notesum.ai
Published at November 15Towards Automatic Evaluation of Task-Oriented Dialogue Flows
cs.CL
cs.AI
Released Date: November 15, 2024
Authors: Mehrnoosh Mirtaheri1, Nikhil Varghese2, Chandra Khatri2, Amol Kelkar2
Aff.: 1USC Information Sciences Institute; 2Got It AI
| STAR | FINANCE | |||||||||||
| Supervised | Unsupervised | Supervised | Unsupervised | |||||||||
| Model | FF1 | FuDGE | Complexity | FF1 | FuDGE | Complexity | FF1 | FuDGE | Complexity | FF1 | FuDGE | Complexity |
| ALG1-Min | 0.59 | 0.08 | 0.57 | 0.73 | 0.26 | 0.28 | 0.71 | 0.03 | 0.44 | 0.03 | 0.03 | 0.99 |
| ALG1-Centroid | 0.58 | 0.12 | 0.71 | 0.34 | 0.71 | .03 | 0.03 | 0.09 | ||||
| ALG2-Min | 0.75 | 0.27 | 0.23 | 0.79 | 0.24 | 0.18 | 0.81 | 0.21 | 0.18 | 0.67 | 0.23 | 0.41 |
| ALG2-Centroid | 0.71 | 0.35 | 0.73 | 0.34 | 0.77 | 0.27 | 0.65 | 0.27 | ||||