Dr. Detective: The Synonyms Game Thesis Design

It is estimated that 20% of medical diagnoses is incorrect [7, 9]. This is caused by limited human cognitive abilities. We are unable to know everything about every human on earth, and therefore cannot reason over this data. Cognitive systems such as IBM Watson can expand these cognitive abilities. These systems are able to reason over data, and interpret more effective than humans can. They are able to process and answer questions in natural language by finding evidence for the correct answer. However, it is difficult for these systems to interpret natural language because they need to be told what is right and what is wrong. Current methods for automating the extraction of this data have shown to be incomplete [14]. Crowdsourcing can solve this by gathering a complete ground truth [12]. In order to capture positive and negative synonym relations in medical texts, we propose a crowdsourcing experiment using the Crowd-Watson framework in order to create efficient templates for these annotation tasks.