Tech

Face Hugging and Services Now launching BigCode, an open source AI system project • TechCrunch


Code generation systems like DeepMind’s AlphaCode, Amazon’s CodeWhisperer, and OpenAI’s Codex, which offer GitHub’s Copilot service, provide an interesting glimpse into what’s possible with AI today in programming. computer. But so far, only one a fistful of hands such AI systems have been freely available to the public and are open-sourced – reflecting the commercial preferences of the companies that build them.

In an effort to change that, face-hugging AI startup and ServiceNow Research, the R&D arm of ServiceNow, today launched BigCode, a new project that aims to develop “modern” AI systems for code in an “open and responsible” way. The ultimate goal is to release a dataset large enough to train the code generation system, which will then be used to prototype – a 15 billion parameter model, larger in size than the Codex (12 billion arguments). number) but smaller than AlphaCode (~41.4 billion parameters) – using ServiceNow’s internal graphics card assembly. In machine learning, parameters are parts of an AI system that are learned from historical training data and essentially determine the skill of the system on a problem, such as code generation.

Inspired by Hugging Face’s BigScience The organizers say that in an effort to create highly complex open source text generation systems, BigCode will be open to anyone with a professional AI research background and able to commit time to the project, the said the organizer. Form went live this afternoon.

“Generally, we expect candidates to be affiliated with a research institution (either in academia or in industry) and working on the technical/ethical/legal aspects of [large language models] for cryptographic applications,” ServiceNow wrote in a blog post. “Once [code-generating system] trained, we will assess its capabilities… We will try to make the assessment easier and broader so that we can learn more about it. [system’s] ability. “

In the collaborative development of a code generation system, which will be open source under a license allowing developers to reuse it under certain terms and conditions, BigCode is looking to address some of the issues. Controversy arose around the workings of AI – powered code generation – especially with regard to fair use. Non-profit organization Protecting Software Freedom among others be censured GitHub and OpenAI to use the public source code, not all under an easy license, to train and monetize the Codex. The Codex is available through OpenAI’s paid API, while GitHub recently started charging for access to Copilot. For their part, GitHub and OpenAI continue to assert that Codex and Copilot do not violate any license terms.

The organizers of BigCode say they will work to ensure that only files from licensed repositories are allowed into the said training dataset. On their way, they say, they will work to establish “responsible” AI practices for training and sharing code generation systems of all types, soliciting feedback from stakeholders before policy statements.

ServiceNow and Hugging Face gave no timeline for when the project could be completed. But they hope it will explore several forms of code generation over the next few months, including systems that autocomplete and synthesize code from snippets and natural language descriptions, and work across a wide range of domains. , tasks and programming languages.



Source link

news7h

News7h: Update the world's latest breaking news online of the day, breaking news, politics, society today, international mainstream news .Updated news 24/7: Entertainment, Sports...at the World everyday world. Hot news, images, video clips that are updated quickly and reliably

Related Articles

Back to top button