notesum.ai

Published at December 3

ESA: Example Sieve Approach for Multi-Positive and Unlabeled Learning

cs.LG

Released Date: December 3, 2024

Authors: Zhongnian Li¹, Meng Wei, Peng Ying, Xinzheng Xu

Aff.: ¹China University of Mining and Technology, Xuzhou, China

Arxiv: http://arxiv.org/pdf/2412.02240v1

Refer to caption

Dataset	(MP), N	UPU	NNPU	MPU	NMPU	CoMPU	ESA
	(1, 2, 3), 0	50.78 $\pm$ 0.05	50.80 $\pm$ 1.10	97.89 $\pm$ 0.50	98.03 $\pm$ 0.55	98.01 $\pm$ 0.32	99.30 $\pm$ 0.12
	(1, 2, 3), 5	49.57 $\pm$ 0.07	48.75 $\pm$ 0.27	96.91 $\pm$ 0.51	97.96 $\pm$ 0.28	97.37 $\pm$ 0.10	99.33 $\pm$ 0.07
MNIST	(1, 2, 3), 9	50.87 $\pm$ 0.17	49.91 $\pm$ 0.39	97.26 $\pm$ 0.25	97.57 $\pm$ 0.34	97.34 $\pm$ 0.25	99.06 $\pm$ 0.29
	(4, 5, 6), 2	51.56 $\pm$ 0.10	49.14 $\pm$ 2.41	97.49 $\pm$ 0.54	98.06 $\pm$ 0.25	97.98 $\pm$ 0.22	99.12 $\pm$ 0.02
	(R, R, R), R	51.28 $\pm$ 0.24	46.21 $\pm$ 1.48	97.25 $\pm$ 0.29	97.59 $\pm$ 0.65	97.78 $\pm$ 0.18	99.27 $\pm$ 0.04
	(1, 2, 3), 0	48.65 $\pm$ 0.31	47.33 $\pm$ 0.69	93.70 $\pm$ 0.20	93.71 $\pm$ 0.35	93.45 $\pm$ 0.21	95.01 $\pm$ 0.08
	(1, 2, 3), 5	49.98 $\pm$ 0.01	49.57 $\pm$ 0.06	96.83 $\pm$ 0.19	96.91 $\pm$ 0.13	96.28 $\pm$ 0.17	97.75 $\pm$ 0.08
Fashion	(1, 2, 3), 9	49.98 $\pm$ 0.01	46.52 $\pm$ 5.24	96.90 $\pm$ 0.32	96.98 $\pm$ 0.03	96.32 $\pm$ 0.22	97.85 $\pm$ 0.19
	(4, 5, 6), 2	43.04 $\pm$ 0.42	34.41 $\pm$ 1.77	81.20 $\pm$ 1.62	79.79 $\pm$ 1.65	80.34 $\pm$ 0.24	87.09 $\pm$ 0.61
	(R, R, R), R	49.88 $\pm$ 0.06	49.75 $\pm$ 0.28	90.34 $\pm$ 0.60	90.32 $\pm$ 1.11	90.12 $\pm$ 0.51	92.95 $\pm$ 0.33
	(1, 2, 3), 0	48.50 $\pm$ 0.15	46.64 $\pm$ 0.86	87.79 $\pm$ 1.44	89.09 $\pm$ 2.20	88.83 $\pm$ 0.30	94.20 $\pm$ 0.22
	(1, 2, 3), 5	45.40 $\pm$ 0.15	43.64 $\pm$ 0.13	84.90 $\pm$ 0.83	85.83 $\pm$ 1.38	85.02 $\pm$ 0.27	90.85 $\pm$ 0.52
	(1, 2, 3), 9	46.63 $\pm$ 0.24	42.45 $\pm$ 1.12	85.01 $\pm$ 1.32	86.23 $\pm$ 1.32	85.35 $\pm$ 0.02	93.09 $\pm$ 0.22
Kuzushi	(4, 5, 6), 2	45.87 $\pm$ 0.13	41.33 $\pm$ 0.40	82.10 $\pm$ 0.63	83.15 $\pm$ 1.63	82.47 $\pm$ 0.05	90.50 $\pm$ 0.70
	(R, R, R), R	44.68 $\pm$ 0.14	38.73 $\pm$ 3.06	83.75 $\pm$ 1.04	84.42 $\pm$ 0.57	84.05 $\pm$ 0.27	91.61 $\pm$ 0.77
	(R, R, R), R	46.24 $\pm$ 0.09	40.46 $\pm$ 2.52	86.40 $\pm$ 0.78	84.02 $\pm$ 3.26	85.37 $\pm$ 0.23	91.77 $\pm$ 0.19
	(R, R, R), R	46.78 $\pm$ 0.47	43.92 $\pm$ 1.10	81.77 $\pm$ 2.50	83.29 $\pm$ 2.14	82.47 $\pm$ 0.70	91.93 $\pm$ 0.80