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

Split Penalty	Standard Deviation	L1 Penalty	Train LL	Validation LL	Test LL	Node#	Edge#	Feature#	Learning Time
10	0.1	1	-9.773708	-9.714603	-9.881591	193871	441407	1656	2045.529031
10	0.1	2	-9.781623	-9.717445	-9.882083	196670	442431	1618	1696.981019
2	0.5	0.1							439116
10	0.5	0.1	-9.743912	-9.703063	-9.865517	276640	633251	1776	2252.503566
10	0.5	2	-9.752609	-9.704323	-9.869543	289599	664095	1700	2217.806841
10	0.1	0.1	-9.800006	-9.734374	-9.900226	217019	491826	1648	2336.374816
5	0.1	1	-9.709895	-9.669034	-9.828498	300430	679091	1984	2713.694455
10	0.5	1	-9.748313	-9.689279	-9.865991	299648	686276	1730	2435.096808
5	0.1	2	-9.729186	-9.679207	-9.836669	332491	739786	1954	2884.281522
5	0.1	0.1	-9.721211	-9.674156	-9.833971	370414	839127	2096	3349.273832
5	0.5	1	-9.698630	-9.659768	-9.826437	371920	845666	2144	3870.107653
5	0.5	2	-9.692375	-9.648737	-9.807169	370328	831736	2052	4386.055217
5	0.5	0.1	-9.672891	-9.647532	-9.809205	435123	974151	2232	5027.436713
2	0.1	2	-9.667572	-9.630129	-9.795032	390584	885588	2592	6080.671596
2	0.1	1	-9.633067	-9.619278	-9.778397	396310	895987	2836	6592.429797
2	0.1	0.1	-9.627384	-9.618267	-9.782756	441762	1008982	3308	8715.629022
2	0.5	2	-9.627057	-9.617145	-9.772134	550566	1249474	2762	10027.006662
2	0.5	1	-9.577228	-9.586455	-9.745818	570558	1287389	3390	11167.571270