F2A-ObjectNav: Adaptive Agentic Architecture for Object-Goal Navigation
Overview
F2A-ObjectNav introduces an adaptive agentic architecture for object-goal navigation, using a rule-based Supervisor to coordinate exploration, object detection, and target validation. Evaluated in AI2-THOR, RoboTHOR, and ProcTHOR, it achieved success rates of 0.84, 0.40, and 0.44, respectively, with behavioral metrics analyzing adaptive interventions.
Generated from attributed reports · Updated 2 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Reported quantity · success rate: 0.84 other · Basis not reported · Differing source assertions
“, and ProcTHOR, achieving success rate (SR) / success-weighted path length (SPL) values of 0.84/0.52, 0.40/0.22, and 0.44/0.37, respectively. In addition to task-level metrics, architecture-specific behavioral metrics are used to characterize semantic-guidance usage, strategy switching, recovery activation, memory-supported behavior, and decision stability. The results show different patterns of adaptive intervention”
Exact source · revision 1Source owner not reported
Reported quantity · success-weighted path length: 0.52 other · Basis not reported · Differing source assertions
“, and ProcTHOR, achieving success rate (SR) / success-weighted path length (SPL) values of 0.84/0.52, 0.40/0.22, and 0.44/0.37, respectively. In addition to task-level metrics, architecture-specific behavioral metrics are used to characterize semantic-guidance usage, strategy switching, recovery activation, memory-supported behavior, and decision stability. The results show different patterns of adaptive intervention”
Exact source · revision 1Source owner not reported
Reported quantity · success rate: 0.4 other · Basis not reported · Differing source assertions
“, and ProcTHOR, achieving success rate (SR) / success-weighted path length (SPL) values of 0.84/0.52, 0.40/0.22, and 0.44/0.37, respectively. In addition to task-level metrics, architecture-specific behavioral metrics are used to characterize semantic-guidance usage, strategy switching, recovery activation, memory-supported behavior, and decision stability. The results show different patterns of adaptive intervention”
Exact source · revision 1Source owner not reported
Reported quantity · success-weighted path length: 0.22 other · Basis not reported · Differing source assertions
“, and ProcTHOR, achieving success rate (SR) / success-weighted path length (SPL) values of 0.84/0.52, 0.40/0.22, and 0.44/0.37, respectively. In addition to task-level metrics, architecture-specific behavioral metrics are used to characterize semantic-guidance usage, strategy switching, recovery activation, memory-supported behavior, and decision stability. The results show different patterns of adaptive intervention”
Exact source · revision 1Source owner not reported
Reported quantity · success rate: 0.44 other · Basis not reported · Differing source assertions
“, and ProcTHOR, achieving success rate (SR) / success-weighted path length (SPL) values of 0.84/0.52, 0.40/0.22, and 0.44/0.37, respectively. In addition to task-level metrics, architecture-specific behavioral metrics are used to characterize semantic-guidance usage, strategy switching, recovery activation, memory-supported behavior, and decision stability. The results show different patterns of adaptive intervention”
Exact source · revision 1Source owner not reported
Reported quantity · success-weighted path length: 0.37 other · Basis not reported · Differing source assertions
“, and ProcTHOR, achieving success rate (SR) / success-weighted path length (SPL) values of 0.84/0.52, 0.40/0.22, and 0.44/0.37, respectively. In addition to task-level metrics, architecture-specific behavioral metrics are used to characterize semantic-guidance usage, strategy switching, recovery activation, memory-supported behavior, and decision stability. The results show different patterns of adaptive intervention”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsF2A-ObjectNav: An Adaptive Agentic Architecture for Object-Goal Navigation
This research presents F2, an adaptive agentic architecture for object-goal navigation, which uses a rule-based Supervisor to coordinate specialized components for exploration, object detection, and target validation. The system is evaluated in AI2-THOR, RoboTHOR, and ProcTHOR, achieving success rates of 0.84, 0.40, and 0.44, respectively, with additional behavioral metrics analyzing adaptive interventions.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).