NLP in Clojure - Session 1 - summary & recording

[daslu]

Following a few recent suggestions, we organized a study meeting about NLP, possibly a first in a series.
This event is related to the upcoming ds4clj course – it is part of our exploration of topics that may later become part of the course.

Moderated: @kiramclean

The session began with an intro by Dimid Duchovny about Linguistics and Natural Language Processing.

Then, @daslu presented the use of the spaCy library, first from Python, and then from Clojure through Libpython-clj.

Towards the end, @Carsten_Behring added some comments about the broader picture of interop and machine learning in Clojure.

Zulip chat of the session

Video

NLP in Clojure Session 1

Notebook

(in editing, coming soon)

Next times

to be announced soon

[daslu]

Here is the intro part I presented (the hand-on part is coming soon):

nlp4clj session 1

Goals

• begin a series of study/exploration sessions around NLP in Clojure

• welcome friends from other tech backgrounds

• explore topics & methods towards the ds4clj course

• begin with a few basic ideas that some NLP systems use as their first layers

Disclaimers

• not an NLP expert

• just an intro to some basic methods

Resources for this session

• libpython-clj

• Tablecloth

• spaCy

• spaCy glossary

• Carin Meier’s blog

• some tutorials by Ethan Miller (need to check about the link)

• ATIS dataset (thanks, Dimid Duchovny)

Broader picture

• NLP for Developers by Rachael Tatman of the Rasa project

• DataLinguist for a Clojure library that covers similar areas (no Python needed)

• For the curious: the historical Zulip topic thread where Alan Marazzi and Chris Nuernberger were looking into spaCy, and then began looking into a Clojure-Python bridge

Let us explore

[beders]

This looks like a great start. Couldn’t join unfortunately, but will happily consume all this juicy content.
I’ve been working on a tool to make my mail inbox usable again and was planning on parsing and indexing all my emails - detecting concepts and relationships while doing so, so I can use RDF or something similar to query my inbox. (full text search won’t quite cut it)

[daslu]

@beders thanks!

Sounds wonderful. Please tell if you’d like to share your experience & practices in a future session.