CS 395T - Deep learning seminar - Fall 2018

meets MW 2:00 - 3:30pm in GDC 4.304

instructor Philipp Krähenbühl
email philkr (at) utexas.edu
office hours by appointment (send email)

TA Chia-Wen Cheng
email cwcheng (at) cs.utexas.edu
office Desk 3 in GDC 1.302
TA hours MW 4:00 - 5:00pm

Please use canvas for assignment questions.

Prerequisites

Class overview

This class covers advanced topics in deep learning, ranging from optimization to computer vision, computer graphics and unsupervised feature learning, and touches on deep language models, as well as deep learning for games. This is meant to be a very interactive class for upper level students (MS or PhD). For every class we read two recent research papers (most no older than two years), which we will discuss in class.

Goals of the class

After this class you should be able to

Grading

To map percentages to letter grades we use the following python script

def grade(p):
  from math import floor
  if p < 60: return 'F'
  v = (100-p) / (10 + 1e-5)
  return chr(ord('A')+int(v)) + ['+','','','-'][int((v-floor(v))*4)]

Schedule

Date Topic Papers Presenters Notes and due dates
Aug 29 Administrative and intro (Linear models)   Philipp  
Sep 3 no class (labor day)      
Sep 5 Gradient based optimization Large-scale machine learning with stochastic gradient descent, Bottou 2010 Philipp paper selection Th Sep 5, 1am canvas
Sep 10 Deep networks and backpropagation Efficient BackProp, LeCun etal. 1998

Deep learning, LeCun, Bengio and Hinton 2015
Xing Han
Su Wang

Mei Wang
Xirou Wang
 
Sep 12 Dropout and batch normalization Dropout: A Simple Way to Prevent Neural Networks from Overfitting, Srivastava etal. 2014

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift, Ioffe etal. 2015
Yixuan Ni
Xinrui Hua


Elisa Ferracane
Jiacheng Zhuo
project 1 out
Sep 17 Advanced optimization and initialization Adam: A Method for Stochastic Optimization, Kingma and Ba 2015

On the Convergence of Adam and Beyond, Reddi etal 2018
Nidhi Kadkol
Akshay Kamath

Jiacheng Zhuo
Elisa Ferracane
 
Sep 19 Convolutional Networks for image classification Gradient-based learning applied to document recognition, LeCun etal. 1998 (sec 4+ optional)

ImageNet Classification with Deep Convolutional Neural Networks, Krizhevsky etal. 2012
Subham Ghosh
Yixuan Ni

John Fang
Madhumitha Sakthi
 
Sep 24 Convolutional Networks for object detection and pixel-wise prediction Rich feature hierarchies for accurate object detection and semantic segmentation, Girshick etal. 2014

Fully Convolutional Networks for Semantic Segmentation, Long etal 2015
Brahma Pavse
Hsin-Ping Huang

Wonjoon Goo
Xingyi Zhou
 
Sep 26 Advanced deep network architectures Deep Residual Learning for Image Recognition, He etal. 2016

Densely Connected Convolutional Networks
Venkata Ravi Ailavarapu
Eddy Hudson

Xingyi Zhou
Ankur Garg
Project 1 QA
Oct 1 Visualizing deep networks Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps, Simonyan etal. 2014

Inverting visual representations with convolutional networks, Dosovitskiy 2016
Ankur Garg
Uday Kusupati

Jerry Tang
Nidhi Kadkol
 
Oct 3 Image generation Generative Adversarial Nets, Goodfellow etal. 2014

Wasserstein GAN, Arjovsky etal 2017
Akshay Kamath
Prateek Shrishail Kolhar

Prateek Shrishail Kolhar
Kurtis David
 
Oct 8 Project 1 presentations - Yearbook/Geolocation presentation schedule 7 min per team   Project 1 due 1am
Oct 10 Project 1 presentations - Yearbook/Geolocation presentation schedule 7 min per team   Project 2 out
Oct 15 Image generation II Auto-Encoding Variational Bayes, Klingma etal. 2014

