Lesson 8 in-class

This is a good place to start: https://www.reddit.com/r/MachineLearning/

They have a what are you reading thread each week that is interesting.

1 Like

Will Generative cover Generative adversarial networks?

How is colorising b/w images NOT just a regression problem?

@jpuderer Yes

1 Like

http://www.arxiv-sanity.com/

7 Likes

You can find research papers on ArXiv.org. Insteresting sections for this course would be:

An easy way to get updated on what is new is to subscribe to the arXiv RSS feed of these sections. To get the rss feed for a specific section you just replace ‘list’ by ‘rss’ in the URL.

To get a ‘feel’ for a field that you don’t know, you can subscribe to the corresponding arXiv section using Feedly where you can see what other people like and look at the most liked recent articles.

4 Likes

How do we download the python notebooks from here?

http://www.platform.ai/part2/lesson1/

Also: https://www.reddit.com/r/MachineLearning/

@harveyslash I’m curious. How can you formulate it as a regression problem? Could you please elaborate.

wget -r -nH -nd -np -R index.html* http://www.platform.ai/part2/lesson1/

3 Likes

given an input, predict the color(R,G,B) value of each pixel.
So your output will be image_width * image_height * 3.
Its just my guess

@harveyslash this would just result in a lookup table with 255 entries, your output will be limited to just 255 colors.

Thank you!!

1 Like

this wouldn’t take into account surrounding pixels and translation invariance (the way CNNs can)

2 Likes

Does prisma app use something similar to come up with these cool filters?

Where can I find the Jupyter theme Jeremy is using?

3 Likes

Is it better to calculate f_content for a higher layer for VGG and use a lower layer for f_style sine the higher layer abstracts are captured in the higher layer and the lower layer captures textures and “style”?

http://www.platform.ai/part2/lesson1/neural-style.ipynb

@nikesh Jupyter color themes: https://github.com/dunovank/jupyter-themes

1 Like

i had an error running it, did anyone else have this issue?

ModuleNotFoundError: No module named 'xgboost'