Current and Recent Research Projects
The Natural Language Group at ISI has a wide range of ongoing projects, including those related to statistical machine translation, tree transducers, parsing and synchronous tree grammars, discriminative word alignment, question answering, summarization, ontologies, information retrieval, and natural language generation.
Principal Research Areas
Statistical Machine Translation
syntax-based translationtree transducers statistical word alignment morphological modeling language modeling parsing |
Ontologies, Text Mining, and Lexical Semantics
semantic representationsemi-automated ontology creation question answering large-scale text annotation semantic clustering |
Machine Learning
multi-dimensional search and optimizationstatistical inference knowledge prediction |
Digital Government
database alignmentpublic commentary analysis automated ontology creation |
Information Retrieval and Extraction
semantic clusteringrelation extraction opinion identification large-scale data processing biomedical information extraction |
Discourse and Summarization
discourse parsingtext summarization discourse theory language generation automated summarization evaluation |
|
USC is home to one of the fastest and most powerful supercomputing clusters in the world, and has been ranked among the top ten most powerful computer system in an academic setting in the nation.
The USC Center for High-Performance Computing and Communications (HPCC) has quickly risen to become a leader in research computing.
The NL group uses this large computing grid to support ongoing research projects.
The Natural Language Group at USC/ISI has a long and unbroken history of informal weekly seminars. Researchers, faculty, and students from ISI, as well as visiting researchers from other institutions, meet to present ideas, report on interesting experiments, and even give practice talks for upcoming conference meetings. The current organizer of the NL Seminar is Jason Riesa. Please contact Jason to schedule a talk.
Our student-led NLP reading group is generally held on Fridays, every two weeks. We read and discuss current and classic papers in natural language processing and related fields.
Since 1995, the Intelligent Systems Division at USC/ISI has run an AI seminar series, in which excellent speakers, often from outside the university, visit, give talks, and meet with other ISI researchers.
|
Downloadable Software
Summarization Evaluation
[Tratz and Hovy, 2008]
Finite-state string transducer tookit
[Knight and Graehl, 1999]
Finite-state tree transducer toolkit
[May and Knight, 2006]
Semantic abduction engine
[Hobbs et al., 1993]
Phrase-based decoder for Statistical Machine Translation
[Koehn et al., 2003]
Word-level decoder for Statistical Machine Translation
[Germann et al., 2001]
Summarization evaluation
[Lin and Hovy, 2004, 2006]
|