Machine learning algorithms have been shown to generalize poorly on combinatorially demanding tasks. Recent research has demonstrated that merging combinatorial optimization with machine learning methods enables solving problems that require non-trivial combinatorial generalization beyond pattern matching. In this spirit, this workshop aims to bring the communities (machine learning and combinatorial optimization, operations research) together in order to motivate further research at the intersection. This involves:
- Machine learning approaches aimed at improving combinatorial algorithms/solvers.
- Machine learning techniques to directly learn solvers for combinatorial problems.
- Hybrid architectures; pipelines containing both algorithmic/combinatorial and standard NN building blocks.
- Applications of the above.

We will be accepting abstracts (4 pages excluding acknowledgment section, references and appendix) that fit the theme of the workshop. Please use the standard NeurIPS template for submitting the abstracts, the submission may optionally contain an appendix, no broader impact section is required. Authors of accepted abstracts are expected to provide a short video (5 min) describing their work and to participate in a short poster session on the day of the workshop.
Important dates:
Submissions open: Sep 03 2020 11:59PM UTC
Submissions close: Oct 05 2020 11:59PM UTC
Notification of acceptance: Oct 29 11:59PM UTC
Deadline for recording: Nov 14 2020 11:59PM UTC
Workshop Dates: Dec 11 or 12 2020 (TBD)

We are using OpenReview as our submission system, the review procedure is going to be double-blind and all review related material will remain private. Accepted abstracts are going to be listed on the conference website and OpenReview unless the authors request otherwise. Contact us: lmcaorganizers@gmail.com
Style Files: https://nips.cc/Conferences/2020/PaperInformation/StyleFiles

Submission: https://openreview.net/group?id=NeurIPS.cc/2020/Workshop/LMCA