Parisa Kordjamshidi

Dr. Kordjamshidi’s research focuses on the extraction of semantics from natural language text, question answering, information extraction, and the interaction of learning and reasoning for language understanding. The goal is to develop machine-learning techniques that can exploit the structure of the language and the domain knowledge to help various natural language semantic extraction tasks in different situations and in different modalities, including the extraction of information from various resources on the web such as social media. Social media is very noisy as it includes errors of various types (misspelling, fragments, odd syntactic constructions) which must be dealt with for information extraction. The REU students will be involved in data preparation, semantic annotation, human evaluation of natural language understanding models, development of demo for our research group NLP systems and possibly the development of parts of our language processing pipelines. Parisa has several years of experience with mentoring undergrad students. During three years of her tenure-track faculty position at Tulane University (2016-2019), she mentored an Honors Undergraduate thesis,14 undergraduate multidisciplinary capstone interdisciplinary projects and also worked with undergraduate summer interns at Tulane and published one text analysis conference (TAC-2018) paper. Three of her undergraduate mentees won the best poster award for their capstone project in Tulane Engineering Capstone Design Expo, Spring of 2019.