ResearchBenchmarksReinforcement Learning on Atari-57 30 random no-op start (test)Reinforcement Learning on Atari-57 30 random no-op start (test)Follow1,189Mean Human-Normalized ScoreRAINBOW189.56449.03708.5967.97Jun 14, 2018Evaluation resultsMethodMethodLinksMean Human-Normalized ScoreMedian Human-Normalized ScoreHuman GapRAINBOWSeeds=2Seeds=22018.061,18923014.4IQNSeeds=5Seeds=52018.061,01921814.1QR-DQNSeeds=3Seeds=32018.0686419316.5C51Seeds=1Seeds=12018.0670117815.2PRIOR.Seeds=1Seeds=12018.0643412417.8DQNSeeds=1Seeds=12018.062287933.4