IBM CodeNet teaches AI to translate programming languages


Bobby Hellard

11 May, 2021

IBM unveiled a suite of new artificial intelligence (AI) services on Monday ahead of its virtual IBM Think 2021 conference. 

The new features include software that teaches AI how to translate code, a cloud migration service, and more functions for the firm’s Watson Assistant. 

The most interesting of the announcements was a Rosetta Stone-like service for programming code called ‘CodeNet‘, which is an expansive dataset designed to teach AI and machine learning systems how to translate code. It comprises some 14 million snippets and 500 million lines of code spread across more than 55 legacy and active languages, such as Cobol, Java, C++, and Python. The dataset is constructed in a way that allows for bidirectional translation, enabling Cobol legacy code to be translated into Java, for example. 

“Given its wealth of programmes written in a multitude of languages, we believe Project CodeNet can serve as a benchmark dataset for source-to-source translation and do for AI and code what the ImageNet dataset did years ago for computer vision,” the firm said. 

AI featured heavily in the updates, with new features for Cloud Pak for Data, an SaaS integration platform called ‘Watson Orchestrate’, and a Watson Assistant collaboration with healthcare providers handling increased workloads due to the pandemic. The tech giant also revealed ‘Maximo Mobile’, a mobile platform based on IBM’s Maximo asset management service. 

The updates also included a new cloud migration service called ‘Mono2Micro’ that also uses AI to analyse large enterprise applications and provide recommendations on how to best adapt them for the move to the cloud. IBM said it can simplify and speed up an error-prone process, which can reduce costs and maximise returns on investment.

Beyond artificial intelligence, the announcements included a hybrid cloud-based financial services platform, built with Red Hat OpenShift. There was also a $1 billion investment to support IBM’s partner ecosystem which includes skills training, and a streamlined service for quantum computing called Qiskit Runtime.