Hi fellow hackers,
Anyone here has experience using TensorFlow.js or has worked on Natural Language Processing area?
I have a few questions for you today:
1.What are your go to sources for data sets to training ML model?
2. What are some of the biggest challenges or this to watch out for when find a solution for NLP?
3. Are there many people or startups out there working on NLP (besides Google, Microsoft, Amazon)?
https://github.com/niderhoff/nlp-datasets
1- Datasets can be reached from different sources, depending on the problem you want to solve. Kaggle has public datasets and there some independent people/companies publishing datasets.
2- Again heavily depends on the problem you want to adress, the techniques you want to use.
3- There are lots of companies, institutions, people are working on NLP. Both in industry and academia.
I believe there are many companies in the NLP space. As an example I worked for Memsource where AI team worked on automated QA checks, trained on translation data from the customers.
Also, you haven't said anything specific. There are many different problems in NLP. What are you after?
As far as public datasets go, I don't have any experience here but I am curious too. I think books might be one of the obvious sources.