In 18 months, the OpenAI team developed GPT-4o

Ultraman: Without his (Prafulla Dhariwal’s) vision, talent, belief and determination, there would be no GPT-4o.

“GPT-4o would not have been possible without the vision, talent, belief and long-term determination of @prafdhar. It was these efforts (and the work of many others) that enabled I hope it will be the result of a revolution in the way computers are used.”

Two days after OpenAI released the new generation flagship generation model GPT-4o, OpenAI CEO Altman commented on one of the people in charge of the project. road.

After 18 months of working with multiple teams at OpenAI, co-founder Greg Brockman said: “GPT-4o is the result of an entire team effort. In particular, Thanks to @prafdhar, who believed in building a comprehensive model and worked with multiple teams at OpenAI over the past 18 months to achieve this goal.

Seeing this, everyone is probably curious about who prafdhar is. He is Prafulla Dhariwal, the leader of the Omni team, and the co-creator of GPT-3 and DALL-E 2. Dhariwal said, "GPT-4o (o stands for omni) is the first model launched by the Omni team and OpenAI's first native, fully multi-modal model. This release is the result of the joint efforts of the entire organization." At the same time, Dhariwal mentioned that he wanted to give a special thanks to some of the amazing members of the team who made this amazing model possible!

Who are the outstanding team members that Dhariwal calls and who are the contributors to the entire team? Next, we will reveal them one by one.

Prafulla Dhariwal

Prafulla Dhariwal is a research scientist at OpenAI (also the person whom Altman was particularly grateful to in the previous article), The research directions are mainly generative models and unsupervised learning. He joined OpenAI in 2017 and has been there for more than 7 years.

Prafulla Dhariwal received her undergraduate degree from MIT. At MIT, Prafulla Dhariwal studied computer science, mathematics, physics, and served as an undergraduate researcher at the MIT Center for Brains, Minds, and Machines.

According to Google Scholar, Dhariwal’s papers have been cited more than 72,200 times, and the research he participated in includes DALL-E 2, DALL-E 3, Glow, Consistency Models, etc., all of which are familiar to everyone.

James Betker

James Betker is a researcher at OpenAI Engineer, focusing on generative modeling of images and audio.Before joining OpenAI, James Betker worked as a software engineer at GARMIN International Avionics Switzerland GmbH for 12 years and worked as a software engineer at Google.

Rowan Zellers

Rowan Zellers is a researcher at OpenAI, building systems that understand people and the world through language, vision, and sound. machine.

Prafulla Dhariwal said: "Rowan Zellers' research work allows GPT-4o to watch videos as naturally as humans." Rowan Zellers is also a member of the GPT-4 research team.

Alexis Conneau

##Alexis Conneau is now the Audio AGI Research Lead at OpenAI. Dhariwal said Conneau came up with the vision for HER earlier than anyone at OpenAI and followed through with it!

Gabriel Goh

##Gabriel Goh is a machine learning researcher at OpenAI. He is interested in interpretability, machine learning, data visualization and convex Interested in optimization. He has participated in the research of DALL・E 1, DALL・E 3, GPT-4, etc.

Ishaan Gulrajani

Ishaan Gulrajani is now an OpenAI machine learning researcher. He previously graduated with a Ph.D. from Stanford University and has participated in projects such as Alpaca and Stanford alpaca. Research.

Alex Nichol

## Alex Nichol works on deep learning research at OpenAI, co-creator of DALL-E 2, and specializes in coding, artificial intelligence, and 3D printing.

Li JING

Li JING is a researcher at OpenAI. He mainly works on image/3D generation in GPT-4o work. Research.

Li JING is also a member of the research team of DALL・E 3 and Sora. Li JING's research work is dedicated to helping artificial intelligence understand the real world and create new intelligence.

Casey Chu

Casey Chu is a researcher at OpenAI, studying multi-modal artificial intelligence systems, and has participated in GPT- 4. Research on DALL・E 2. Previously, he received his PhD from Stanford University.

Mark Chen

##Mark Chen is the director of cutting-edge research at OpenAI. He was present at the early morning press conference on May 14. Mark Chen has participated in research on DALL・E, GPT-4, etc.Jiahui Yu

##Jiahui Yu leads the perception team at OpenAI. Previously, he received a PhD from the University of Illinois at Urbana-Champaign. Ph.D. Before joining OpenAI, Jiahui Yu participated in research at Google including Gemini, PaLM 2, etc.

Huiwen Chang

Huiwen Chang joined OpenAI in 2023. She previously worked as a senior research scientist at Google and graduated from Tsinghua University with her undergraduate degree. University.

A. Jabri

##A. Jabri is a technical staff member of OpenAI. His main research field is using Scalable goals and architectures for self-supervised and unsupervised learning. Previously interned at Google DeepMind.

Christine McLeavey

##Christine McLeavey is a member of the technical team at OpenAI, mainly engaged in multi-modal research. Her research interests lie mainly in music generation, and she is also a member of the OpenAI music model MuseNet research team.

Yu Zhang

Yu Zhang is a member of the OpenAI technical team and has served as a research scientist at Google DeepMind for 6 years.

Yu Zhang graduated from the Computer Science Department of Shanghai Jiao Tong University with a bachelor's degree and a Ph.D. from the Computer Science Department of MIT.

In addition to the researchers mentioned above, OpenAI has also officially released a complete list:

Full list address: https://openai.com/ gpt-4o-contributions/

Among them, the core contributors in language include:

The core contributors in multimodality include:

Interested readers can view the complete list for more information.

The above is the detailed content of In 18 months, the OpenAI team developed GPT-4o. For more information, please follow other related articles on the PHP Chinese website!