Drug perturbation response prediction on 4i 20% (test)
1.3Sinkhorn Divergence (eps=0.1)HyCNN
Evaluation results
| Method | Links | |
|---|---|---|
| HyCNNTreatment=Trametinib, Sample size=2637, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.3 | |
| HyCNNTreatment=Tram-dabra, Sample size=2958, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.3 | |
| HyCNNTreatment=Tram-panob, Sample size=2377, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.31 | |
| HyCNNTreatment=Tram-dabra, Sample size=2958, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.32 | |
| ICNN(sp+q)Treatment=Tram-dabra, Sample size=2958, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.34 | |
| ICNN(sp+q)Treatment=Tram-panob, Sample size=2377, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.34 | |
| HyCNNTreatment=Tram-erlot, Sample size=2452, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.36 | |
| HyCNNTreatment=Trametinib, Sample size=2637, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.38 | |
| HyCNNTreatment=Tram-erlot, Sample size=2452, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.39 | |
| ICNN(sp+q)Treatment=Trametinib, Sample size=2637, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.41 | |
| ICNN(sp+q)Treatment=Tram-erlot, Sample size=2452, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.47 | |
| HyCNNTreatment=Melphalan, Sample size=2352, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.53 | |
| HyCNNTreatment=Lenalidomide, Sample size=2524, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.55 | |
| HyCNNTreatment=Dabrafenib, Sample size=2675, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.56 | |
| HyCNNTreatment=Lenalidomide, Sample size=2524, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.56 | |
| HyCNNTreatment=Mln2480, Sample size=2367, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.56 | |
| HyCNNTreatment=Panobinostat, Sample size=2492, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.56 | |
| HyCNNTreatment=Mln2480, Sample size=2367, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.57 | |
| HyCNNTreatment=Tram-panob, Sample size=2377, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.57 | |
| HyCNNTreatment=Sorafenib, Sample size=2763, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.61 | |
| ICNN(sp+q)Treatment=Dabrafenib, Sample size=2675, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.64 | |
| HyCNNTreatment=Dacarbazine, Sample size=2485, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.64 | |
| ICNN(sp+q)Treatment=Mln2480, Sample size=2367, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.64 | |
| HyCNNTreatment=Cisplatin, Sample size=2693, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.65 | |
| HyCNNTreatment=Dabrafenib, Sample size=2675, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.65 | |
| ICNN(sp+q)Treatment=Lenalidomide, Sample size=2524, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.65 | |
| ICNN(sp+q)Treatment=Melphalan, Sample size=2352, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.65 | |
| HyCNNTreatment=Sorafenib, Sample size=2763, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.65 | |
| HyCNNTreatment=Ulixertinib, Sample size=2858, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.66 | |
| HyCNNTreatment=Dacarbazine, Sample size=2485, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.68 | |
| HyCNNTreatment=Everolimus, Sample size=2592, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.68 | |
| ICNN(sp+q)Treatment=Cisplatin, Sample size=2693, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.69 | |
| ICNN(sp+q)Treatment=Sorafenib, Sample size=2763, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.69 | |
| ICNN(sp+q)Treatment=Panobinostat, Sample size=2492, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.7 | |
| ICNN(sp+q)Treatment=Dacarbazine, Sample size=2485, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.71 | |
| HyCNNTreatment=Melphalan, Sample size=2352, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.71 | |
| ICNN(sp+q)Treatment=Regorafenib, Sample size=2524, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.71 | |
| HyCNNTreatment=Decitabine, Sample size=2648, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.73 | |
| HyCNNTreatment=Palbociclib, Sample size=2322, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.74 | |
| HyCNNTreatment=Cisplatin, Sample size=2693, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.75 | |
| HyCNNTreatment=Midostaurin, Sample size=2425, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.75 | |
| HyCNNTreatment=Regorafenib, Sample size=2524, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.76 | |
| HyCNNTreatment=Regorafenib, Sample size=2524, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.76 | |
| HyCNNTreatment=Erlotinib, Sample size=2406, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.77 | |
| HyCNNTreatment=Ulixertinib, Sample size=2858, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.79 | |
| ICNN(sp+q)Treatment=Decitabine, Sample size=2648, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.8 | |
| HyCNNTreatment=Vem-cobimet, Sample size=2695, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.81 | |
| HyCNNTreatment=Olaparib, Sample size=2540, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.82 | |
| ICNN(sp+q)Treatment=Palbociclib, Sample size=2322, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.83 | |
| HyCNNTreatment=Dexamethasone, Sample size=2796, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.84 | |
