IC Colloquium: From Programs to Interpretable Deep Models, and Back
Event details
Date | 30.10.2018 |
Hour | 11:00 › 12:15 |
Location | |
Category | Conferences - Seminars |
By: Eran Yahav - Technion Israel Institute of Technology
Video of his talk
Abstract:
In this talk, we demonstrate how deep learning over programs is used to provide (preliminary) augmented programmer intelligence. In the first part of the talk, we show how deep learning over programs is used to tackle tasks like code completion, code summarization, and captioning.
We describe a general path-based representation of source code that can be used across programming languages and learning tasks, and discuss how this representation enables different learning algorithms. In the second part, we describe techniques for extracting interpretable representations from deep models, shedding light on what has actually been learned in various tasks.
Bio:
Eran Yahav is a faculty member at the Computer Science Department, Technion, Israel. His research interests include program synthesis, machine learning and information-retrieval techniques for PL, program analysis, abstract interpretation, verification, programming Languages, and software engineering. He is a recipient of an European Research Council (ERC) grant PRIME (Programming with Millions of Examples). He is also a senior technology advisor at Codota. Prior to joining Technion, he was a research staff member at IBM Research.
More information
Video of his talk
Abstract:
In this talk, we demonstrate how deep learning over programs is used to provide (preliminary) augmented programmer intelligence. In the first part of the talk, we show how deep learning over programs is used to tackle tasks like code completion, code summarization, and captioning.
We describe a general path-based representation of source code that can be used across programming languages and learning tasks, and discuss how this representation enables different learning algorithms. In the second part, we describe techniques for extracting interpretable representations from deep models, shedding light on what has actually been learned in various tasks.
Bio:
Eran Yahav is a faculty member at the Computer Science Department, Technion, Israel. His research interests include program synthesis, machine learning and information-retrieval techniques for PL, program analysis, abstract interpretation, verification, programming Languages, and software engineering. He is a recipient of an European Research Council (ERC) grant PRIME (Programming with Millions of Examples). He is also a senior technology advisor at Codota. Prior to joining Technion, he was a research staff member at IBM Research.
More information
Practical information
- General public
- Free
- This event is internal
Contact
- Host: Viktor Kuncak