Advancing Preference Learning in AI: Beyond Pairwise Comparisons
Table 2.
Comparison of Preference Elicitation Methods.
|
| Method |
Information Captured |
User Effort |
Scalability |
|
|
| Pairwise Comparisons |
Binary preferences between two items |
High |
Limited |
|
| Ranking Multiple Items |
Ordinal preferences among several items |
Moderate |
Moderate |
|
| Numerical Ratings |
Degree of preference on a scale |
Low |
High |
|
| Natural Language Feedback |
Detailed, nuanced preferences |
Variable |
High |
|
| Hybrid Approaches |
Combined strengths of multiple methods |
Variable |
High |
|
|
1 Information summarized from various studies on preference learning methods.
|
|
