The NIPS Workshop on Multi-Class and Multi-Label Learning with Millions of Categories
Monday, 9th December 2013, Lake Tahoe, Nevada, United States
Extreme Delights

The workshop venue is Harrah's Fallen+Marla. Please note that the morning session will start at 7:00 AM (not 7:30 AM) and the afternoon session at 3:00 PM (not 3:30 PM).

07:00 - 07:05 Introduction Manik Varma and John Langford

Session 1: Extreme Multi-Label Learning with Applications to Ranking and Recommendation

07:05 - 07:35 Manik Varma (MSR) Reformulating Ranking and Recommendation as Multi-Label Learning with Millions of Labels
07:35 - 08:05 Jason Weston (Google) Label Partitioning for Sublinear Ranking
08:05 - 08:35 Hsiang-Fu Yu (UT Austin) Large-Scale Multi-Label Learning with Missing Labels
08:35 - 09:05 Discussion Chair - John Platt
 
09:05 - 09:30 Coffee break + posters

Session 2: Extreme Multi-Class Classification - Taxonomies, Trees and Embeddings

09:30 - 10:00 Paul Bennett (MSR) Extreme Classification in Large Taxonomies
10:00 - 10:30 Anna Choromanska (Columbia) Extreme Multi-Class Classification
10:30 - 11:00 Bin Zhao (CMU) Structured Sparse Output Coding for Scalable Multi-Class Classification
11:00 - 11:30 Discussion Chair - John Langford
 
11:30 - 11:45 Spotlights
 
11:45 - 15:00 Lunch

Session 3: Extreme Deep and Visual Learning

15:00 - 15:30 Samy Bengio (Google) DeViSE: A Deep Visual-semantic Embedding Model
15:30 - 16:00 Rob Fergus (NYU) Visualizing and Understanding Convolutional Neural Networks
16:00 - 16:30 Jay Yagnik (Google) Fast, Accurate Detection of 100,000 Object Classes on a Single Machine
16:30 - 17:00 Yoshua Bengio (Montreal) From Deep Learning and Multi-Task Transfer to Structured Probabilistic Outputs with GSNs
17:00 - 17:30 Discussion Chair - Yann Lecun
 
17:30 - 19:00 Coffee break + posters

Spotlights

Heejin Choi & Nathan Srebro Hierarchical Classification with Structured SVMs
Maya Gupta, Samy Bengio & Jason Weston Training Highly Multi-Class Linear Classifiers
Harish Guruprasad, Shivani Agarwal & Ambuj Tewari Consistent Algorithms for Extreme Multi-Class Problems with a Label Hierarchy
Akshay Balsubramani & Omid Madani An Empirical Comparison of Sparse vs. Embedding Techniques on Many-Class Text Classification
Partha Pratim Talukdar & William Cohen Scaling Graph-Based Semi Supervised Learning to Large Number of Labels Using Count-Min Sketch
Nicholas Rhinehart, Jiaji Zhou, Martial Hebert & J. Andrew Bagnell Fine-Grained Detection via Efficient Extreme Classification