Papers with open-world models
Benchmarking the Covariate Shift Robustness of Open-world Intent Classification Approaches (2022.aacl-short)
Copied to clipboard
| Challenge: | Existing benchmarks for open-world intent classification focus on the second aspect of the problem. |
| Approach: | They propose two datasets that include utterances useful for evaluating the robustness of open-world models to covariate shift. |
| Outcome: | The proposed datasets provide a more realistic evaluation scenario for open-world intent classification in task-oriented dialog systems. |