New content search algorithm for multi-documents w/ pytorch JupyterLab HuggingFace UMAP

Deep learning example w/ pytorch on JupyterLab for multi-document content analysis.

1. My multi-document: 125 pre-print from arxiv CS.CL on graph neural networks (full text)
3. My approach: apply all 45.500 sentences to my manifold of code, visualize emerging clusters and select specific clusters for further subclustering (a highly selective mode of NEW complex content search algorithm)
4. Benefits: topologically comparable to word2vec on a higher complexity manifold with semantic encoding on sentence level.

#sbert
#nlproc
#nlptechniques
#clustering
#semantic
#bert
#climatechange
#3danimation
#3dvisualization
#topologicalspace
#deeplearning
#machinelearningwithpython
#pytorch
#sentence
#embedding
#complex
#umap
#insight
#algebraic_topology