How to Predict Music You Love (LIVE)

In this video, we're going to look at several different type of recommender systems in an iPython notebook. Popularity based, item-item collaborative, then user-item collaborative. Then we'll touch on the bleeding edge in deep learning at the end. Also I freestyle. Twice lol.

Code for this video:

More learning resources:

Join us in our Slack channel:

Please Subscribe. And Like. And comment. That's what keeps me going.

And please support me on Patreon:
Follow me:
Signup for my newsletter for exciting updates in the field of AI:

[shreyas]

You're a legend man! I've learnt a hell lot from your videos than I've learned from my classes

[rocklobster]

i still think you're doing FANTASTIC. you real stike me as an absolute star in the making. in both your productions and cause.
i still haven't gotten the balls to play with python, but youre effectively nudging me closer and closer
Thank you

[mockingbird]

How can I install Recommender and Evaluation

[talfstad]

EPPIIIC. Thank you so much!!! I REALLY appreciate you teaching all of this!

[TpBrass]

I prefer live coding more!(Although I can only watch the recorded version)
With live coding, I'm not only learn the code itself but also know how you solving a problem or achieving a goal, this is more worth to me.

[ravitanwar]

beautiful, you should come up with more of these

[SirCutRy]

An interesting thing to do would be to have an unsupervised learning AI find how similar music files are and then recommend new pieces of music to a user based on what they previously liked and what is most similar to that music.

[eturnerx]

Don't live code the data loading and massaging parts. Live code the ML please.

[rouanloup]

Hi Siraj, I really like your videos. Would be nice if you can share experiences or explore binary neural net, using XOR function instead sigmoids to accelerate inference :)

[prani]

So this code does not use machine learning or any deep learning techniques, am i right???

[florentinalexandruiftimie]

Hey Siraj, can you make a video about IBM Watson? Like how does it work in the Jeopardy competition?

[secondsandthings]

You've probably got a lot of videos you want to do, but one going into detail for one shot neural nets (Memory Augmented Neural Networs) would be interesting!

[tanvirhossain]

i tried to run it on google colab . got many errors. what should i do?

[javisartdesign]

See you in Udacity QA. I have recently heard a podcast about something related with spreadsheets and Markov chain models for this type of predictions you're talking about. Do you think it works better than normal NN? I would like the retake again with Deep Learning, I learned Theano framework, because Tensorflow wasn't released yet. Do you recomend to switch to TF? thanks for your time!! you're amazing man

[wafgear]

19:33 Should percentage sum to one? For me it seems the sum should be 100 rather than 1 since it's percentage, not the ratio

[sudharsan]

Hey siraj,

Can we use your recommender file as benchmark mark design to build custom recommender system for various business case ?
I hope you help me

[sweetyadwani]

Hey Siraj, can you tell us how the score and rank for the songs is calculated? Also, on adding new songs, the predictive model fails, why is it so?

[AdarshSingh-qkrj]

hey u didnt exlplained the Evaluation part of the program...:(

[joegarza]

hi, siraj . very nice . u got more then 3000.

[priyankavaidya]

Hey Siraj, can you pls tell how and on which platform should i run the code ???