speech recognition, neural networks, and theoretical use of statistical models to radically different applications. 

The presentation begins at the 21:20 mark of the recorded livestream, linked below, with the Q&A session beginning at starting at ~47:50. To learn more about the NAIC, visit
Jay Falker, Peter Norvig, Director of Research, Googleon NIAC2014 on Livestream - Watch live streaming Internet TV. Broadcast your own live streaming videos, like NIAC2014 in Widescreen HD. Livestream, Be There.
Vlad S's profile photoMisbah Ashraf's profile photoJirka Daněk's profile photoKrishna Sankar's profile photo
Excellent so it seems the robots are for mapping and localization work
They are replacing and supporting human decision with automated algorithms. Also, they are enabling experimentation to discover needs, expose variability, and improve performance. At the end, the use of big data will underpin new waves of productivity growth and consumer surplus. (Google owns the Matrix right now) 
Vlad S
tl;dr: Study machine learning!
The questions at the end are quite funny. SVM, aliens and talking to lions. ;) Btw, Thad Starner, the Glass guy who works for Google, had a PhD student who developed a method to extract patterns from signals and was applying it to recordings of dolphins.
Yep, interesting talk indeed. I think they are using DeepLearning not SVD. The translator discussed here, like Google search, is pattern matching with some math. It really cannot understand something that hasn't been understood somewhere in the universe. But it is interesting to note that we cannot even understand intelligent animals who have been living with us for centuries ... in that sense the Google technologies are very primitive ... and that is the best we have ;o(
Add a comment...