# Luke Zettlemoyer

Luke Zettlemoyer is a professor at the University of Washington and a senior researcher at Meta AI, specializing in natural language processing and large language models.

Luke Zettlemoyer is a professor in the Paul G. Allen School of Computer Science & Engineering at the University of Washington and a senior research scientist at [Meta AI](https://www.wikiprompt.org/wiki/meta-ai). His research focuses on [natural language processing](https://www.wikiprompt.org/wiki/natural-language-processing) (NLP), particularly on developing methods for semantic parsing, grounded language understanding, and efficient training of [large language models](https://www.wikiprompt.org/wiki/large-language-model). He is known for contributions to weakly supervised learning and for co-developing the [DeBERTa](https://www.wikiprompt.org/wiki/deberta) model architecture, which achieved state-of-the-art results on several NLP benchmarks.

Zettlemoyer received his Ph.D. from the Massachusetts Institute of Technology in 2007, where he worked under the supervision of Michael Collins. Before joining the University of Washington in 2013, he was an assistant professor at the University of Illinois at Urbana-Champaign. He has also held research positions at Microsoft Research and [Amazon AI](https://www.wikiprompt.org/wiki/amazon).

## Research Contributions

Zettlemoyer's early work focused on learning semantic parsers from natural language utterances. He introduced the use of combinatory categorial grammars (CCG) for weakly supervised semantic parsing, allowing models to learn from sentence-level annotations rather than detailed logical forms. This line of research influenced subsequent work in grounded language acquisition and question answering.

In the 2010s, he contributed to the development of neural network-based models for NLP, including work on attention mechanisms and memory-augmented networks. His group at the University of Washington produced influential papers on multi-task learning, domain adaptation, and efficient inference for sequence models.

## Large Language Models and Meta AI

At Meta AI, Zettlemoyer leads a team working on large-scale language models. He was a key contributor to the [OPT](https://www.wikiprompt.org/wiki/opt) and [LLaMA](https://www.wikiprompt.org/wiki/llama) model families, which are open-source large language models released by Meta. These models have been widely adopted by researchers and practitioners due to their competitive performance and open accessibility. Zettlemoyer's work on efficient training methods, such as FlashAttention and gradient checkpointing, has helped reduce the computational cost of training large models.

He has also been involved in efforts to improve the factual accuracy and reasoning capabilities of language models, including work on retrieval-augmented generation and chain-of-thought prompting.

## Teaching and Mentorship

At the University of Washington, Zettlemoyer teaches graduate courses on natural language processing and machine learning. He has advised numerous Ph.D. students who have gone on to positions in academia and industry, including researchers at [OpenAI](https://www.wikiprompt.org/wiki/openai), [Google DeepMind](https://www.wikiprompt.org/wiki/google-deepmind), and [Anthropic](https://www.wikiprompt.org/wiki/anthropic). His mentorship style emphasizes rigorous experimentation and clear communication of research ideas.

## Awards and Recognition

Zettlemoyer has received several awards for his research, including a National Science Foundation CAREER Award and a Sloan Research Fellowship. He is a frequent area chair for major NLP conferences such as ACL and EMNLP, and he has served on the editorial board of the Transactions of the Association for Computational Linguistics. His work has been cited tens of thousands of times, reflecting its broad impact on the field.

## Selected Publications

Among his most cited papers are "DeBERTa: Decoding-enhanced BERT with Disentangled Attention" (2021), which introduced a novel attention mechanism that improved performance on the [SuperGLUE](https://www.wikiprompt.org/wiki/superglue) benchmark, and "OPT: Open Pre-trained Transformer Language Models" (2022), which detailed the training and release of a suite of open-source language models. He has also published influential work on semantic parsing, including "Learning to Map Sentences to Logical Form" (2007) and "Weakly Supervised Learning of Semantic Parsers" (2012).

Zettlemoyer continues to be an active researcher, with recent interests in multimodal models and the alignment of language models with human values. His work bridges academic research and industrial deployment, making him a prominent figure in the [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) community.

---
Source: https://www.wikiprompt.org/wiki/luke-zettlemoyer
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-09-08T15:39:07.995127+00:00