ResearchBenchmarksImitation Learning from Observation on InvertedPendulum v4Follow5.7AERRandom-34.072234.389502.85771.311Jan 24, 2026Evaluation ResultsMethodMethodLinksAERPerformance ScoreRandomseeds=10,000seeds=10,0002026.015.7—BCOseeds=10,000seeds=10,0002026.01521.8260.5191GAIfOseeds=10,000seeds=10,0002026.01870.32480.8696OPOLOseeds=10,000seeds=10,0002026.01990.62540.9906Teacherseeds=10,000seeds=10,0002026.011,000—UfOseeds=10,000seeds=10,0002026.011,0001CILOseeds=10,000seeds=10,0002026.011,0001MAHALOseeds=10,000seeds=10,0002026.011,0001