CS-498 Applied Machine Learning - Homework 8
CS-498 Applied Machine Learning
D.A. Forsyth --- 3310 Siebel Center
daf@uiuc.edu, daf@illinois.edu
13:00 - 14:15 OR 1.00 pm-2.15 pm (in old fashioned time)
WF
1404 Siebel Center
TA's:
Tanmay Gangwani gangwan2@illinois.edu
Jiajun Lu jlu23@illinois.edu
Jason Rock jjrock2@illinois.edu
Anirud Yadav ayadav4@illinois.edu
Office Hours
- DAF: Mon 10-11 and Fri 2:30 - 3:30
- Jason: Mon 11-12 and Tues 1:30 - 2:30
- Jiajun: Tues 2:30 - 3:30 and Wed 5 - 6
- Anirud: Wed 1-2 and Thur 4 - 5
- Tanmay: Mon 4 - 5 and Fri 4 - 5
DAF Mon - 14h00-15h00, Fri - 14h00-15h00
or swing by my office (3310 Siebel) and see if I'm busy
Evaluation is by: Homeworks and take home final.
I will shortly post a policy on collaboration and plagiarism
Homework 8: Due 8 May 2017 23h59 (Mon; midnight)
You should do this homework in groups of up to three; details of how to submit have been posted on piazza. You must use tensorflow.
Details and description subject to minor changes
Convolutional neural networks: Tensorflow is an environment for building, among other things, neural networks. You can build multilayer networks; convolutional layers; and residual networks using tensorflow.
There is a tool called tensorboard, which will display the progress of learning.
- Obtain and install Tensorflow, here. Go through the "MNIST for beginners" tutorial here. Now go through the "DEEP MNIST for experts" tutorial here. Now go through the tensorboard tutorial here. Finally, insert appropriate lines of code into the tensorflow example to log the accuracy on tensorboard every 100 batches, for at least 2000 batches. You should screen capture the accuracy graph from tensorboard, and submit this.
- Modify the architecture that is offered in the DEEP MNIST tutorial to get the best accuracy you can. I made three convolutional layers of smaller depth (i.e. the 32 went to 8), dropped the max pooling, and used three layers.
Submit a screen capture of tensorboard graphs of accuracy. We will run a leaderboard of these graphs.
- Go through the CIFAR-10 tutorial here, and ensure you can run the code. Finally, insert appropriate lines of code into the tensorflow example to log the accuracy on tensorboard every 100 batches, for at least 2000 batches. You should screen capture the accuracy graph from tensorboard, and submit this.
- Modify the architecture that is offered in the CIFAR-10 tutorial to get the best accuracy you can. Anything better than about 93.5% will be comparable with current research. Be aware that people with bigger computers will likely do better at this exercise (so I won't make grades depend on accuracy). Submit a screen capture of tensorboard graphs of accuracy. We will run a leaderboard of these graphs.