IBM Project CodeNet to teach AI to code


In an effort to make debugging, maintaining, and updating code easier, IBM unveiled Project CodeNet, a open-source dataset To advance AI understanding and translating the code. The project was announced at this week’s Think conference as part of IBM’s AI for Code initiative, which aims to help developers improve their productivity by further automating their engineering process.

“We are entering a new era where it is essential to leverage today’s powerful technologies such as artificial intelligence (AI) and hybrid cloud to create new solutions capable of modernizing processes across the business pipeline. information technology (IT) ”, Richir Puri, Chief Scientist at IBM Research, written in a blog post. “The CodeNet project can specifically drive algorithmic innovation to extract this context with sequence-to-sequence models, just like what we have applied in human languages, to make a more significant breach in machine understanding of code as opposed to processing. code machine. “

The CodeNet project includes 14 million code samples and 500 million lines of code from programming languages ​​such as C ++, Java, Python, Go, COBOL, Pascal, and FORTRAN. The project also includes high-quality metadata and annotations, as well as input and output examples to help researchers program intent when translating from one programming language to another.

“Our team is excited to provide researchers and developers with an easy-to-use and easy-to-understand dataset and technology set, while simultaneously assisting in the development of algorithms that will advance AI for code. With Project CodeNet, we hope to generate sustainable business value as companies embark on their IT modernization journey, ”Puri wrote.

Potential future use cases for the project include code research and cloud detection, automatic code correction, regression studies, and prediction.

The company also announced a new milestone with Cloud Pak for Data; new interactive artificial intelligence capabilities in Watson to increase the personal productivity of professionals; the easy-to-deploy Maximo Mobile mobile platform; Mono2Micro for the optimization and modernization of applications for the hybrid cloud; and Qiskit Runtime improvements.

“We will look back on this year and last as the moment when the world entered the digital century with full force,” said Arvind Krishna, CEO of IBM. “In the same way that we have electrified factories and machines over the past century, we will use hybrid cloud to infuse AI into 21st century software and systems. And one thing is certain: it is a future that must be built on the basis of deep industrial collaboration. No one understands this better than IBM, which is one of the reasons we are increasing investments in our partner ecosystem. Also at Think 2021, we’re unveiling our latest innovations in hybrid cloud and artificial intelligence – the very technologies that serve as cornerstones of a new IT architecture for businesses. “

Leave A Reply

Your email address will not be published.