Results

In Tables 4, 5 and 6 the execution times of WARMR on Mutagenesis are given, with maximal search depth varying from 3 for the large language to 9 levels for the small language. Here, 'total' is the total execution time and 'exec' is the time needed to test the queries against the examples. In Table 7 the execution times of WARMR on Bongard are given.

Table 4: Results for WARMR on the Mutagenesis dataset using a small language.

Level	No packs		With packs ILPROLOG		speedup ratio
	total	exec	total	exec	net	exec
1	0.35	0.23	0.18	0.15	1.94	1.53
2	6.27	5.60	4.56	4.12	1.38	1.36
3	36.93	31.49	14.01	9.87	2.64	3.19
4	117.33	84.45	45.14	16.27	2.60	5.19
5	215.95	104.36	129.37	20.78	1.67	5.02
6	336.35	111.28	249.41	22.39	1.35	4.97
7	569.14	115.80	497.86	24.63	1.14	4.70
8	902.72	120.99	831.30	25.98	1.09	4.66
9	1268.16	119.60	1148.23	32.28	1.10	3.71

Table 5: Results for WARMR on the Mutagenesis dataset using a medium language.

Level	No packs		With packs ILPROLOG		speedup ratio
	total	exec	total	exec	net	exec
1	2.58	2.27	2.16	2.09	1.19	1.09
2	112.98	42.32	34.35	13.39	3.29	3.16
3	735.19	128.67	262.83	34.70	2.80	3.71
4	4162.15	287.72	1476.06	54.10	2.82	5.32
5	17476.98	444.44	6870.16	73.11	2.54	6.08
6	65138.72	866.85	25921.73	104.81	2.51	8.27

Table 6: Results for WARMR on the Mutagenesis dataset using a large language.

Level	No packs		With packs ILPROLOG		speedup ratio
	total	exec	total	exec	net	exec
1	2.82	2.42	2.28	2.11	1.24	1.15
2	408.85	102.38	102.29	50.67	4.00	2.02
3	27054.33	1417.76	3380.19	370.44	8.00	3.83

Table 7: WARMR results on Bongard.

Level	No packs		With packs ILPROLOG		speedup ratio
	total	exec	total	exec	net	exec
1	0.24	0.22	0.24	0.23	1.00	0.96
2	0.83	0.75	0.77	0.68	1.08	1.10
3	3.28	2.82	2.34	1.92	1.40	1.47
4	11.56	9.31	6.08	4.28	1.90	2.18
5	38.34	28.11	16.20	8.15	2.37	3.45
6	75.51	46.97	36.57	12.22	2.06	3.84
7	135.64	71.60	68.96	15.59	1.97	4.59
8	186.23	84.93	102.46	17.82	1.82	4.77
9	210.82	88.97	120.76	18.52	1.75	4.80
10	216.61	89.38	125.84	18.88	1.72	4.73