AI for Programming Education

Опубликовано 4 января 2022, 16:44

Education needs technology-based disruption for personalization. Programming education needs most attention, given the increasing skills gap in the IT industry and the need to introduce computational thinking in schools. Programming education is also challenging as there is no unique solution to a problem and today’s programming toolkit is far from the challengers that learners face. First, precise syntax requirement of programming languages (made worse by confusing and unhelpful errors messages from compilers) throws off some students who fall behind or drop off. Then the next challenge for students is to fix semantic/logical bugs in their programs. Providing them a sample solution isn’t sufficient since what they really need to know is, how their solution, which reflects their creative process, can be lifted into a correct solution. There is potential to also help students with performance and design issues in their code. With recent ongoing advances in program analysis and synthesis techniques, there is a huge opportunity to build AI-powered bots for programming education that can provide personalized and interactive feedback/hints to students just like a human instructor. This can enable more students to learn, and learn better and faster. In this talk, I will showcase some of these advances and potential.

Speaker: Sumit Gulwani, Partner Research Manager, Microsoft Research
microsoft.com/en-us/research/p...

PROSE group: microsoft.com/en-us/research/g...