Monday, October 23, 2017
6:00 PM to 8:00 PM
ACL Office, 14th Floor
980 Howe St., Vancouver, BC
Agenda:
• Give a general overview of the meetup and the course
• Briefly step through a skip-gram model for word2vec and the mathematics behind the error derivatives (Material in Lecture 2)
• Step through advanced word vector representations - Glove vectors (Material in Lecture 3)
• Tour of questions in Homework 1
How to Prepare?
Participants are expected to watch video lecture 1, video lecture 2 and video lecture 3 so as to not get lost during the discussions. We'll be going through some of this material at a brisk pace and concentrate on discussions as we step through some of the material.
Oct 30 (next meetup): Solutions to Homework 1 + Material that will lead up to Homework 2.
What we're open to (sponsors welcome):
• Ideas for prizes for the final winning project
• Ideas for obtaining GPU access
Discussions:
Join us on our Vancouver technology (VanTech) Slack channel #rogue-ml . To sign up for the VanTech Slack channel click here or go to vantech.herokuapp.com.
General Description:
The goal of this meetup is to get together to work through the material of CS224n from Stanford : Deep Learning for Natural Language Processing. The video lectures and materials for this course are available online: http://web.stanford.edu/class/cs224n/. We will come together on a weekly/bi-weekly (TBD) basis, review some of the lecture material and dive right into solutions of the homework problems and projects. This is a project heavy meetup where you would be expected to come and talk about your solution to a homework project and what worked and what didn't. There will be a final project that each student is expected to work on, and present at the end of the course. CS224n offers a default project as Assignment 4. A list of other projects done by Stanford students who took the course is available on the course website. A prize would be awarded for the best project from amongst the participants of this meetup. The best project will be judged by Natural Language Processing practitioners from the industry and the academia.