Machine learning has seen a surge of interest in both research and practice. From natural language processing to self-driving cars, machine learning is creating new possibilities that are changing the way we live and interact with computers. However, the impact of these advances on programming languages remains mostly untapped. Yet, incredible research opportunities exist when combining machine learning and programming languages in novel ways.
This symposium seeks to bring together programming language and machine learning communities to encourage collaboration and exploration in the areas of mutual benefit. The symposium will include a combination of rigorous peer-reviewed papers and invited events. The symposium welcomes papers on all aspects that combine programming languages and machine learning including (and not limited to):
- Machine learning methods trained over source code, including applications to software development tools and program synthesis
- Application of machine learning to compilation and run-time scheduling
- Collaborative human / computer programming (i.e., conversational programming)
- Deterministic and stochastic program synthesis
- Infrastructure and techniques for mining and analyzing large code bases
- Interoperability between machine learning frameworks and existing code bases
- Probabilistic and differentiable programming
- Programming language and compiler support for machine learning applications
- Programming language support and implementation of machine learning frameworks
- Neurosymbolic and intentional programming
Mon 13 JunDisplayed time zone: Pacific Time (US & Canada) change
09:00 - 10:00 | |||
09:00 45mKeynote | Competitive Programming with AlphaCodevirtual MAPS Yujia Li Deepmind | ||
10:30 - 12:00 | |||
10:30 45mKeynote | Improving Software Reliability using Machine Learningvirtual MAPS Baishakhi Ray Columbia University | ||
11:15 15mTalk | Productivity Assessment of Neural Code Completion MAPS Albert Ziegler GitHub, Eirini Kalliamvakou GitHub, X. Alice Li GitHub, Andrew Rice GitHub, Devon Rifkin GitHub, Shawn Simister GitHub, Ganesh Sittampalam GitHub, Eddie Aftandilian GitHub Pre-print | ||
11:30 15mTalk | Predictive Synthesis of API-Centric Code MAPS Daye Nam CMU, Carnegie Mellon University, Baishakhi Ray Columbia University, Seohyun Kim Meta, xianshan qu , Satish Chandra Facebook Pre-print | ||
11:45 15mTalk | From Perception to Programs: Regularize, Overparameterize, and Amortize MAPS | ||
13:30 - 15:00 | |||
13:30 45mKeynote | Can Transformers Code?virtual MAPS Łukasz Kaiser OpenAI | ||
14:15 15mTalk | Syntax-Guided Program Reduction for Understanding Neural Code Intelligence Modelsvirtual MAPS Md Rafiqul Islam Rabin University of Houston, Aftab Hussain University of Houston, Amin Alipour University of Houston DOI Pre-print | ||
14:30 15mTalk | A Systematic Evaluation of Large Language Models of Codevirtual MAPS Frank F. Xu Carnegie Mellon University, Uri Alon Carnegie Mellon University, Graham Neubig Carnegie Mellon University, Vincent J. Hellendoorn Carnegie Mellon University | ||
14:45 15mPoster | Poster Session MAPS | ||
15:30 - 17:00 | |||
15:30 45mKeynote | Unsupervised Program Synthesis: Hierarchy and Perception MAPS Kevin Ellis Cornell University | ||
16:15 15mTalk | ExeBench: An ML-scale dataset of executable C functions MAPS Jordi Armengol-Estapé University of Edinburgh, Jackson Woodruff University of Edinburgh, Alexander Brauckmann University of Edinburgh, José Wesley de Souza Magalhães University of Edinburgh, Michael F. P. O'Boyle University of Edinburgh | ||
16:30 15mTalk | Automatically Debugging AutoML Pipelines Using Maro: ML Automated Remediation Oracle MAPS | ||
16:45 15mTalk | A Graph Neural Network-based performance model for Deep Learning Applications MAPS Shikhar Singh University of Texas, James Hegarty Facebook, Hugh Leather University of Edinburgh, UK, Benoit Steiner Facebook | ||
22:30 - 00:00 | |||
22:30 45mKeynote | Improving Software Reliability using Machine Learningvirtual MAPS Baishakhi Ray Columbia University | ||
23:15 15mTalk | Productivity Assessment of Neural Code Completion MAPS Albert Ziegler GitHub, Eirini Kalliamvakou GitHub, X. Alice Li GitHub, Andrew Rice GitHub, Devon Rifkin GitHub, Shawn Simister GitHub, Ganesh Sittampalam GitHub, Eddie Aftandilian GitHub Pre-print | ||
23:30 15mTalk | Predictive Synthesis of API-Centric Code MAPS Daye Nam CMU, Carnegie Mellon University, Baishakhi Ray Columbia University, Seohyun Kim Meta, xianshan qu , Satish Chandra Facebook Pre-print | ||
23:45 15mTalk | From Perception to Programs: Regularize, Overparameterize, and Amortize MAPS | ||
Tue 14 JunDisplayed time zone: Pacific Time (US & Canada) change
01:30 - 03:00 | |||
01:30 45mKeynote | Can Transformers Code?virtual MAPS Łukasz Kaiser OpenAI | ||
02:15 15mTalk | Syntax-Guided Program Reduction for Understanding Neural Code Intelligence Modelsvirtual MAPS Md Rafiqul Islam Rabin University of Houston, Aftab Hussain University of Houston, Amin Alipour University of Houston DOI Pre-print | ||
02:30 15mTalk | A Systematic Evaluation of Large Language Models of Codevirtual MAPS Frank F. Xu Carnegie Mellon University, Uri Alon Carnegie Mellon University, Graham Neubig Carnegie Mellon University, Vincent J. Hellendoorn Carnegie Mellon University | ||
02:45 15mPoster | Poster Session MAPS | ||
03:30 - 05:00 | |||
03:30 45mKeynote | Unsupervised Program Synthesis: Hierarchy and Perception MAPS Kevin Ellis Cornell University | ||
04:15 15mTalk | ExeBench: An ML-scale dataset of executable C functions MAPS Jordi Armengol-Estapé University of Edinburgh, Jackson Woodruff University of Edinburgh, Alexander Brauckmann University of Edinburgh, José Wesley de Souza Magalhães University of Edinburgh, Michael F. P. O'Boyle University of Edinburgh | ||
04:30 15mTalk | Automatically Debugging AutoML Pipelines Using Maro: ML Automated Remediation Oracle MAPS | ||
04:45 15mTalk | A Graph Neural Network-based performance model for Deep Learning Applications MAPS Shikhar Singh University of Texas, James Hegarty Facebook, Hugh Leather University of Edinburgh, UK, Benoit Steiner Facebook | ||
Accepted Papers
Call for Papers: 6th Annual Symposium on Machine Programming (previously Machine Learning and Programming Languages)
(Deadline extended! See new deadline at right.)
