TTIC Student Workshop

February 8, 2019


Congratulations to the award winners!

Best Talk Award: Pedro Savarese and Omar Montasser
Best Poster Award: Shubham Toshniwal

The 3rd TTIC Student workshop will take place on February 8, 2019 on the 5th floor at TTIC. Talks and panel will be in room 526. Posters and refreshments will be in the 5th floor common area. There are a total of 9 talks and 4 poster presentations. Talks will be 25 minutes long: 20 min talk, 4 min questions, 1 min turnover. There will be awards for the best poster and best talk.

Organizing Committee: Blake Woodworth, Shubham Toshniwal, Karen Livescu, Madhur Tulsiani, Erica Cocom
Talk Award Committee: Sam Wiseman, Steve Hanneke, Mesrob Ohannessian
Poster Award Committee: Michael Yu, Arturs Backurs


09:00-09:20 Breakfast refreshments
09:20-09:30 Opening Remarks
09:30-10:30 Talk Session 1 (Chair: Matthew Walter)
09:30-09:55 Kevin Stangl Fairness Constraints with Concrete Bias Models
09:55-10:20 Pedro Savarese How do infinite-width, bounded-norm ReLU networks look in function space?
10:20-10:45 Sumeet Katariya Adaptive Sampling for Ranking and Clustering
10:45-11:30 Poster Session
Hai Wang Evidence Sentence Extraction for Machine Reading Comprehension
Ruotian Luo Context-aware zero-shot recognition
Zewei Chu Zero shot text classification with a million Wikipedia categories
Shubham Toshniwal Inter-Language Relationships in Cross-Lingual Embedding Spaces
11:30-12:45 Talk Session 2 (Chair: Kevin Gimpel)
11:30-11:55 Davis Yoshida Analyzing Natural Language Understanding Models
11:55-12:20 Nick Kolkin Style Transfer by Relaxed Optimal Transport and Self-Similarity
12:20-12:45 Shane Settle Acoustically Grounded Word Embeddings for Improved Acoustics-to-Word Speech Recognition
12:45-13:45 Lunch + Poster session (same posters)
13:45-15:00 Talk Session 3 (Chair: Madhur Tulsiani)
13:45-14:10 Falcon Z. Dai Maximum Expected Hitting Cost of a Markov Decision Process and its Conditioning by Potential-based Reward Shaping
14:10-14:35 Mrinalkanti Ghosh Approximability of Matrix p to q Norms
14:35-15:00 Omar Montasser Adversarially Robust Learnability
15:00-15:20 Break + refreshments
15:20-15:30 Awards
15:30-whenever Panel discussion: Allyson Ettinger, David McAllester, Greg Shakhnarovich, Rebecca Willett