11-891 Neural Code Generation
6–12 units, Carnegie Mellon University
This course is a graduate-level course introducing research topics in neural code generation: modeling and synthesizing programs using deep-learning methods, with an emphasis on neural language models. The course covers (1) foundational advances in neural language models that are relevant to code generation (2) code-specific formal methods for training, inference, and evaluation and (3) research frontiers such as long document lengths, domain adaptation, and downstream interactive and production-ready uses of the final code generation models. The course combines lectures and paper discussions, as well as a final research project for students taking the course for 12 units. This is a graduate course designed primarily for PhD and research master students at LTI, S3D, MLD, CSD, HCII and RI; others, for example (undergraduate) students of CS or from professional master programs, are advised to seek prior permission of the instructors. It is required for students to have taken an introductory ML course or an ML project-based course such as 10-401, 10-601, 10-701, 11-663, 11-441, 11-641, 11-667, 11-741, 11-711, or 11-777. Prior knowledge of deep learning is recommended.
Prerequisites: none listed
Offered: Fall 2025
Courseweave needs JavaScript to show sections, workloads and planning tools.