Machine learning has seen a surge of interest in both research and practice. From natural language processing to self-driving cars, machine learning is creating new possibilities that are changing the way we live and interact with computers. However, the impact of these advances on programming languages remains mostly untapped. Yet, incredible research opportunities exist when combining machine learning and programming languages in novel ways.
This symposium seeks to bring together the programming languages and machine learning communities to encourage collaboration and exploration in the areas of mutual benefit. The symposium will include a combination of rigorous peer-reviewed papers and invited events. The symposium welcomes papers on all aspects that combine programming languages and machine learning including (and not limited to):
- Machine learning methods trained over source code, including applications to software development tools and program synthesis
- Application of machine learning to compilation and run-time scheduling
- Collaborative human / computer programming (i.e., conversational programming)
- Deterministic and stochastic program synthesis
- Infrastructure and techniques for mining and analyzing large code bases
- Interoperability between machine learning frameworks and existing code bases
- Probabilistic and differentiable programming
- Programming language and compiler support for machine learning applications
- Programming language support and implementation of machine learning frameworks
- Neurosymbolic and intentional programming
- Recommendation systems for programming
Important Dates
Submission deadline: Mar 18, 2022 Author notification: Apr 21, 2022 Final papers due: May 5, 2022 Workshop date: Jun 13, 2022
Evaluation Criteria
As in previous years, reviewers will evaluate each contribution for its significance, originality, and clarity to the topics listed above. Submissions should clearly state the novelty of their research contributions and how they improve upon any existing work.
Evaluation will be double-blind and papers must be properly anonymized. This means that author names and affiliations must be omitted from the submission. Additionally, if the submission refers to prior work done by the authors, that reference should be made in third person. These are firm submission requirements. If you have questions about making your paper double blind, please contact the Program Chair.
Review Process
The review process will be double-blind, managed through Open Review. During the review process, submissions and reviews will be confidential. After the acceptance notification, reviews of accepted papers will be made public through the Open Review web site, and public discussion comments will be enabled before and during the symposium to allow for broader discussion. Anonymous discussion comments will not be allowed.
All accepted papers will appear in the published proceedings and be available on the ACM Digital Library. Authors must be familiar with and abide by SIGPLAN’s republication policy, which forbids simultaneous submission to multiple venues and requires disclosing prior publication of closely related work.
Broader Impact
Due to the growing concerns regarding potential positive and negative impacts of any research work, as in previous years, the authors of MAPS submissions are asked to include a section on the potential broader impact of their work. This section should highlight an evaluation of potential misuses and negative impacts of the presented technology on its users and those indirectly affected, such as their friends and families, communities, society, and the planet. Authors should ponder and discuss the negative outcomes of their research 1) using its current form, 2) if enhanced in the future with new capabilities. They should also discuss potential ways to mitigate those harms (policy, law, alternative design choices, etc.).
The broader impact section will be outside the page limit of the original paper. This section should be at least one paragraph but should not exceed 1 page. Although this section is a must-have, this year, the decision to accept the paper will not be influenced by the discussed negative impacts. However, it might influence the acceptance decision in future MAPS.
Some helpful tips to think about broader impact.
Paper Submissions
Submissions must be in English. Papers should be in PDF and format and no more than 8 pages in standard two-column SIGPLAN conference format including figures and tables but excluding references and appendices. Submissions must be made through Open Review: https://openreview.net/group?id=PLDI.sigplan.org/2022/Workshop/MAPS.
| Mon 13 Jun 2022 Workshop date |
| Thu 5 May 2022 Final papers due |
| Thu 21 Apr 2022 Author notification |
| Fri 25 Mar 2022 Submission deadline |
Farhana Aleen
NVIDIA
Uri Alon
Carnegie Mellon University
United States
Osbert Bastani
University of Pennsylvania
United States
Satish Chandra
Facebook
United States
Xinyun Chen
Vinod Grover
NVIDIA
United States
Vincent J. Hellendoorn
Carnegie Mellon University
United States
Jeevana Priya Inala
MIT
United States
Aditya Kanade
Indian Institute of Science, Bangalore
Baishakhi Ray
Columbia University
United States
Xujie Si
McGill University, Canada
Rishabh Singh
Google Brain
United States
Earl T. Barr
University College London
Abdul Wasay
Intel Labs and UCSD
United States