Complete result settings for Pumsb-star 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	2	-23.142205	-23.346364	-23.231494	853818	1959912	2490	15145.553525
10	0.5	2	-22.753335	-22.981794	-22.890634	848835	1961240	3076	21385.698878
10	0.5	1	-23.053142	-23.289884	-23.187517	789841	1816053	2582	22311.904074
10	0.5	0.1	-22.950342	-23.216972	-23.106033	814724	1877142	3262	25122.032869
10	0.1	1	-23.425112	-23.684005	-23.507116	844474	1947954	3238	38077.002415
10	0.1	0.1	-23.419272	-23.662684	-23.515307	860946	1982435	3400	40067.005888
5	0.1	0.1	-23.321106	-23.527645	-23.446619	829937	1908563	4360	49645.983662
5	0.1	2	-22.892420	-23.113942	-23.025891	819506	1882215	3402	43379.745276
5	0.5	1	-22.682819	-22.915504	-22.814058	842729	1951847	3802	48945.622134
5	0.5	0.1	-22.908061	-23.197415	-23.079176	843489	1956163	4464	47831.675480
5	0.5	2	-22.996338	-23.309834	-23.137740	786362	1817803	3894	54928.023670
5	0.1	1	-23.125917	-23.414013	-23.250661	848090	1956171	4290	51750.594712
2	0.1	2	-22.894247	-23.178860	-23.021153	857473	1968492	4446	50674.153357
2	0.5	1	-22.657612	-22.988254	-22.811411	858786	1982444	4572	74435.082147
2	0.5	0.1	-23.009054	-23.340071	-23.182582	432023	973350	6204	75921.589163
2	0.5	2	-22.766691	-23.040309	-22.876999	858528	1962257	5324	75894.821232
2	0.1	0.1	-23.308965	-23.515601	-23.437483	397732	890982	5932	75887.439354
2	0.1	1	-23.148656	-23.468281	-23.273852	763066	1727298	5768	76897.386820