The PortNLP Lab at Portland State University investigates how computers can understand, process, and generate human language. Led by Associate Professor Ameeta Agrawal in the Department of Computer Science, the lab advances research in natural language processing (NLP), a field of artificial intelligence focused on the interaction between computers and human language.

What We Research

Language is one of the most important ways people communicate and share knowledge. However, human language is also complex. Meaning depends on context, wording, relationships between ideas, and differences across languages and cultures. The PortNLP Lab develops computational approaches that help language technologies interpret this complexity while remaining effective, efficient, fair, and human-centered.

Natural Language Understanding

Natural language understanding focuses on helping computers interpret the meaning and structure of written language. Unlike systems that simply identify individual words, modern language technologies attempt to recognize relationships, context, and the information expressed across sentences and documents. The PortNLP Lab investigates methods for analyzing complex text and improving how language models represent and process information.

This research contributes to technologies that can organize knowledge, assist with information retrieval, and support interactions between people and computing systems. See recent lab papers

Text Summarization

A major research area of the lab is automatic text summarization. Large collections of written information can be difficult to review efficiently. Research articles, reports, and other documents may contain extensive details that readers need to understand within limited time. Text summarization methods aim to produce shorter representations that preserve the most important information.

Developing reliable summarization systems requires more than selecting a few sentences. Researchers must investigate how to identify central ideas, preserve context, and generate summaries that accurately represent the original material. The lab studies computational approaches to these challenges across different types of text, including work on fair summarization that balances quality and diversity. Read “Fair Summarization: Bridging Quality and Diversity in Extractive Summaries”

Multilingual Language Technologies

People communicate in thousands of languages, but language technologies do not always provide the same level of performance across them. The PortNLP Lab investigates multilingual language processing and summarization, with the goal of supporting technologies that can work across linguistic boundaries. Multilingual systems face challenges involving differences in grammar, vocabulary, available training data, and the ways information is expressed.

Research in this area can contribute to broader access to language technologies and help address limitations affecting communities whose languages have received less computational attention. The lab’s work includes evaluating how well multilingual language models handle long documents across languages. Read “Evaluating Multilingual Long-Context Models for Retrieval and Reasoning”

Effective and Human-Centered AI

The PortNLP Lab emphasizes developing language technologies that are useful and trustworthy. A language system must do more than generate fluent text. It should also communicate information accurately, respond appropriately to context, and be designed with the needs of its users in mind. The lab’s research considers effectiveness, computational efficiency, fairness, and human-centered design. These priorities are important as language technologies become increasingly integrated into information systems and everyday applications.

Interdisciplinary Research

Natural language processing brings together machine learning, computer science, linguistics, and the study of human communication. The PortNLP Lab’s research addresses both technical questions about algorithms and broader questions about how language technologies interact with people. The group is also affiliated with the Coco Lab, a related research group exploring connections among computing, management, and anthropology. These connections reflect the importance of studying artificial intelligence within the human contexts where it is used.

Faculty Research: Ameeta Agrawal

Associate Professor Ameeta Agrawal leads PortNLP and its research in natural language understanding, summarization, and multilingual language technologies. She earned her MSc and PhD in Computer Science from York University and was named a 2025 to 2027 Wedge Vision Professor in the Maseeh College of Engineering and Computer Science. Her multilingual language model research is part of NSF’s NAIRR Pilot. View her PSU profile

Since its launch in 2020, the lab has grown from a small research group focused on text summarization and multilingual models to a team investigating a wider range of language processing challenges. The laboratory brings together students and researchers interested in developing computational methods that improve how people interact with written information.

Research and Learning

PortNLP provides a research environment focused on the development and evaluation of language technologies. Its work connects the technical foundations of natural language processing with questions about accessibility, reliability, and responsible artificial intelligence. Undergraduate, master’s, and PhD students contribute to the lab, and the lab has mentored students through PSU’s summer research program.