Machine Learning Discussion Group

From Cornell CS Wiki
Jump to: navigation, search

The Machine Learning Discussion Group (MLDG) is an informal group for discussing the latest work in the field of Machine Learning. We usually discuss a paper from a recent conference (NIPS, ICML, etc) each meeting.

The organizers are Sarah Tan and Ashudeep Singh. This semester (Fall 2017) we meet on Wednesdays (5.30 - 6.30 pm) in Gates 416 (Ithaca) and Bloomberg 398 (Cornell Tech, New York City).

Schedule of papers

Feel free to volunteer for any available week! Send an email to ht395 [at] cornell.edu and as3354 [at] cornell.edu.

Date Discussion Leader Topic, Paper links Area Other activities and Deadlines
30 Aug, 2017 Organizational Meeting
6 Sept, 2017 Wesley YellowFin and the Art of Momentum Tuning and On the importance of initialization and momentum in deep learning Deep Learning Sep 7 - 11 EMNLP conference, Sep 8 - AAAI deadline
13 Sept, 2017 James Deep Learning for Identifying Metastatic Breast Cancer Deep Learning
20 Sept, 2017 Dieqiao Neural Turing Machines Deep Learning Sep 21 - Colloquium Speaker: Chris Re, Stanford
27 Sept, 2017 Luke and Eric Recurrent Convolutional Network for Video-based Person Re-Identification Deep Learning
4 October, 2017 Dipendra Value Iteration Networks Reinforcement Learning
11 October, 2017 Alane From Language to Programs: Bridging Reinforcement Learning and Maximum Marginal Likelihood Reinforcement Learning/NLP Oct 13 - AISTATS deadline, Oct 13 - AI Seminar Speaker: Andrew McCallum, UMass Amherst
18 October, 2017 Andrew Reinforcement Learning for Bandit Neural Machine Translation with Simulated Human Feedback Reinforcement Learning/NLP Oct 19 - Colloquium Speaker: Tom Kalil, Office of Science and Tech Policy
25 October, 2017 Arzoo Automatic Rule Extraction from Long Short Term Memory Networks NLP Oct 26 - WWW deadline, Oct 27- ICLR deadline
1 Nov, 2017 Ashudeep The Mythos of Model Interpretability and Understanding Black-box Predictions via Influence Functions Fairness and Transparency in ML
8 Nov, 2017 Sarah Interpretation of Neural Networks is Fragile and Feature Visualization: How neural networks build up their understanding of images Fairness and Transparency in ML
15 Nov, 2017 Angela Learning Cost-Effective Treatment Regimes using Markov Decision Processes Fairness and Transparency in ML Nov 15 - CVPR deadline
22 Nov, 2017 Thanksgiving Nov 30 - Colloquium Speaker: Scott Aaronson, UT Austin
29 Nov, 2017 Ayush The Landscape of Empirical Risk for Non-convex Losses Theory Dec 4-9 NIPS conference

FAQ

Who attends the MLDG?

The group is mainly attended by graduate students and undergraduate students actively interested in the area of machine learning. Professors are also welcome. Beginning of every semester, we meet to discuss areas of interests to each of the people attending and decide on a few themes for the semester. Suggestions for topics or papers to discuss may be sent to the mailing list and are always welcome.

How do I join the mailing list?

You can join the list by sending an email with subject "join" (no quotes) to cornell-mldg-l-request@cornell.edu.

How do I get involved in MLDG?

Stop by and check us out any week! You can find our meeting time and location at the top of this page. Also, if you have anything interesting to share or announce, you can use the mldg mailing list, cornell-mldg-l@cornell.edu.

Archive of previous MLDG pages