Machine Learning for Programming Languages Workshop

Machine Learning for Programming Languages Workshop
Name	Machine Learning for Programming Languages Workshop
Status	Active
Discipline	Computer science
Frequency	Annual
Venue	Varies
First	2015
Organizer	Academic and industrial research groups

Machine Learning for Programming Languages Workshop is an interdisciplinary scientific meeting that explores intersections between Alan Turing-inspired computation theory, Geoffrey Hinton-era neural representation, and practical systems from Ken Thompson-style compiler engineering. The workshop convenes researchers from institutions such as Massachusetts Institute of Technology, Stanford University, University of Cambridge, Carnegie Mellon University, and companies including Google, Microsoft, Facebook, DeepMind to exchange work bridging formal languages, Yoshua Bengio-influenced machine learning, and systems research. It attracts authors who have published at venues like NeurIPS, ICML, ICLR, PLDI, POPL, ICSE and OOPSLA.

Overview

The workshop focuses on techniques that combine ideas from John McCarthy-style symbolic AI, Donald Knuth-era algorithmics, and Leslie Lamport-level formal verification with contemporary Yann LeCun-linked deep learning models. Topics include probabilistic programming influenced by Peter Norvig, program synthesis drawing on work by Ravi Chandra, differentiable interpreters following research at University of California, Berkeley, neural program induction connected to Judea Pearl-adjacent causality, and hybrid systems inspired by Robin Milner. Attendees often come from labs at ETH Zurich, Princeton University, University of Oxford, Imperial College London, and companies such as Amazon, IBM, Apple, NVIDIA, Intel.

History and Origins

The workshop lineage traces to early meetings in the mid-2010s influenced by results from Google DeepMind teams and research groups at MIT CSAIL, Microsoft Research, and Berkeley AI Research (BAIR). Foundational antecedents include program induction work by researchers associated with University of Toronto and theory advances from California Institute of Technology. Initial organizers included faculty from Harvard University, Cornell University, Brown University, and industrial researchers from OpenAI. Early keynote speakers have included scholars affiliated with Royal Society, recipients of the Turing Award, and contributors to ACM conferences.

Topics and Themes

Common themes span neural architectures influenced by Ian Goodfellow, probabilistic inference tied to Thomas Bayes-style modelling, and symbolic reasoning rooted in Alonzo Church-derived lambda calculus. Research presented connects program synthesis variants related to work at SRI International, program repair reminiscent of projects at NASA, and semantic parsing influenced by publications from Google Research. Other recurring themes relate to automated theorem proving from Stanford University groups, type systems advanced by University of Washington researchers, and compiler optimization studies originating in Bell Labs-like traditions.

Organization and Format

Typically collocated with major conferences such as NeurIPS, ICML, ICLR, PLDI, POPL, or ACL, the workshop features invited talks, paper presentations, poster sessions, and dedicated discussion panels. Organizing committees have included faculty from Dartmouth College, University of California, San Diego, University of Illinois Urbana-Champaign, University of Toronto and representatives from Salesforce Research and Huawei. Program chairs have been affiliated with prizes and honors tied to the ACM SIGPLAN community and institutions like Max Planck Institute and École Polytechnique Fédérale de Lausanne.

Key Papers and Contributions

Notable contributions include advances in neural code generation influenced by models developed at OpenAI and algorithmic learning techniques similar to work by Leslie Valiant. Papers have explored differentiable interpreters, learned type inference, and neural program repair with connections to legacy research at AT&T Bell Laboratories and modern projects at Facebook AI Research. Benchmarks and datasets introduced by workshop authors have been adopted by groups at MIT, Stanford, Carnegie Mellon University, and Tsinghua University. Methodological links appear to research agendas at Google Brain, Microsoft Research Redmond, Baidu Research, and academic groups such as University of Edinburgh.

Participants and Community

The community comprises academics, industry researchers, and students from institutions including University of California, Los Angeles, Yale University, Columbia University, New York University, University of Michigan, University of California, Irvine, Peking University, Zhejiang University, Seoul National University, and KAIST. Industry participation has included teams from Dropbox, Adobe Research, LinkedIn, Intel Labs, Samsung Research, and Siemens. The workshop fosters collaborations with research consortia and funding bodies such as National Science Foundation-supported projects and partnerships with centers like Alan Turing Institute.

Impact and Future Directions

The workshop has influenced subsequent work at flagship venues including NeurIPS and PLDI, informed industrial tooling at GitHub and Microsoft Visual Studio, and inspired open-source efforts hosted by organizations like Apache Software Foundation and Linux Foundation. Future directions emphasize tighter integration with formal methods from INRIA, scalable systems reflecting designs from Oracle Corporation, and interdisciplinary collaboration with researchers at Broad Institute and Scripps Research. Anticipated trends include incorporation of causal inference approaches associated with Causality Research Group initiatives, improved benchmarks from international labs, and broader uptake by practitioners across publishing venues such as AAAI and IJCAI.

Category:Workshops in computer science