15 / 2020-06-17 16:07:02

(Re)CALLing the Human Phenotype Ontology: What Advances in Biomedical Informatics Tell Us About ICALL

Second language learning,Corpora,Computer Assisted Language Learning,Syntactic Analysis

应用语言学前沿问题研究 > 语言智慧教学

Biomedical informatics is implementing NLP, developing universal corpora, and moving toward interoperability and scalability faster than any other field. For example, systems based on the Unified Medical Language System (UMLS) and the Human Phenotype Ontology (HPO) influence millions of critical medical decisions each year even as foreign language instructors and curriculum designers debate the efficacy of various iterations of CALL and corpora-based activities. What are the technologies being deployed in Bioinformatics and what can the paths and pitfalls shaping their trajectory tell us about problems and possibilities for ICALL, and even classroom practice? From revealing the perplexing details of the HPO, to the current reach of the UMLS and what researchers are doing with it, to the convolutional networks, random forests, and bi-directional recurrent modules they deploy to instantiate queries and extract answers, what is happening in biomedical informatics today pushes the very limits of our understanding of what we can do with language, and tells us a great deal about what we may be able to do with ICALL in the future, while also revealing some cautionary tales. This paper delves into the latest findings of Biomedical Informatics to recover lessons for language learning and teaching from a practitioner's perspective, and, informed by Halliday's semiotics, draws three key paradigms of practice from the investigation, as well as developing a call for collaborative attention to address the problems that surfaced in this investigation,