Computational Linguistics and Information Processing (CLIP)

The CLIP Laboratory at the University of Maryland is engaged in designing algorithms and methods that allow computers to effectively and efficiently perform human language-related tasks.  CLIP is using computational methods to improve our scientific understanding of the human capacity for language and to explore heterogeneous datasets at scale; for example, how to design an online search engine that searches content across human languages?

CLIP research covers major areas of computational research on language, including deep learning, human-in-the-loop machine learning, multilingual text processing and low-resource languages, machine translation, and summarization. We also study speech retrieval and cross-language information retrieval when performing online searches. Additional challenges include ranking and personalization, computational social science for mental health, data science for finance, data science for social good, computational psycholinguistics, and e-discovery.

With faculty, researchers, and students spanning the Department of Computer Science, the Department of Linguistics, the iSchool, and the Robert H. Smith School of Business, we are a group known not only for high-quality research, but also for intellectual diversity and strong collaborations.

CLIP Laboratory Website

Research Projects

IARPA BETTER: Multilingual Fine-grained Decompositional Analysis

Principal Investigator(s): Jordan Boyd-Graber
Investigator(s): Benjamin Van Durme (Johns Hopkins University), Aaron Steven White (University of Rochester), Eugene Agichtein (University of North Carolina, Chapel Hill)
Developing enhanced methods for personalized, multilingual semantic extraction and retrieval from text, in support of IARPA's goal of providing users with a system that quickly and accurately extracts complex semantic information, targeted for a specific user, from text.