DL2.0 Unit, I2R-TokyoTech Co-workshop on Deep Learning

Day 1 7th March [PM] Resonance@L13N
14:00 - 14:30 30 mins Topic: Towards Robust Compressed Convolutional Neural Networks

Presenter: Assoc. Prof Tsuyoshi Murata (Tokyo Tech)
14:30 - 15:00 30 mins Topic: Deep Learning Frameworks Portability Study: Post-K Perspective

Presenter: Dr Aleksandr Drozd (Tokyo Tech)
15:00 - 15:30 30 mins Topic: Optimizing The Bit Allocation for Compression Of Weights And Activations Of Deep Neural Networks

Presenter: Wang Zheng (DL2.0)
15:30 - 16:00 30 mins Topic: Dataflow-based Joint Quantization of Weights and Activations for Deep Neural Networks

Presenter: Geng Xue (DL2.0)
16:00 - 17:30 1.5 hours Deep dive on Model Compression [Neural Networks]


Day 2 8th March [AM] Franklin@11S
9:00 - 9:10 10 mins Opening speech - Foo Chuan Sheng
9:10 - 9:40 30 mins Introductory Talk: JST CREST Project Fast and cost-effective deep learning algorithm platform for video processing in social infrastructure.

Speaker - Prof Koichi Shinoda (Tokyo Tech)
9:40 - 10:00 20 mins Topic: Free Discussion Session
10:00 - 10:15 15 mins Tea Break (Snack)
10:15 - 10:45 30 mins Topic: Second Order Optimization for Distributed Data-parallel Deep Learning on 4000 GPUs

Presenter: Assoc Prof Rio Yokota (Tokyo Tech)
10:45 - 11:05 20 mins Topic: Accelerating Deep Learning with Distributed Computation and Second-Order Optimization

PhD Sharing: Mr. Yohei Tsuji (Tokyo Tech)
11:05 - 11:25 20 mins Topic: Accelerating Deep Learning with Distributed Computation and Second-Order Optimization: Future work

PhD Sharing: Mr. Kazuki Osawa (Tokyo Tech)
11:25 - 11:55 30 mins Topic: 100x speed up for Homomorphic Encryption Meets Deep Learnings

Presenter: Khin Mi Mi Aung (DL2.0)
12:00 - 13:00 1 hours Deep dive on Acceleration Homomorphic Encryption
13:00 - 14:30 1.5 hour Lunch


Day 2 8th March [PM] Resonance@L13N
14:30 - 15:00 30 mins Topic: Towards Large-Scale Few-Shot Learning

Presenter: Asst Prof Nakamasa Inoue (Tokyo Tech)
15:00 - 15:20 20 mins Topic: Approaches for Data Augmentation for Videos Using GAN

PhD Sharing: Mr. Pablo Cervantes (Tokyo Tech)
15:20 - 15:40 20 mins Topic: Second Order Methods and Other Techniques for Fast Training of GAN

PhD Sharing: Mr Yasin Yazici (DL2.0)
15:40 - 16:00 20 mins Tea Break (Snack)
16:00 - 16:30 30 mins Topic: Efficient FPGA Implementation for Deep Learning Based Applications

Presenter: Assoc Prof Hiroki Nakahara (Tokyo Tech)
16:30 - 17:00 30 mins Topic: Generalization error analysis of deep learning and its application to model compression

Presenter: Assoc Prof Taiji Suzuki (Tokyo Tech)
17:00 - 17:30 30 mins Topic: Video Classification on YouTube-8M Dataset from Raw Video Data

Presenter: Wang Anran (DL2.0)
17:30 - 18:00 30 mins Topic: Training Very Deep Object Detectors with Large Minibatch SGD

Presenter: Cai Lile (DL2.0)
18:00 - 18:30 30 mins Topic: Towards Convergence of Extreme Computing and Big Data Centers

Presenter: Prof. Satoshi Matsuoka (Tokyo Tech)
18:30 - 19:00 30 mins Deep Dive on Deep Learning (GANS, Learning with Less Data, Distributed Deep Learning, Error Analysis, Deep Learning for Visual Classifications & Detection)
Day 3 9th March [AM] Resonance@L13N
10:00 - 11:30 1.5 hours Topic: Discussion on Collaboration Direction & Partnership

Attendees:
(Tokyo Tech)
Prof Koichi Shinoda, Assoc Prof Rio Yokota and other professors and students
(I2R)
Dr Vijay Chandrasekhar, Dr Foo Chuan Sheng and other researchers
11:30 - 12:00 30 mins Closing on Visit