FLAIR

Program

Future Leaders of AI Retreat

NYU Shanghai

 

Click here to download FLAIR Handbook

Program

Wednesday, December 20

08:30am-09:00am     Registration

09:00am-09:10am     Opening

Session I   "Theory and Core Algorithms"

09:10am-09:40am     On the Power of Randomly Initialized Gradient Descent for Learning Convolutional Neural Networks (by Shaolei Du)

09:40am-10:10am     Opening the black of deep neural networks via information (by Ravid Schwartz-Ziv)

10:10am-10:30am     Tea break

10:30am-11:00am     Nonconvex Sparse Blind Deconvolution: Global Geometry and Efficient Methods (by Yuqian Zhang)

11:00am-11:30am     Adversarial regularized autoencoders and error encoding network (by Junbo Zhao)

11:30am-12:00pm     Enhancing Iterative Algorithms using Neural Networks (by David Wipf) (Keynote)

12:00pm-12:15pm     Q&A

12:15pm-01:15pm     Lunch

Session II   "Application: Game and Generalizable Intelligence"

01:15pm-01:45pm     Sample-Efficient Deep RL for Robotics: Generalizing On-policy, Off-policy, and Model-based Approaches (by Shixiang Gu)

01:45pm-02:15pm     Towards a wholistic approach to dexterous and cluttered manipulation (by Jiaji Zhou)

02:15pm-02:45pm     Towards Generalizable Robot Learning with Perception (by Yuke Zhu)

02:45pm-03:15pm     Effective Master-Slave Communication for Multi-Agent Deep Reinforcement Learning (by Xiangyu Kong)

03:15pm-03:45pm     AI in Games: Achievements and Challenges (by Yuandong Tian) (Keynote)

03:45pm-04:00pm     Q&A

04:00pm-04:15pm     Tea break

Session III   "Application: Natural Language Processing and Computer Vision"

04:15pm-04:35pm     Present and Future of Natural Language Processing - With Dialogue as Example (by Hang Li) (Keynote)

04:35pm-05:05pm     Think out of black-box (by Zhengdong Lu)

05:05pm-05:35pm     Learning to Reason with End-to-End Neural Module Networks (by Ronghang Hu)

05:35pm-06:05pm     On unifying deep generative models (by Zhiting Hu)

06:05pm-06:35pm     Deep Multitask Learning for Semantic Dependency Parsing (by Hao Peng)

06:35pm-06:50pm     Q&A

Thursday, December 21

08:30am-09:00am     Registration

Session IV   "System"

09:00am-09:30am     TVM: An End to End IR Stack for Deep Learning Systems (by Tianqi Chen)

09:30am-10:00am     Tofu: Distributing Tensor Computation Automatically for Deep Learning Systems (by Minjie Wang)

10:00am-10:15am     Q&A

10:15am-10:30am     Tea break

Session V   "Creativity and Music with AI"

10:30am-11:00am     Empowering music intelligent systems with expression and interactivity (by Gus Xia)

11:00am-11:30pm     Creativity in Musical Intelligence (by Zhengshan Shi)

11:30am-12:00pm     Crowd Sourcing Clothes Design Directed by Adversarial Neural Networks (by Hiroyuki Osone & Daitetsu Sato)

12:00pm-12:15pm     Q&A