UtahPOET: Disorder Mention Identification and Context Slot Filling with Cognitive Inspiration

Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015), pages 399–405, Denver, Colorado, June 4-5, 2015. c⃝2015 Association for Computational Linguistics UtahPOET: Disorder mention identification and context slot filling with cognitive inspiration Kristina Doing-Harris Sean Igo Department of Biomedical Informatics Department of Biomedical Informatics University of Utah Health Sciences Center University of Utah Health Sciences Center 421 Wakara Way, Ste 140 421 Wakara Way, Ste 140 Salt Lake City, UT 84108 USA Salt Lake City, UT 84108 USA kristina.doing-harris@utah.edu Sean.igo@utah.edu Jianlin Shi John Hurdle Department of Biomedical Informatics Department of Biomedical Informatics University of Utah Health Sciences Center University of Utah Health Sciences Center 421 Wakara Way, Ste 140 421 Wakara Way, Ste 140 Salt Lake City, UT 84108 USA Salt Lake City, UT 84108 USA Jianlin.shi@utah.edu John.hurdel@utah.edu Abstract We describe the performance of UtahPOET on SemEval 2015 Task 14. UtahPOET is a cognitively inspired system designed to ex- tract semantic content from general clinical texts. We find that our system performs much better on the context slot-filling aspects of Tasks 2A and 2B than the disorder CUI map- ping of Tasks 1 and 2B or the body location CUI mapping of Task 2B. Our problems with CUI mapping suggested several possible sys- tem improvements. An alteration in the corre- spondence between the system architecture and psycholinguistic findings is also indicat- ed. 1 Introduction We note at the outset that our team approaches clinical NLP using a new, cognitively inspired ar- chitecture. We value dataset independence, so our design priorities do not completely overlap those encompassed by the goals of Task 14. We share the SemEval vision of extracting the full semantic content of clinical text. Our short-term goal, how- ever, was to field test an early prototype of our new architecture and Task 14 provided a conven- ient and well-designed use case. 1.1 Cognitive inspirations Only the human brain is currently able to extract full semantic content from text. We propose an intermediate step between artificial neurons (Mer- olla et al., 2014; Sowa, 2010) and statistical ma- chine learning (ML). We use ML and rule-based NLP components with demonstrated success in clinical information extraction arranged in an ar- chitecture inspired by well-documented findings with respect to cortical processing. Briefly, UtahPOET is inspired by findings re- lated to: layered cognitive processes, the distinc- tion between the dorsal and ventral language processing streams, and the phenomenon of itera- tive refinement. The type of layered (i.e., staged or hierarchical) processing we use shares much in common with traditional NLP and biologically inspired cognitive architectures (Chella, Cossenti- no, Gaglio, & Seidita, 2012; Indurkhya & Damerau, 2010; Sowa, 2010). We will discuss our system’s layering in the system description below. Our distinctive model of dorsal-ventral pro- cessing streams comes from psycholinguistic find- ings. The interpretation of unfamiliar or ungrammatical constructions, rule-based pro- cessing, and learning have been linked to dorsal processing streams in the brain. Ventral processing streams handle familiar, expected, regular con- 399 structions as well as heuristic-type processing (Dominey & …

Content truncated to 500 words