Cohere announced the launch of a nonprofit research lab called Cohere For Artificial Intelligence. Cohere will work to solve some of the industry's toughest challenges by contributing "fundamental research" to the open source community.
As we broaden how and where research is done, we are excited to be leading a new nonprofit artificial intelligence research lab. According to Hooker, the focus will be on collaborating openly and contributing fundamental research. A key part of our work will be to expand the community and train the next generation of talent, creating new entry points to work on core research.
Not enough funding is being set aside for artificial intelligence research outside of wealthy corporations, according to a study. A study shows that over the course of a decade, 52 tenure-track artificial intelligence professors were removed from academic and nonprofit work by the companies they worked for.
Concentrating power within corporations has a number of obvious drawbacks, but one of the most alarming is that it tends to under emphasize certain values. According to a number of experts, corporate artificial intelligence projects have lead to an "unscientific fixation" on projects only possible for people with access to powerful data centers. It can take years to see the light of day if work is closely guarded.
Hooker said that the agenda was centered on advancement of machine learning questions and community focused research. We want to have a research agenda that will allow us to identify major challenges before they become problems. A core piece of the research is the robust use of models, and we are focused on a variety of different disciplines.
According to Hooker, access to compute resources is one of the core components of Cohere For Artificial Intelligence. Trends in language models show that the role of compute access is changing. The creation of a highly sophisticated language model required a lot of compute resources. The barriers to entry have been lowered thanks to the work of the open source community.
Cohere was founded in February of this year by a team that includes the co-director of UC Berkeley's artificial intelligence lab and other well-known figures. The academic paper "Attention Is All You Need" was written by Gomez. The GPT 3 and Codex are based on the Transformer architecture. The two are contributors at For.ai, an open artificial intelligence research collective.
Hooker said that for.ai was designed to help early career enthusiasts engage with more experienced researchers. Many of the founding members went on to further their education. One of the first research groups to support independent researchers was For.ai. The original concept is being reintroduced with more resources built out from Cohere.
Hooker said that Cohere For Artificial intelligence will offer ways for data scientists to meet and collaborate. Promoting stewardship of open source scientific practices and the "responsible" release of code will be part of this.
Cohere For Artificial intelligence is an ambitious research lab that is contributing to the research community, but also priorities how to better involve a diverse set of voices. We want to change the way research is done.
Despite its lofty goals, Cohere For Artificial Intelligence is likely to invite skepticism from researchers who are wary of Cohere. Cohere has raised $170 million to date from institutional venture capital firms including Tiger Global Management. Early backers of Cohere were the chief scientist of the cloud artificial intelligence division of the company, as well as a fellow of the company, and two other people. Large language models can be trained on the company's dedicated hardware infrastructure.
One week after forming it, the artificial intelligence advisory board was dissolved by the search engine giant. Timnit Gebru was fired by the company in 2020 for sending an email critical of the company. Margaret Mitchell, an ethicist who had publicly denounced the company's handling of the situation, was dismissed by the company.
There have been some mixed bag initiatives to fund artificial intelligence research.
The Allen Institute for Artificial Intelligence, founded by the late Microsoft co-owner Paul Allen, is one of the success stories. Anthropic, launched by former OpenAI executives, has raised over half a billion dollars to research and develop reliable, interpretable, and steerable artificial intelligence systems.
For every Anthropic and Artificial Intelligence, there is an OpenAI, which began as a nonprofit before transitioning to a capped-profit. The White House Office of Science and Technology Policy came under scrutiny after it was reported that EricSchmidt wields heavy influence over the office. One of the first recipients asked that her name not be considered.
Gebru's distributed artificial intelligence research is a new collective that has shown promise. Projects like Hugging Face's BigScience and Eleutherai are examples of what can be done in the field of artificial intelligence.
Hooker said that it was up to them to prove that Cohere Forai wouldn't venture over time. Though Cohere will rely on Cohere for resources and funding, a separate separation has been created between the two to preserve its independence. The separation is necessary so that it can continue to contribute to and serve the larger community. Cohere For Artificial intelligence is a nonprofit that was designed to work with many different organizations. The work will be open source in order to give more people access to it.
All Rights Reserved © 2024
