Abstract Detail

Teaching

Endara, Lorena [1], Burleigh, Gordon [2], Cui, Hong [3].

Using natural language processing tools to facilitate learning taxonomy, nomenclature, and botany in the classroom.

We developed an undergraduate class in which we used online Natural Language Processing (NLP) software to extract phenomic data from taxonomic descriptions and create phylogenetic character matrices. Our objective was to create a classroom environment in which undergraduates learn about basic concepts of taxonomy, morphology, nomenclature, and phylogenetics while ultimately producing datasets to be used in authentic research projects.
In the first classes, the students learned about plant classification and species concepts across time. Using live plants and herbarium specimens, students described organisms with their own words. Then we transformed the student descriptions into ‘telegraphic syntax’ (the prevalent format of formal taxonomic descriptions). This activity helped the students learn how the NLP software works. In this class, we used the Explorer Taxon Concepts (ETC) online application (http://etc.cs.umb.edu/etcsite/start.html#HomePlace:).
Students used a model system to learn about the functionality of the online NLP software pipeline. They initiated their individual research projects that focused on extracting character information for the conifers and Gnetlaes. Students were assigned different groups of Gymnosperms, and they were responsible for: (A) gathering digital taxonomic descriptions available from online monographic treatments or floras, (B) cleaning and formating the descriptions, (C) uploading them in the NLP software and (D) parsing the text with the NLP software.
During the parsing step, the system requires human input to categorize words that are unknown to the system. Students researched the meaning of the technical terms using glossaries and publications. We designed an activity in which students mapped the structural terms extracted by the system into a sketch of their organisms. Through this activity we sought to convert the student into the “expert” of the group, and also find terms that are interchangeably used by different authors and that need to be synonymized for the parsing analysis.
Matrices derived by parsing analyses and the characters and character states were analyzed and compared in class. Preliminary phenomic-based phylogenetic reconstructions using the datasets generated by students were performed in class to demonstrate the use of this approach.
Our students prepared an abstract and poster to disseminate the results of their research projects at the Botany 2016 meetings, as we considered this an important step to expose students in early stages of their career to modern botanical research.

Related Links:
Explorer of Taxon Concepts
MatrixConverter

1 - University of Florida, Biology, Carr Hall, 217, PO Box 118525, Gainesville, FL, 32611, USA
2 - University Of Florida, P.O. Box 118526, Gainesville, FL, 32611, USA
3 - University of Arizona, School of Information Resources and Library Science, Tucson, AZ, 85719, USA

Keywords:
Natural Language Processing
Nomenclature
Taxonomic descriptions
Botany
Phenotypic traits
Matrices
phylogeny
Undergradute education.

Presentation Type: Oral Paper
Session: 33, Teaching Section Papers Session II
Location: 101/Savannah International Trade and Convention Center
Date: Tuesday, August 2nd, 2016
Time: 2:15 PM
Number: 33002
Abstract ID:153
Candidate for Awards:None