Deep Learning for Vision
NUS SoC, 2017/2018, Semester I
CS 6101 - Exploration of Computer Science Research
September 7, 2017 - Updated venue and assignments.
July 5, 2017 - Initial port of last year's home page.
This is a section of the CS 6101 Exploration of Computer Science Research at NUS. CS 6101 is a 4 modular credit pass/fail module for new incoming graduate programme students to obtain background in an area with an instructor's support. It is designed as a "lab rotation" to familiarize students with the methods and ways of research in a particular research area. The lab rotation is hosted by the Web IR / NLP Group (WING) at NUS, led by Min-Yen Kan.
Our section will be conducted as a group seminar, with class participants nominating themselves and presenting the materials and leading the discussion. It is not a lecture-oriented course and not as in-depth as Fei-Fei Li et al.'s original course at Stanford, and hence is not a replacement, but rather a class to spur local interest in Deep Learning for Vision. Unlike the original course, we do not require you to do projects, assignments or homework, although those that attempt this will be encouraged to present in the project showcase at the end of the course to the general public.
This "course" is offered twice, for Session I (Weeks 3-7)and Session II (Weeks 8-13), although it is clear that the course is logically a single course that builds on the first half. Nevertheless, the material should be introductory and should be understandable given some prior study.
A discussion group will be on Slack 
. Students and guests, please login when you are free.
If you have a @comp.nus.edu.sg, @u.nus.edu,
@nus.edu.sg, @a-star.edu.sg, @dsi.a-star.edu.sg or
@i2r.a-star.edu.sg email address you can create your Slack account
for the group discussion without needing an invite.
Meeting Venue and Time
18:00-20:00, Tuesdays for both Sessions I (Weeks 3-7) and Session II (Weeks 8-13).Updated Venue is Active Learning Lab (COM1 #B1-03).
For directions to NUS School of Computing (SoC) and COM1: please read the directions here, to park in CP15 or take the shuttle bus to SoC.
Please eat before the course or during (we don't mind -- like a brown-bag seminar series).
People
Welcome. If you are an external visitor and would like to join us, please email Kan Min-Yen to be added to the class role. Guests from industry, schools and other far-reaching places in SG welcome, pending space and time logistic limitations. The more, the merrier.
External guests will be listed here in due course once the course has started. Please refer to our Slack after you have been invited for the most up-to-date information.
NUS (Postgraduate): Session I (Weeks 3-7): Siddharth Aravindan, Devamanyu Hazarika, Nguyen Van Hoang, Xie Yaqi, Ziwei Xu, Yew Zi Jian
NUS (Postgraduate): Session II (Weeks 8-13): Cheng Chen, Xinyu Chen,Remmy A. M. Zen, Ridi Hossain, Chengxi Xue, Meng-Jiun Chiou
Other NUS: Wesley Chui Lui Goi, Jethro Kuan, Nandha Kumar, Kyaw Zaw Lin, Joel Lee, Panpan Qi, Krishnendu Sanyal, Shivshankar Umashankar, Ming Rui Wang
WING: Muthu Kumar Chandrasekaran, Divish Dayal, Min-Yen Kan, Animesh Prasad
Guests: Ming Liang Ang, Nipun Batra, Wesley Chui Lui Goi, Canh Tran Duy, Christabella Irwanto, Siow Meng Low, Jishnu Mohan, Nyan Tun Zaw, Karthik Raja Periasamy, Sherly Sherly, Kok Keong Teo, Shivshankar Umashankar, Sriram Vaikundam, Daniel Wong
Syllabus
We're happy to report that the Stanford course CS231n: Convolutional Neural Networks for Visual Recognition, taught by Fei-Fei Li, Justin Johnson and Serena Yeung has agreed to let us run an iteration of their course at NUS. Please refer to their page for the detailed syllabus.
Please note that the original course contains additional lectures that are prerequisites to the materials that we are using. Our course starts on Lecture 4, but participants may want to start with the pre-flight Lectures 1-3 that we may cover optionally before the course formally starts. Please check in Slack for more details.
| Session | Date | Description | Readings | ||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pre-flight
Video Conference Room |
Weeks 1-2
17 Aug | Course Introduction:
Computer vision overview Historical context Course logistics Image Classification: The data-driven approach K-nearest neighbor Linear classification I Loss Functions and Optimization: Linear classification II Higher-level representations, image features Optimization, stochastic gradient descent Presenters: Min |
| ||||||||||||||||||||||||||||||
| Session I
Weeks 3-7
| Week 3 (29 Aug)
| Introduction to Neural Networks:
Backpropagation Multi-layer Perceptrons The neural viewpoint Presenters: Min |
| ||||||||||||||||||||||||||||||
| Week 4 (5 Sep)
| Convolutional Neural Networks:History Convolution and pooling ConvNets outside vision Presenters/Questioners: Devamanyu Hazarika, Siddharth Aravindan, Ming Rui Wang / Xie Yaqi, Kenneth Tan, Kok Keong Teo |
|
Student Projects
Student projects are required for all external guests to the course, to students in the course with conflicting lecture timings. All students in the course are highly encouraged to take on a project to gain practical experience in tandem with the class lecture. All student projects can be done in any sized team, and will be featured in the 11th SoC Term Project Showcase (STePS).
The below listing is tentative, please refer to the Slack group or to the STePS website for authoritative information.
- Ang Mingliang: Deepore: Deep learning for base calling MinION reads
- Christabella Irwanto / Jethro: A Neural Network Approach to Product Similarity and Pricing Suggestions - 1st Place winner at 11th STePS!
- Kok Keong Teo: How Transferrable are Features in Deep Neural Networks?
- Joel Lee / Le Trung Hieu: Yet Another attempt on the CDiscount Kaggle Competition
- Swee Kiat Lim: Deep Reinforcement Learning with StarCraft II - 2nd Place winner at 11th STePS!
- Yunshan Ma / Ziwei Xu: E-Commerce Product Classification on both Image and Text Information
- Tram Anh Nguyen: Right Whale Challenge
- Kenneth Tan: Drug Guru
- Kok Keong Teo: How Transferrable are Features in Deep Learning Model?
- Siow Meng: ImageNet Object Detection
- Shivshankar Umashankar / Wesley Goi: Multi-class classification of e-commerce products
- Suchun Wong: Bag-Of-Words Representation of Instances from Faster-RCNN
- Deric Yeak: FaceLog Mobile Solution - 3rd Place winner at 11th STePS!
Other Links
- Deep Learning for NLP using Python (Git Repo) - https://github.com/rouseguy/europython2016_dl-nlp
- Deep Learning, an MIT Press book, by Ian Goodfellow, Yoshua Bengio and Aaron Courville - http://www.deeplearningbook.org/
- Neural Networks and Deep Learning - free e-book by Michael A. Nielsen - http://neuralnetworksanddeeplearning.com/