Hacker News Reader

Comment by gtani | Hacker News Reader

Frankly if you give more detail on what kind of NLP, we can give more pointers, but for comprehensive resource, 3rd draft Jurafsky/martin standard text, missing pieces TBA: LSTM's, vector embeddings (word2vec etc), seq2seq/neural translation etc

https://web.stanford.edu/~jurafsky/slp3/

(you can email questions anytime)

View on HN

Thanks for the lectures, I'll look through it, I've started to watch their lecture form Jurafsky and Manning that is available on yt.

Currently It's mostly about topic/category recognition in articles and the other part are entities, how to do entity linking and the entities relevance for the articles.

Maybe later on I'll have to look for synonyms (wordNet and word embeddings can be useful for this), but this is something for the future.

Things like topic recognition, entity and reference linking/disambiguation, synonym/paraphase are well researched tasks with purpose-made datasets for competitions like TREC. The most common approaches will be to choose a word embedding (one of the word2vec or Glove are good starts) and run that into a CNN or RNN variant: https://arxiv.org/abs/1702.01923

https://arxiv.org/abs/1611.09100

https://metamind.io/research/learning-when-to-skim-and-when-...

https://arxiv.org/abs/1611.06792

https://arxiv.org/abs/1611.03305

probably if you search arxiv and the ACL anthology you should find good lit searches/historical summaries: http://aclweb.org/anthology/ Karpathy's arxiv sanity preserver is a big help. You can put together a list of maybe a few dozen twitter accounts and hashtags to watch, that'll keep you pretty current.

Reply on news.ycombinator.com