Density estimation using Real NVP, Dinh etal 2016
Kurtis David
Mei Wang

Su Wang
Haresh Karnan
 
Oct 17 Image translation High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs, Wang etal 2018

Semi-parametric Image Synthesis, Qi etal. 2018
Madhumitha Sakthi
Wei-Jen Ko

Hsin-Ping Huang
-
 
Oct 22 Unpaired image translation Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks, Zhu etal. 2017 Wei-Jen Ko
Tanya Goyal
Project 2 flash presentations (2min)
Oct 24 Compression Full Resolution Image Compression with Recurrent Neural Networks, Toderici etal 2016

Real-Time Adaptive Image Compression, Rippel and Bourdev 2017
Abhinav Singh
Neeharika Immaneni

Neeharika Immaneni
Abhinav Singh
 
Oct 29 Adversarial attacks Explaining and Harnessing Adversarial Examples, Goodfellow etal. 2014

Towards Evaluating the Robustness of Neural Networks, Carlini and Wagner 2016
Quang Duong
Subham Ghosh

Xianda Zhou
Quang Duong
 
Oct 31 Recurrent models Generating Sequences With Recurrent Neural Networks, Graves 2013

Pixel Recurrent Neural Networks, Oord etal 2016
Ryan Wolter
Ryan Rock

Xinrui Hua
Xianda Zhou
 
Nov 5 Language models Sequence to sequence learning with neural networks, Sutskever etal. 2014

Neural machine translation in linear time, Kalchbrenner etal. 2016
Aditya Gupta
Ryan Wolter

Daniel Crockett
Jerry Tang
 
Nov 7 Video Two-Stream Convolutional Networks for Action Recognition in Videos, Simonyan and Zisserman 2014

Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset, Carreira and Zisserman 2017
Uday Kusupati
Venkata Ravi Ailavarapu

Ryan Rock
Wonjoon Goo
 
Nov 12 Atari games Playing Atari with Deep Reinforcement Learning, Mhin etal. 2013

Human-level control through deep reinforcement learning, Mhin etal. 2015
Pratyush Kar
Brahma Pavse

Eddy Hudson
Pratyush Kar
 
Nov 14 Video games Playing for Benchmarks, Richter etal. 2017 Xiruo Wang
Abduallah Mohamed
Project 2 - QA
Nov 19 Alpha GO Mastering the game of Go with deep neural networks and tree search, Silver etal. 2016

Mastering the game of Go without human knowledge
-
Gabriel Aptekar

Gabriel Aptekar
John Fang
 
Nov 21 no class (Thanksgiving)      
Nov 26 Not RL Learning to act by predicting the future, Dosovitskiy etal. 2016

Evolution Strategies as a Scalable Alternative to Reinforcement Learning, Salimans etal. 2017
Daniel Brown
Dian Chen

Dian Chen
Daniel Brown
 
Nov 28 Fresh off the press Deep contextualized word representations, Peters etal 2018

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding, Devlin etal 2018
Tanya Goyal
Aditya Gupta

Manish Ravula
Daniel Crockett
 
Dec 3 Final project presentations presentation schedule 7 min per team   Project 2 due 1am
Dec 5 Final project presentations presentation schedule 7 min per team    
Dec 10 Fresh off the press Large Scale GAN Training for High Fidelity Natural Image Synthesis, Brock etal 2018

Neural Ordinary Differential Equations, Chen etal 2018
Abduallah Mohamed
Xing Han

Haresh Karnan
Manish Ravula
 

Paper assignment

All papers are assigned by a Deep Networks based on student preferences. If you don’t like your paper, please blame deep learning, tensorflow or your preference list (instead of the instructor). The code for the assignment problem will be make available.

Expected workload

Estimates of required effort to pass the class are:

General tips

Tips for reading/reviewing a paper

Tips for presentations

Notes

Syllabus subject to change.