Combining CNNs and Pattern Matching for Question Interpretation in a Virtual Patient Dialogue System
| Autor: | Douglas R. Danforth, Lifeng Jin, Evan Jaffe, Laura Zimmerman, Michael White |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Training set
Interpretation (logic) 020205 medical informatics business.industry Computer science 02 engineering and technology Machine learning computer.software_genre Convolutional neural network Identification (information) Virtual patient 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Pattern matching Artificial intelligence business Baseline (configuration management) computer Word (computer architecture) |
| Zdroj: | BEA@EMNLP |
| DOI: | 10.18653/v1/w17-5002 |
| Popis: | For medical students, virtual patient dialogue systems can provide useful training opportunities without the cost of employing actors to portray standardized patients. This work utilizes word- and character-based convolutional neural networks (CNNs) for question identification in a virtual patient dialogue system, outperforming a strong word- and character-based logistic regression baseline. While the CNNs perform well given sufficient training data, the best system performance is ultimately achieved by combining CNNs with a hand-crafted pattern matching system that is robust to label sparsity, providing a 10% boost in system accuracy and an error reduction of 47% as compared to the pattern-matching system alone. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|