ruborcalor

How do you think this compares to coworking?

Hello HN!

The goal of Artifai is to help people generate art from photos. I had a problem: My walls were bare. I wanted to decorate them with memories, but I wanted the decorations to be artistic at the same time.

This lead me to discover neural-style, an algorithm that takes two images as input, and copies the style from one onto the other while retaining its content. Neural-style lets you take a meaningful photo and turn it into awesome art! There were several implementations available on Github, but I didn't see a way for non-technical users to easily use it and have the result shipped to their home. And so I began building Artifai with a group of friends; We are a group of university students and this was our first fullstack app!

You can read more about the technical details here: https://docs.google.com/document/d/1Q63JH2CnRWZbhO7oDHEeAWJT...

Thanks for checking it out and let us know if you have any feedback! :)

Wow very cool startup; wish you guys the best of luck!

Definitely a good tip. I'll try not to make that mistake again.

You can now find the jupyter notebook code here: https://github.com/Ruborcalor/Age-Prediction-Via-Blood-Sampl...

what a meme haha :)

after many similar comments I plan on implementing a simpler model and seeing how it compares

Yeah i'd agree that ML hype is too real I guess I'm part of the problem

It's more fun to use a neural net :) but after many similar comments I plan on implementing a simpler approach and seeing how it compares

Wow very cool tool! I didn't know about this somehow. Thanks for sharing. Cool that you've tried it on yourself.

What would you prefer to call it? Machine learning?

yes valid point haha. It's more fun to use a neural net :) but after many similar comments I plan on implementing a simpler approach and seeing how it compares

Thanks I really appreciate it.

I'll try and get back to you with the performance of k-folds validation and shuffling.

I don't think it can be learning the order or samples because the train and test data sets are separated very early on. If it were learning order or samples of the training set it would have to perform very poorly on the test set.

Yes I agree you want number of samples to be greater than number of features. This is why the number of features were reduced from over 400,000 to 25. After this reduction the number of features is less than the number of samples (~700).

Honestly I didn't prioritize hyperparameter tuning enough. I pretty much went with one of the first models I identified.

Could you elaborate on the idea of not capturing dependent features please?

I'll try out these models and get back to you with their effectiveness.

I apologize for the confusion, i've since removed this typo.

The data is only split once, before using a correlation test to select the features that the model would be trained on. As far as I can tell there is no data snooping occurring because the data is split into train and test sets before any decision are made.

Haha yes good point take everything with a grain of salt

Yes the Horvath Clock seems to be the standard.

Thanks for sharing the book i'll have to sheck it out!

Interesting point, the idea of transfusing blood from healthy subjects had never occurred to me.

I'm not sure I understand; what was the claim from the startups, and what does the rumor mill say is happening at a low level?