Ce Soir: Paris Machine Learning Meetup #14 Season 3: Dare Mighty Things

What a season and what an end of a season we have for you. With no advertizing. no public funding, just a bottom-up wave, we have gathered 3780 people interested in Machine Learning in Paris. That makes it one of the largest and most attended meetup on Earth. Wow!

Tonight's program will feature issues such as how Machine Learning is helping us understand Mars better with Raymond out of JPL. Lenka will provide us some guiding principle on what statistical physics can bring to Machine Learning. Antoine will guide us in the newclass of memory networks. Mark will teach us how to teach bots to be good moral citizens. Philippe will show a conversation engine for bots. Stephane will talk to us about OSs and ML and finally Robert will be seeking some input from the community as he tries to benchmark several AI services. The meetup is hosted and sponsored by Lagardere. Thank you to them and particularly Olivier for making it happen.

The following video is a short explanation of how Mars exploration is difficult, in part due to the time delay between Earth operations at JPL and Mars. Machine Learning is one of these technique that enable autonomous operations on that planet. The following video details the autonoous systems needed when Curiosity entered Mars atmsophere and landed on Mars in 2012. Enjoy your seven minutes of Terror of  EDL.

 The following is the video of the streaming for tonight's meetup that should start at 7:00PM Paris time.

Here is the program (with the slides -all the slides should be there before the meetup starts-)

• Olivier Sorba, Lagardere, presentation
Rony Rozas, LeGuide Lagardere, presentation2

• Lenka Zdeborova, ( @zdeborova ) Institut de Physique Theorique de Saclay/CNRS, What can physics bring to machine learning?

• Mark Reidl, (@mark_riedl ),Georgia Tech, Using Stories to Teach Human Values to Artificial Agents

This talk describes ongoing research to develop an AI system that learns human values and social conventions in order to avoid amoral behaviors.

• Antoine Bordes Facebook AI Research lab (FAIR), Memory Networks pour questions-réponses et systèmes de dialogues

Cette présentation va discuter des Memory Networks, une nouvelle classe de réseaux de neurones pouvant manipuler une mémoire symbolique. Nous montrerons comment ils peuvent être utilisés pour la conception de systèmes de questions-réponses et de dialogues.

• Raymond Francis (@CosmicRaymond) NASA/JPL, AEGIS Autonomous Targeting Software Deployed for the Curiosity Rover's ChemCam Instrument (ppt version)

The Mars Science Laboratory mission has deployed new software that allows the rover to autonomously recognize & select rock targets for its ChemCam laser spectrometer, acquire data, and transmit the results to Earth, all without human intervention.

• Philippe Duhamel,(@teamkadeal ) Clustaar, Moteur de conversation pour les bots/robots 
(ppt)

• Stephane Bura, (@s_bura ) Weave.ai, Machine Learning and OS (ppt ) (see also this blog entry on AI and the future of the Operating System)



+ Robert Salita, Comparing, Benchmarking AI Services

Just hacked a tool to compare and benchmark AI services. All major providers are supported. I have some preliminary hard data but it raises more questions than it answers. I need community feedback.

 

Join the CompressiveSensing subreddit or the Google+ Community or the Facebook page and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.