Think about approaches to this problem

csndl-iitd / realtime-sleep-staging

Using ML for identification of sleep stages in real time in humans

MIT License

0 stars 0 forks source link

Think about approaches to this problem #4

Closed gsaurabhr closed 5 months ago

gsaurabhr commented 7 months ago

We discussed these two approaches today. You can learn more about them and how applicable they are to our data. Include your observations in comments below (one comment for each approach that summarizes that approach).

[ ] 3D CNNs: capture spatio-temporal dynamics across scales for effective classification. There might be other cool applications of 3D CNNs for EEG data, so you can think about those also!
[ ] Sequence to sequence transformers

gsaurabhr commented 6 months ago

Some other resources / techniques people have used: https://github.com/akaraspt/deepsleepnet https://github.com/akaraspt/tinysleepnet

gsaurabhr commented 5 months ago

Tanvi to work on 3D CNNs based approach #10.

Ayush to work on transformers based approach #11.