|Implementing a type system in Python|
|Organisatsiooni nimi||ATI keeletehnoloogia|
|Kokkuvõte||Implementing a type system in Python (BSc/MSc)|
An implementation of NLC†, a type system for natural language compositionality, has been done in Coq (https://gitlab.com/eluuk/nlc) but in order to be usable in a real-word scenario, it should be implemented in a language well integrated w/ NLP, i.e. Python. The obvious problem is that Coq is an interactive theorem prover and a functional programming language with a complex static type system while Python is an imperative scripting language w/ only dynamic (duck) typing. The lack of static types in Python has to be leveraged somehow (e.g. w/ custom classes and/or set-theoretic operations) to accommodate the types of NLC. There are some ideas and code snippets on how this might work but you should contribute w/ more ideas, testing, analysis, and (eventually) developing. Ideally, you would feel comfortable w/ Python before choosing this topic. Depending on the projected result, this could be a BSc or MSc thesis topic. A knowledge of Estonian is a plus (but not required), as the system may be also tested on Estonian examples.
#Python #natural_language #type_system
|Lõputöö kaitsmise aasta||2019-2020|
|Suhtlemiskeel(ed)||eesti keel, inglise keel|
|Märksõnad||#Python #natural_language #type_system|