e-infrastructures for data management and deep learning for NLP
This week, I spent two days at the Workshop on e-Infrastructures for Massively Parallel Sequencing in Uppsala. It was a nice event with sessions on workflow systems, virtual machines, DNA sequencing data management experiences in various countries etc. Rather than summarizing the takeaways, I’d like to point to a Storify made by Surya Saha which includes a lot of tweets from yours truly. I sometimes live-tweet talks from events but I don’t think I’ve ever gotten so many retweets and new followers before. I guess I underestimated how interested people are on workflow management and pipelining systems.
Tonight, I went to a meetup on deep learning and natural language processing hosted by the Stockholm NLP Meetup group. It was a fun session led by Roelof Pieters which included hands-on programming with Theano and some tantalizing glimpses on what you can do with word embeddings and distributed semantics. A recording can be found here and the slides are here. There’s also a GitHub repo with some iPython notebooks containing useful example code in Theano.