I am currently a data scientist in an AI start up (campus.swarma.org), and will join in the MACSS program this fall. I am interested in understanding the boundary between different fields, and transfer learning on mind.
Forced-directed Layout
Text Embedding and Evolution of the knowledge sky
Recommend System
Keyword Extraction
realized repulsive force to arrange the layout, adding the clearity of knowledge map in 2-D.
used BOW, Word2Vec to embed the text into vectors, T-sne+ to lower down dimension and evolute the knowledge sky .
(click here for a clip on Youtube)
implementing and modifying Deep Learning approaches(including GCN, GraphSAGE) to learn the pattern of learning habbit of users on a website(campus.swarma.org).
constructing the knowledge structure around 'complex system' and 'artificial intelligence' based on wikipedia pages.
To understand the individual and group's learning shock with Deep Learning including NLP and Graph neural network, especially to understand how analogy works for learning.
To understand how cities are shaped. (have explored London's traffic network, touched the interesting but seemingly superficial pattern of city's growth, so what else?)
Expressive ways to visualize research findings