Combining SEPIA and ML4PG
نویسنده
چکیده
SEPIA is an approach that infers state based models from existing proof corpora. These models can then be used for the development of new proofs. As with similar approaches, selecting the best facts to use in new proofs is challenging. We investigate the potential use of ML4PG as a relevance filter for our approach achieving a model only inferred from the lemma suggestions by ML4PG. These reduced models still contain the neccesary tactics to achieve proofs.
منابع مشابه
HoTT formalisation in Coq: Dependency Graphs \& ML4PG
This note is a response to Bas Spitter's email of 28 February 2014 about ML4PG:"We (Jason actually) are adding dependency graphs to our HoTT library: https://github.com/HoTT/HoTT/wiki I seem to recall that finding the dependency graph was a main obstacle preventing machine learning to be used in Coq. However, you seem to have made progress on this. What tool are you using? https://anne.pacalet....
متن کاملML4PG in Computer Algebra Verification
ML4PG is a machine-learning extension that provides statistical proof hints during the process of Coq/SSReflect proof development. In this paper, we use ML4PG to find proof patterns in the CoqEAL library – a library that was devised to verify the correctness of Computer Algebra algorithms. In particular, we use ML4PG to help us in the formalisation of an efficient algorithm to compute the inver...
متن کاملProof Pattern Search in Coq/SSReflect
ML4PG is an extension of the Proof General interface of Coq, allowing the user to invoke machine-learning algorithms and find proof similarities in Coq/SSReflect libraries. In this talk, we will show the recent ML4PG features in action, using examples from the standard SSReflect library and HoTT library. We will compare ML4PG with traditional Coq searching tools and dependency graphs.
متن کاملMachine Learning in Proof General: Interfacing Interfaces
We present ML4PG — a machine learning extension for Proof General. It allows users to gather proof statistics related to shapes of goals, sequences of applied tactics, and proof tree structures from the libraries of interactive higher-order proofs written in Coq and SSReflect. The gathered data is clustered using the state-of-the-art machine learning algorithms available in MATLAB and Weka. ML4...
متن کاملFunctional properties of fish protein hydrolysates from Cuttlefish (Sepia pharaonis) muscle produced by two commercial enzymes
Fish protein hydrolysates were prepared from Cuttlefish (Sepia pharaonis) muscle using alcalase and protamex methods. By conducting the Cuttlefish protein hydrolysate using alcalase (CPHA), it showed higher degrees of hydrolysis, proximate composition, yield, emulsifying activity index (EAI), emulsion stability index (ESI), foaming activity index (FAI), solubility, water holding capacity (WHC) ...
متن کامل