Select your localized edition:

Close ×

More Ways to Connect

Discover one of our 28 local entrepreneurial communities »

Be the first to know as we launch in new countries and markets around the globe.

Interested in bringing MIT Technology Review to your local market?

MIT Technology ReviewMIT Technology Review - logo

 

Akash Lal, 29

Improving software quality using automated verification

Microsoft Research

THE POIROT Akash Lal has made significant advances on the problem of analyzing and verifying proper¬ties of concurrent programs. His approach is about 30 times faster than existing tools.

With software becoming all pervasive in our daily lives, it must become all the more reliable and shouldn't crash or behave unexpectedly. And the way to address this is through automated veri­fication of programs. Though automated verification can significantly improve the quality of software with little manual effort, the available verification tools have been used to test sequential programs. This is a concern as modern software is invariably concurrent.

The problem of automatically veri­fying concurrent programs is two-fold: one has to explore different inputs to the program, and also explore differ­ent interleavings that happen between threads in the program for each input. Akash Lal's work shows that verifica­tion of real-world concurrent programs is possible. His innovation bridges the worlds of sequential and concurrent pro­grams. He has shown how a concurrent program can be efficiently transformed to a sequential one under certain con­ditions. Lal, along with his colleagues at Microsoft Research Lab in Banga­lore, has built a tool called Poirot that is capable of finding bugs in real-world concurrent programs that are missed by conventional testing.

"Akash's research has created an opportunity for developing precise and scalable tools for analyzing concurrent software, and many researchers includ­ing me are building on top of his work," says Shaz Qadeer, Lal's colleague at Microsoft, Redmond. Using a collec­tion of models of concurrent programs, Lal's approach is 30 times faster than any existing tool for verifying concur­rent programs.

2011 TR35 Winners

Pieter Abbeel (video)

Robots that learn from people

June Andronick (video)

Software that can’t crash

Jernej Barbič (video)

Speeding up simulations of complex objects

Dan Berkenstock (video)

Cheaper satellite pictures

Brian Gerkey (video)

A common language for robots

Jeff Hammerbacher (video)

Managing huge data sets

Gert Lanckriet (video)

Teaching computers to classify music

Alina Oprea (video)

Guaranteeing cloud security

Aishwarya Ratan (video)

Converting paper records to digital in real time

Noah Snavely (video)

Synthesizing 3-D models from 2-D photographs

Piya Sorcar (video)

Software that can be localized to teach taboo topics

Kun Zhou

Creating movie-quality graphics in real time

Advertisement

More Innovators Under 35: