Complete result settings for Book 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 Initial Evidence Train CLL Validation CLL Test CLL Node# Edge# Feature# Learning Time
2 0.5 0.5 10 -15.958571 -16.949854 -17.031504 18301 24441 6644 62787.019921
5 0.5 0.5 10 -15.973572 -16.930987 -17.017814 18076 24126 6554 63282.855542
2 0.5 2 10 -15.966958 -16.904997 -16.967332 18311 24455 6648 63713.677048
2 0.5 1 10 -15.963710 -16.927939 -17.008467 18286 24420 6638 62791.745203
2 0.5 2 20 -15.733747 -16.764247 -16.894331 30501 41564 11522 67028.442127
2 0.5 1 20 -15.688642 -16.829720 -16.969579 30541 41577 11540 68445.116759
10 0.5 0.5 10 -15.991116 -16.931588 -17.022231 17966 23972 6510 68616.289737
2 0.5 0.5 20 -15.671721 -16.879221 -17.021503 30551 41591 11544 67993.835364
5 0.5 0.1 10 -15.962489 -16.920255 -17.010723 18151 24231 6584 62977.861908
2 0.5 0.1 10 -15.945347 -16.952646 -17.038222 18431 24623 6696 62940.959518
10 0.5 1 10 -16.005141 -16.921693 -17.004543 17946 23944 6502 65220.383994
5 0.5 1 10 -15.989450 -16.916421 -17.000274 18026 24056 6534 63457.613975
2 0.5 0.1 20 -15.677697 -16.864122 -17.008626 30698 41887 11590 68622.018867
2 0.5 1 40 -15.385454 -16.828867 -17.022359 55336 76290 21458 82765.127787
2 0.5 0.5 40 -15.322831 -16.938727 -17.168627 55336 76290 21458 82287.146451
5 0.5 0.1 20 -15.708652 -16.842778 -16.987742 30446 41444 11502 76085.743208