Huang Receives Microsoft Award for Innovative Work in Sequential Decision-Making

Furong Huang, an assistant professor of computer science with an appointment in the University of Maryland Institute for Advanced Computer Studies, has received an award from Microsoft’s Accelerating Foundational Models Research initiative for her novel work in sequential decision-making.

The award recognizes and funds innovative projects that push the boundaries of foundational models and their potential applications. Winners can receive Azure credits, equipping them with powerful tools to further their research endeavors.

Huang, who is also a core member of the University of Maryland Center for Machine Learning, says she is deeply honored to win this award.

“This recognition not only affirms the significance of our work in sequence decision-making, but also serves as an inspiration for my team to push the boundaries of what's possible in machine learning and artificial intelligence,” she says.

Huang’s award-winning project, “Building Foundation Models for Efficient Finetuning or Zero-Shot Learning of Sequential Decision-Making,” delves deep into the intricacies of machine learning, focusing on maximizing the efficiency of foundational models. Once fine-tuned, these models can enhance zero-shot learning—a technique that allows machines to recognize and act on data they haven’t been explicitly trained on.

The prize package will grant Huang exclusive access to Microsoft Azure and OpenAI API, which includes innovative technologies such as GPT-4 and DALL-E 2. She will also be able to utilize open-source models like Llama-2; benefit from Azure Cognitive Services covering areas like speech, vision, decision, and translation; and explore Microsoft’s open-source software libraries.

“Our research aims to revolutionize the way foundation models are used in sequence decision-making across various domains, from health care to autonomous systems,” Huang says. “The award provides us with the resources to further develop models that are more effective but also trustworthy and ethical.”

By emphasizing both online and offline learning stages, her work strives to create adaptable AI systems that can quickly pivot to meet specific needs. Ultimately, Huang believes the research will have a “lasting and positive impact” on society, making AI more reliable and beneficial for all.

—This story was adapted from an article by the Department of Computer Science