Portrait of Shuvendu Lahiri

Shuvendu Lahiri

Senior Principal Researcher

News & features

Research Focus: Week of April 29, 2024
Microsoft Research Blog

In this edition: Can LLMs transform natural language into formal method postconditions; Semantically aligned question + code generation for automated insight generation; Explaining CLIP performance disparities on blind/low vision data; plus recent news.

Awards | TOGA: A Neural Method for Test Oracle Generation

An illustration of resolving a bad merge into a safe merge. Moving from left to right, circles on a continuous line represent code commits in a version control system. A circle labeled “Base” is the most common ancestor of the commits marked A and B, respectively. All three commits pass the project’s quality gates, denoted by green check marks alongside each of these commits. The subsequent merge results in a failure of some quality gate, denoted by a blue circle labeled “Bad merge” with a red x above it. The repair uses machine learning, denoted by an abstract image of a neural network, and program verification and synthesis, denoted by a formal inference rule containing math symbols, to construct a safe merge that passes the quality gate, denoted by a circle outlined in green with a green check mark above it.   
Microsoft Research Blog

| Shuvendu Lahiri

Since the computing world began embracing an open-source approach to programming, building software has become increasingly collaborative. Members of development teams with as few as two developers and as many as thousands are simultaneously editing different components in creating software…

Awards | CAV

The CAV award is given annually at the CAV conference for fundamental contributions to the field of Computer-Aided Verification.  In this nomination, CAV proposed to recognize contributors whose pioneering works laid the groundwork for the theory and practice of SMT. …

In the news | Yahoo Finance

Microsoft has released a new open source tool that will allow for the verification and analysis of smart contracts written in the popular Solidity programming language that is commonly used on the Ethereum blockchain.

In the news | Microsoft Research Blog

With a vision of fortifying smart contracts in Azure Blockchain, the team found a partner in researchers at Microsoft Research working on advanced techniques for ensuring correctness of software programs.

In the news | SIGSOFT

For the paper, 'Feedback-Directed Random Test Generation, ICSE-29, 2007.'