Inspecting Neural Networks with CCA - A Gentle Intro (Explainable AI for Deep Learning)

[@gergerger53]

Correct on the Scandinavian points. Written Norwegian is just historically imported Danish (they even call it, "Book language") with a few tiny changes added later (they can almost perfectly read each other's languages) and Swedish is also very similar. They're part of a dialect continuum and not really different "languages" from the perspective of a lack of mutual intelligibility. It's for political and historical reasons we consider them as such, today. Main differences lie in the spoken language, anyway. Written, which is what NLP models see, they're all very, very similar.

[@arp_ai]

Very interesting! Thanks for highlighting that!

[@pubgkiller2903]

Hi Jay... I am very happy to see your new video. Although I didn’t go through yet but I am happy that you start porting videos on fast track ... thanks again

[@Mrbits01]

Great overview. Subbed!

[@AMFLearning]

nice, amflearning by doing

[@brandomiranda6703]

What is the intuition/argument that the SV part of SVCCA does better specifically for NN (or was it noise? or both?)?

[@brandomiranda6703]

I'm also curious, what is the main idea of why PW is more robust to noise?

[@AyushSingh-vj6he]

Hey, this is Ayush, I have something important to ask but may I have your email