MLP 2018 - Machine Learning for Programming

The quest for techniques to reduce software development efforts and costs, by generating correct code or catching bugs as early as possible, has been permanent since the beginning of computer science as a discipline, and the verification community has played a central role in this effort. The last five years have seen an enormous increase in the amount of code that is publicly available (for example, in repository systems such as github) and this gave birth to a research effort known as "Big Code", which consists in applying data mining techniques to software programs. At the same time, progress made in the field of Machine Learning have started to broaden the range of applications of AI to more complex data sets, with logical structure and semantics.
In this context a new area of research emerged, with the aim of applying deep learning and statistical methods trained on the immense amounts of code available. Researchers in the formal methods are increasingly using statistical methods as heuristics driving non-statistical techniques, for instance, to guide the exhaustive exploration of search spaces or derive candidate invariants.
More generally, the interplay between statistical/machine learning techniques and the areas more traditionally associated with program analysis (formal methods, programming languages, software engineering) has sparked fruitful research. While this research has been recognised as relevant and results were published in top venues of verification, programming languages, and verification, there is no venue dedicated exclusively to this topic.
Our aim is to gather the most relevant researchers in the area to co-organise and attend the workshop so that the separate efforts start to build in the context of a community.
Recent events (e.g. ML4PL 2015, Neural Abstract Machines & Program Induction, LiVe 2017, NL+SE) indicate a great interest in this area. We expect that, in further years, the community behind the MLP and related workshops will lead to creation of a new field at the intersection between verification, machine learning, and software engineering; we aim to accelerate this effort and help ensure it leads to industrial applications.