Meet CBI faculty host Graham Neubig

We asked Graham Neubig to tell us a bit about himself and his plans to work with CBI fellow, Xiang Yue.

Introduce yourself and your research interests.                                 

I'm an associate professor at the Language Technologies Institute of Carnegie Mellon University and chief scientist at All Hands AI. My research focuses on language processing, with a particular interest in fundamentals, applications, and understanding of large language models for tasks such as question answering, code generation, and multilingual applications. My final goal is that every person in the world should be able to communicate with each-other, and with computers in their own language. I also believe strongly in open-source and open-science, and try to make almost everything I produce available publicly.

What about the CBI Fellowship program drew your attention?

A couple things. I was looking to hire a post-doc in the areas that CBI supports, and this seemed like an excellent opportunity for the post-doc to be given maximal freedom to pursue directions that they thought would be impactful. Second, I have collaborated with Bosch before, and enjoyed it, so I thought this would be a nice opportunity to renew the collaboration.

What will your CBI fellow be working on?

Xiang Yue is focusing on harnessing the potential of Natural Language Processing (NLP), particularly through the lens of large foundation models. His recent focus is twofold:

1. understanding and enhancing the comprehension capabilities of foundation models
2. ensuring these systems are not just technologically robust but also responsible and reliable in their societal interactions

In particular, he has a strong focus on improving the multimodal abilities and reasoning capabilities of models.

In your view, why is this topic important/pressing/exciting?

This topic is extremely relevant to recent developments in AI, as most large language models are now moving in the direction of multimodality and improved reasoning capabilities. Because of this, rigorous benchmarking and model training methods are necessary.

What about your postdoc fellow makes them the right person to tackle this topic?

Yue obtained his Ph.D. from Ohio State University, where he performed much of the seminal work on these topics.

Yue has a very broad view and carefully keeps his finger on the pulse of recent developments. At the same time, he has strong critical thinking skills and doesn't just take results at face value. Finally, he has strong organizational and mentoring skills, allowing him to effectively lead other students on their projects. The combination of these abilities makes him uniquely positioned to make a big dent in an area that is fast-moving and chaotic by identifying the most important problems to be working on and executing on them in a well-organized and well-motivated way. 

Check back soon to learn more about Xiang Yue and his hopes for his CBI fellowship experience.