Digging Deep Learning at The Crypt

Cem Baris 13.03.2015

Wednesday 11th March, the Morgan McKinley Data and Technology team attended the 17th London Machine Learning Meetup at The Crypt in Clerkenwell.

A popular event on the London Technology scene, the guest list was full within hours of being open including a rather lengthy waiting list that was 100+ strong. With plenty of beers and pizza to go around, the evening was set up for a relaxed conference with the opportunity to network among Machine Learning enthusiasts and learn more about Natural Language Processing.   

I thought the meetup was a great way of getting lots of people who work in data science and machine learning together to discuss ideas.Speaking on the night was Felix Hill (PhD Student in NLP at University of Cambridge) who shared with us his insight on applying Deep Learning to natural language, and how this has resulted in some impressive breakthroughs to certain language processing applications in the last year or two.  The examples he gave were of 'language modelling, machine translation and image caption generation'.
 
Felix presented some of his work relating to understanding how and why neural language models are so effective. The models learn the meaning of words and how to combine these meanings to do the task they are built to do. He explained that the meaning of the words correspond to human intuitions about word meanings. Felix also gave a demo of a new system that has been built that functions as a 'reverse dictionary' (given a definition-like query, can you tell me the word I'm thinking of) and a general-knowledge crossword solver.

Felix Hill shared his thoughts on the night,"I thought the meetup was a great way of getting lots of people who work in data science and machine learning together to discuss ideas. From my perspective I learned about some of the challenges that companies are facing in industrial and commercial scenarios, which can ultimately give me ideas for what to look at in a research environment."London Machine Learning allows me to get to know so many people who are interested in the same niche area as myself and hearing about their stories and their challenges.

Francesco Emanuele Bucci, Co-Organiser London Machine Learning said "It's a great pleasure to organise and participate in such an event. First of all it's a pleasure to get to know people like Felix who are at the forefront of their research which I find thoroughly interesting. His talk was great: I'm into NLP as well and this offers me the opportunity to hear directly from someone who is researching this area and who deals with the innovation in the field everyday is amazing." 

Francesco continued “I enjoyed his discussion on research, researchers and the relationship between academia and the industry.  London Machine Learning allows me to get to know so many people who are interested in the same niche area as myself and hearing about their stories and their challenges."’ 

Cem Baris, Associate Director Technology, Risk and Analytics at Morgan McKinley said "London Machine Learning have a special thing going on, the following is very strong and the speed in which the guest list fills up is evidence of that.  It's great to be at a meet up or talk where the audience are so engaged and participation is high. With an exciting roster of speakers and events in planning we're really looking forward to continuing our relationship with LML."

Cem Baris's picture
Director
cbaris@morganmckinley.co.uk