Trying to use NLP and suitable knowledge representation to make a smart, learning and reasoning friend that is able to answer all your questions.
With today's NLP approaches the processing of human language has made big progress, but the understanding of the contents by semantics is still not achieved. I want to find out where the challenges still are, and which approaches might work and which limits they have.
- build up a knowledge base by processing natural language texts (texts written to be understood/processed by humans, like e.g. articles in Wikipedia or on news-websites) or from textual conversation (pseudo messenger communication interface)
- answer questions by processing/evaluating/searching the knowledge base incl. reasoning
- Python3
- pip3
pip install the following:
- bs4
- html5lib
- matplotlib
- nltk
- word2number
Then open the Python console and do:
import nltk
nltk.download()
Then install all.
If you encounter problems reading webcontent with Python using bs4 (done in some examples), make sure to install the certificates:
- goto Macintosh HD/Applications/Python3.x/
- execute "Install Certificates.command"