Thank you, Chainlinks, Hyperbolic Embeddings, Melanie Weber...

Hey guys,

Here comes my new blogpost. 

1. Thank you - Thanks a lot for messages on A. my field prize prediction, B. My Ricci Flow video & C. My talk abstract which was selected for a satellite conference. 

2. Chainlinks - A lot of you are wondering about my tweet on Hodge Decomposition of information flow over complex networks - The reason was hodge decomposition of information flow over 2 narrow networks - regular economy and crypto economy (including my new favourite ChainLinks) hence EconoPhysics 

3. Hyperbolic Embeddings - https://dawn.cs.stanford.edu/2018/03/19/hyperbolics/ Created by Stanford (I often use this in my projects since it has been pre-trained on UMLS). 

This has also been pre-trained on embeddings from MusicBrainz. Apps & companies like Gaana, Saregamapa - might get some help in improving their recommendation system - The Indian music industry is valued at Rs 1500 Cr and has grown at 10.3% CAGR in the past five years... Spotify, YouTube Music might use this apart from reinforcement learning on user data.

4. Melanie Weber - This paper by Melanie Weber is really good as an application of Ricci Flow in big data - Forman-Ricci Flow for Change Detection in Large Dynamic Data Sets - https://web.math.princeton.edu/~mw25/publication/axioms/