#6: recsys with matrix factorization and information extraction from documents

Our schedule for this time:

1. Recommender systems with matrix factorization (Abhishek Thakur (https://twitter.com/abhi1thakur))

In this talk, we will discuss the basics of recommendation systems and how a simple recommendation system can be constructed using a nearest neighbor method. This will be extended to recommendation systems with matrix factorization and analysis of different methods used for the Netflix Challenge.

2. Information Extraction from scanned documents - from rule based to Machine Learning (Florian Kuhlmann, leverton.de)

Abstract hopefully coming soon.

3. Meet robots halfway (Michael Bucko, 20min)

Gödel once wondered how we can understand each other. Yes, we all have different definitions of love. I will follow this idea and present defin3.com- the tool, which I build to do something with this issue and meet robots halfway. Defin3.com is meant to be a moderated user-based (Wiki-style) dictionary of (connected) meanings. The rule is: to build a higher abstraction word, you need to have its building blocks defined first. In my talk, I am going to focus on the algorithms behind the quality of claims that can be built from the words defined in defin3.com (http://defin3.com/).

Have an interesting topic you want to talk about?

We're always looking for interesting presentations for our meetups. If you have a topic you're excited about and you want to talk about, maybe between 15 and 50 minutes, then send a short abstract to daniel.nouri gmail.com. Thanks!