IC Seminars : Programming Languages and Tools Research @ Microsoft
Event details
| Date | 18.05.2012 |
| Hour | 10:00 |
| Speaker | Dr. Sriram Rajamani, Microsoft Research India – Hosted by Prof. Viktor Kuncak |
| Location | |
| Category | Conferences - Seminars |
Abstract
We present an overview of Programming Languages and Tools Research at Microsoft Research India. Our research group is interested in
verification, program analysis, empirical/statistical techniques, machine learning and distributed systems. After an overview, we will focus on two topics:
1.Interplay between program analysis and machine learning: Every program analysis/verification tool needs annotations. We should how annotations can be inferred using techniques from machine learning, particularly Bayesian inference. We show how programmer intuitions can be represented as probabilistic constraints and solved to yield likely annotations, which can then be checked using a sound verification tool. Going the other way, we also show how to improve the efficiency of probabilistic inference itself using iterative refinement techniques.
2.Programming models for large scale distributed systems: We have been working on programming abstractions for distributed systems, which hides the complexities of partitioning, replication, and failures, while providing simple declarative abstractions to the programmer. In particular, we present the language CScale, its semantics, its runtime, and sample applications we have been able to build with it.
Biography
Sriram Rajamani is Assistant Managing Director of Microsoft Research India. He got his PhD from UC Berkeley, and worked for 6 years in Microsoft Research Redmond before moving to Microsoft Research India in 2005. His research interests are in program analysis, programming languages and software engineering.
>> More information
We present an overview of Programming Languages and Tools Research at Microsoft Research India. Our research group is interested in
verification, program analysis, empirical/statistical techniques, machine learning and distributed systems. After an overview, we will focus on two topics:
1.Interplay between program analysis and machine learning: Every program analysis/verification tool needs annotations. We should how annotations can be inferred using techniques from machine learning, particularly Bayesian inference. We show how programmer intuitions can be represented as probabilistic constraints and solved to yield likely annotations, which can then be checked using a sound verification tool. Going the other way, we also show how to improve the efficiency of probabilistic inference itself using iterative refinement techniques.
2.Programming models for large scale distributed systems: We have been working on programming abstractions for distributed systems, which hides the complexities of partitioning, replication, and failures, while providing simple declarative abstractions to the programmer. In particular, we present the language CScale, its semantics, its runtime, and sample applications we have been able to build with it.
Biography
Sriram Rajamani is Assistant Managing Director of Microsoft Research India. He got his PhD from UC Berkeley, and worked for 6 years in Microsoft Research Redmond before moving to Microsoft Research India in 2005. His research interests are in program analysis, programming languages and software engineering.
>> More information
Practical information
- Informed public
- Free
Organizer
- Simone Muller
Contact
- Simone Muller