RnC_exp_res

	A	B	C	D	E	F	G	H
1								in exp 4-6, RnC(L1) refer to train encoder using RnC and classifier using L1
2		CV						in exp 1-3, just train without using RnC (baselines)
3			Results on AgeDB (batch_size = 128)
4		Experiments	Test MAE	Reported in paper (bs=256)	Error in replication(%)	MAE reduction
5	1	L1	6.544	6.63	-1.297
6	2	L2	6.852	6.57	4.292
7	3	Huber	6.448	6.54	-1.407
8	4	RnC(L1)	6.154	6.14	0.228	-5.959657702	-7.390648567
9	5	RnC(L2)	6.439	6.19	4.023	-6.027437245	-5.783866058
10	6	RnC(Huber)	6.182	6.15	0.520	-4.125310174	-5.963302752
11	7	RnC(feature_sim = 'l1')			#DIV/0!
12	8	RnC(feature_sim = 'l2')			#DIV/0!
13	9	RnC(feature_sim = 'cos')			#DIV/0!
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100