Complete result settings for MSNBC dataset (table columns are sortable)
The experiments are described in Rooshenas and Lowd, Discriminative Structure Learning of Arithmetic Circuits, AIStats 16

Tree#	C	Max. EM Iteration	Train CLL	Validation CLL	Test CLL	Learning Time
1	2.000000	10	-1.794320	-1.795978	-1.801392	185.141931
1	0.500000	20	-1.794322	-1.795984	-1.801391	189.619675
1	2.000000	50	-1.794320	-1.795978	-1.801392	183.311162
1	0.100000	10	-1.794340	-1.796021	-1.801392	185.248915
1	0.500000	10	-1.794322	-1.795984	-1.801391	190.126596
1	1.000000	10	-1.794320	-1.795980	-1.801392	192.854996
1	1.000000	50	-1.794320	-1.795980	-1.801392	184.187671
1	0.500000	50	-1.794322	-1.795984	-1.801391	186.933281
1	1.000000	20	-1.794320	-1.795980	-1.801392	191.966608
1	0.100000	20	-1.794340	-1.796021	-1.801392	188.619983
1	0.100000	50	-1.794340	-1.796021	-1.801392	186.479327
1	2.000000	20	-1.794320	-1.795978	-1.801392	230.745426
2	2.000000	10	-1.653483	-1.653344	-1.661206	40317.303440
2	1.000000	10	-1.655957	-1.655920	-1.663351	40432.161831
2	0.100000	10	-1.659824	-1.660353	-1.666708	40474.912756
2	0.500000	10	-1.656552	-1.656758	-1.663639	41750.021153
2	2.000000	50	-1.653314	-1.653209	-1.661037	43981.610140
2	0.500000	50	-1.656454	-1.656823	-1.663492	44467.057879
2	2.000000	20	-1.653314	-1.653209	-1.661037	45017.933796
2	0.500000	20	-1.656454	-1.656823	-1.663492	46207.249278
2	1.000000	50	-1.653672	-1.653603	-1.661343	56491.115267
2	1.000000	20	-1.653672	-1.653603	-1.661343	57698.660627
3	1.000000	10	-1.602735	-1.602102	-1.614666	74367.213623
3	2.000000	10	-1.602315	-1.601649	-1.614329	74563.065948
3	0.500000	10	-1.603331	-1.602706	-1.615218	74832.006619
3	0.100000	10	-1.607347	-1.606700	-1.619053	79530.419902
2	0.100000	20	-1.659336	-1.659921	-1.666273	80178.926763