| ICNN(sp+q)Treatment=Erlotinib, Sample size=2406, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.84 | |
| HyCNNTreatment=Everolimus, Sample size=2592, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.86 | |
| ICNN(sp+q)Treatment=Olaparib, Sample size=2540, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.86 | |
| ICNN(sp+q)Treatment=Everolimus, Sample size=2592, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.88 | |
| HyCNNTreatment=Palbociclib, Sample size=2322, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.9 | |
| HyCNNTreatment=Erlotinib, Sample size=2406, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.91 | |
| HyCNNTreatment=Decitabine, Sample size=2648, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.92 | |
| HyCNNTreatment=Panobinostat, Sample size=2492, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.92 | |
| ICNN(sp+q)Treatment=Dexamethasone, Sample size=2796, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.93 | |
| HyCNNTreatment=Temozolomide, Sample size=2513, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.93 | |
| HyCNNTreatment=Vem-cobimet, Sample size=2695, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.93 | |
| HyCNNTreatment=Dexamethasone, Sample size=2796, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.94 | |
| ICNN(sp+q)Treatment=Midostaurin, Sample size=2425, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.94 | |
| ICNN(sp+q)Treatment=Ulixertinib, Sample size=2858, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.94 | |
| ICNN(sp+q)Treatment=Vem-cobimet, Sample size=2695, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.94 | |
| HyCNNTreatment=Cisp-olap, Sample size=2245, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.95 | |
| HyCNNTreatment=Temozolomide, Sample size=2513, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.95 | |
| ICNN(sp+q)Treatment=Cisp-olap, Sample size=2245, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 1.97 | |
| HyCNNTreatment=Hydroxyurea, Sample size=2715, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 1.99 | |
| ICNN(sp+q)Treatment=Imatinib, Sample size=2631, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2 | |
| HyCNNTreatment=Imatinib, Sample size=2631, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.01 | |
| HyCNNTreatment=Olaparib, Sample size=2540, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.01 | |
| HyCNNTreatment=Tram-midos, Sample size=2211, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.01 | |
| HyCNNTreatment=Hydroxyurea, Sample size=2715, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.03 | |
| HyCNNTreatment=Cisp-olap, Sample size=2245, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.05 | |
| ICNN(sp+q)Treatment=Tram-erlot, Sample size=2452, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.07 | |
| ICNN(sp+q)Treatment=Tram-midos, Sample size=2211, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.07 | |
| ICNN(sp+q)Treatment=Hydroxyurea, Sample size=2715, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.08 | |
| HyCNNTreatment=Midostaurin, Sample size=2425, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.08 | |
| HyCNNTreatment=Ixa-Lenal-Dexa, Sample size=2522, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.1 | |
| HyCNNTreatment=Ixa-Lenal-Dexa, Sample size=2522, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.16 | |
| HyCNNTreatment=Vindesine, Sample size=2102, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.19 | |
| ICNN(sp+q)Treatment=Ixa-Lenal-Dexa, Sample size=2522, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.2 | |
| ICNN(sp+q)Treatment=Temozolomide, Sample size=2513, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.21 | |
| ICNN(sp+q)Treatment=Tram-dabra, Sample size=2958, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.23 | |
| ICNN(sp+q)Treatment=Ulixertinib, Sample size=2858, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.27 | |
| ICNN(sp+q)Treatment=Vindesine, Sample size=2102, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.27 | |
| HyCNNTreatment=Imatinib, Sample size=2631, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.33 | |
| ICNN(sp+q)Treatment=Lenalidomide, Sample size=2524, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.36 | |
| HyCNNTreatment=Ixazomib, Sample size=2514, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.44 | |
| HyCNNTreatment=Paclitaxel, Sample size=2636, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.45 | |
| ICNN(sp+q)Treatment=Mln2480, Sample size=2367, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.47 | |
| ICNN(sp+q)Treatment=Dacarbazine, Sample size=2485, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.49 | |
| ICNN(sp+q)Treatment=Paclitaxel, Sample size=2636, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.51 | |
| ICNN(sp+q)Treatment=Tram-panob, Sample size=2377, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.52 | |
| ICNN(sp+q)Treatment=Decitabine, Sample size=2648, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.54 | |
| ICNN(sp+q)Treatment=Dabrafenib, Sample size=2675, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.57 | |
| ICNN(sp+q)Treatment=Poma-Carfil-Dexa, Sample size=2545, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.61 | |
| ICNN(sp+q)Treatment=Ixazomib, Sample size=2514, Training budget=T = 50,000, S = 10, Evaluation Protocol=single long-schedule run2026.04 | 2.62 | |
| HyCNNTreatment=Tram-midos, Sample size=2211, Training budget=T = 1,000, S = 5, Evaluation Protocol=cross-seed mean (50 seeds)2026.04 | 2.62 |