Hello! Thanks for your question. I think there is a lot of potential for HTM to do natural language processing (NLP), especially when paired with the work that Cortical.io has been doing. They are a partner of Numenta, and have created an API that will return SDRs (they call them fingerprints) for words or phrases you give it. Here are some further resources you might want to look over:
- https://github.com/numenta/nupic/wiki/Natural-Language-Processing - https://www.youtube.com/watch?v=X4XjYXFRIAQ&start=7084 - https://www.youtube.com/playlist?list=PL3yXMgtrZmDoSxIPw26Y-Ai-X8tD1xcZD Regards, --------- Matt Taylor OS Community Flag-Bearer Numenta On Mon, Jan 25, 2016 at 9:12 PM, Kock Leong Ng <[email protected]> wrote: > Hihi, I'm a newbie to both Machine Learning (HTM, Deep learning etc). > > I'm not very technically inclined. > > Personally, I feel that HTM is a promising area and well done to everyone > involved. > > From the videos, I conclude that HTM algorithms are able to understand > patterns. > > I'm curious about whether HTM is able to understand meaning in > words/sentences. Eg. I give 10 sentences and the machine is able to > paraphrase my questions, or I give any sentence in English and the machine > will be able to convert to Chinese (with the correct meaning and > semantics/syntax). If not, will such development/research be in the > pipeline? > > Thanks > Ng Kock Leong >
