A few weeks ago, I gave a demo on word2vec in the Social Dynamics Laboratory. I based the notebook on the one I used while I was in the Facebook Artificial Intelligence PyTorch Scholarship Challenge, but added code to import all of the data within the notebook using some terminal commands. I've found Google Colab to be quite useful for distributing tutorials like this, but I still think Google has a long way to go to make Colab easier to work with. Importing and exporting data to your drive can be complicated. If anyone has any tips on how to do these tasks easier please share (and thanks in advance)!
If anyone wants to know more about word2vec or any other embedding/language models for language, events, social network graphs, etc., please let me know! I've worked with a range of them over the last two years (e.g., the graph embedding plot in my Research page is a metapath2vec embedding for multi-relational network graphs) and am happy to share my knowledge (and to learn about new techniques as well)! I'll be talking about embedding models and my experiences with them more. The next model I plan to put into use is the DeepInf model for evaluating social influence in dynamic graphs. I'm hoping to add text and other contextual information into the model pipeline in addition to other data related to social interactions over time. Click here to read the original word2vec paper.
0 Comments
Your comment will be posted after it is approved.
Leave a Reply. |
|