Usefulness Prediction on SIGIR 16 benchmark leaderboard · SOTA2 Research

Usefulness Prediction on SIGIR 16

0.53S-rhoTRUERubric+Session

Evaluation Results

Method	Links
TRUERubric+Session2025.04		0.53	0.19	43	43	38	0.88
TRUESession2025.04		0.5	0.13	39	38	32	0.99
TRUERubric+Session2025.04		0.5	0.14	41	40	32	0.98
TRUESession2025.04		0.43	0.12	40	37	32	0.94
GBDT2025.04		0.42	0.24	40	44	44	0.86
TRUESession2025.04		0.42	0.09	37	34	30	1.02
TRUERubric+Session2025.04		0.42	0.1	28	28	23	1.1
TRUERubric+Session2025.04		0.41	0.09	37	34	28	1.11
TRUERubric+Baseline2025.04		0.4	0.05	31	31	20	1.31
CLUE2025.04		0.38	0.18	38	41	38	0.98
TRUEBaseline2025.04		0.38	0.05	32	31	22	1.28
CLUE2025.04		0.37	0.16	34	40	36	0.99
TRUESession2025.04		0.36	0.09	34	34	27	1.16
U_a (Human–Human)2025.04		0.35	0.17	39	38	38	0.92
TRUEBaseline2025.04		0.34	0.1	39	36	29	1.1
TRUERubric+Baseline2025.04		0.32	0.08	36	34	26	1.18
GBDT2025.04		0.31	0.16	40	40	37	1.01
Linear Regression2025.04		0.3	0.02	36	27	17	0.99
TRUEBaseline2025.04		0.29	0.09	32	34	27	1.22
TRUERubric+Baseline2025.04		0.28	0.1	32	34	30	1.03
TRUEBaseline2025.04		0.27	0.05	30	30	23	1.28
TRUERubric+Session2025.04		0.27	0.06	25	25	19	1.32
TRUERubric+Baseline2025.04		0.26	0.11	35	35	30	1.14
TRUESession2025.04		0.23	0.05	24	24	20	1.27
TRUEBaseline2025.04		0.12	0.01	29	26	12	1.69
TRUERubric+Baseline2025.04		0.1	0.01	31	26	9	1.77