JEFL: Joint Embedding of Formal Proof Libraries
This addresses a domain-specific problem for users of interactive proof assistants by enabling better concept retrieval across libraries, though it appears incremental as it builds on existing methods like fasttext.
The paper tackles the problem of discovering similar mathematical concepts across different interactive proof assistant libraries with heterogeneous logical foundations, comparing a previously proposed matching algorithm with an unsupervised embedding approach based on fasttext and tree traversal, and argues that the neural embedding approach has more potential for integration into interactive proof assistants.
The heterogeneous nature of the logical foundations used in different interactive proof assistant libraries has rendered discovery of similar mathematical concepts among them difficult. In this paper, we compare a previously proposed algorithm for matching concepts across libraries with our unsupervised embedding approach that can help us retrieve similar concepts. Our approach is based on the fasttext implementation of Word2Vec, on top of which a tree traversal module is added to adapt its algorithm to the representation format of our data export pipeline. We compare the explainability, customizability, and online-servability of the approaches and argue that the neural embedding approach has more potential to be integrated into an interactive proof assistant.