LADDER contains 145 tasks across SimplerEnv-Long, RobotArena-∞-Long and RoboLab-Long.
Robot policies falter as manipulation tasks stretch across more subgoals
LADDER isolates task horizon and shows that orchestration helps, but physical execution remains the dominant bottleneck.
Big Tech
Petko Petkov · Giuliano Albanese · Nikolay Nikolov · Luc Van Gool · Jonathan Tremblay · Danda Pani Paudel
INSAIT, Sofia University “St. Kliment Ohridski” · NVIDIA
Research Digest··2 min read
Petkov et al.
Why this paper
From INSAIT, Sofia University “St. Kliment Ohridski” and NVIDIA
In one line
Robot manipulation policies succeed on single subgoals at 85.4% but collapse on multi-subgoal tasks; tool-selection orchestration improves success but physical execution errors remain the bottleneck.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors (2 noted)
- ✓Reports numbers on named benchmarks (3 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§