Quality Rating on Human-study paradigms H1–H4
66.7Win RateCUPID
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CUPIDSetup=H3: open preference (text), N=152026.05 | 66.7 | 20 | 13.3 | 1 | 0.058 | 0.59 | |
| CUPIDSetup=H4: open preference (image), N=102026.05 | 50 | 40 | 10 | 0.5 | 0.406 | 0.476 | |
| CUPIDSetup=All pooled (H1–H4), N=512026.05 | 43.1 | 23.5 | 33.3 | 0 | 0.496 | 0.118 | |
| CUPIDSetup=Text pooled (H1–H3), N=412026.05 | 41.5 | 19.5 | 39 | 0 | 0.805 | 0.046 | |
| CUPIDSetup=H1: constraints (text), N=142026.05 | 35.7 | 7.1 | 57.1 | -1 | 0.537 | -0.187 | |
| CUPIDSetup=H2: domain expert (text), N=122026.05 | 16.7 | 33.3 | 50 | -0.5 | 0.289 | -0.5 |