The Regretful Agent (no beam search; greedy action selection)
Reported on 5 benchmarks across 1 task · 1 paper · 1 SOTA
Note: results are matched by exact model name. Different papers may use the same name for different model variants.
Robots5 results
- spl· 2019-03-05SOTA0.4best: 0.76 (human)
- error· 2019-03-055.69best: 1.61 (human)
- length· 2019-03-0513.69best: 1257.38 (Speaker-Follower)
- oracle success· 2019-03-050.56best: 1 (FOAM-Beam Search)
- success· 2019-03-050.48best: 0.86 (human)