Over the past 15 years, artificial intelligence (AI) and particularly deep machine learning have become essential to web search. This technology analyzes human search and browsing behavior to enhance search engines' understanding of user queries and available content, enabling them to deliver more accurate and relevant results.

Georgia Tech School of Electrical and Computer Engineering (ECE) Professor Larry Heck has been at the forefront of machine learning research since his time as an ECE Ph.D. student in the 1980s. Over his 30-year career, he has been a trailblazer in deep machine learning, pioneering algorithms for speech processing in the 1990s, web search relevance in the 2000s, and AI virtual assistants in the 2010s.

He helped develop several prominent programs, including Microsoft Cortana, Google Assistant, Viv Labs, and Samsung Bixby.

His influential research on deep learning for web search was recently honored with the Test of Time Award by the Conference on Information and Knowledge Management (CIKM). This prestigious award recognizes CIKM papers published over a decade ago that have made a significant and lasting impact on the research community.

Heck received the award for his 2013 paper, “Learning deep structured semantic models for web search using clickthrough data.” It is one of the earliest publications in deep learning for web search and resulted in one of the most successful and impactful innovations in Microsoft’s Bing Search engine.

“This paper represents a fantastic collaboration between product engineers and deep learning research scientists and was an early example of the power of big data (100M+ daily search clicks) and massive computing,” Heck said.

The research included innovations in training a deep learning search relevance engine with a discriminative objective function over clickthrough data. Clickthrough data measures how often users click on a document or advertisement link relative to how many times it was viewed.

Heck’s paper was not only one of the earliest in deep learning for web search, but it was also the first to differentiate between good and bad clicks in training a web search relevance model.

Heck has a joint appointment with the School of Interactive Computing and is the chief scientist of Tech AI, executive director of the Machine Learning Center. He holds the Rhesa S. Farmer Advanced Computing Concepts Chair and is a Georgia Research Alliance Eminent Scholar.

He had over 30 years of industry research experience before returning to his alma mater Georgia Tech in 2021. This included positions at many industry leaders such as Google, Microsoft, Yahoo!, and Samsung. He currently directs the AI Virtual Assistant (AVA) Lab which focuses on creating the next generation of AI virtual assistants with conversational AI.

ECE professor Suman Datta previously won the Test of Time Award in 2022.