AsgardBench: A benchmark for visually grounded interactive planning

Microsoft Research introduces AsgardBench, a new benchmark for evaluating embodied AI systems that can perform visually grounded interactive planning. The benchmark focuses on testing robots' ability to observe environments, make decisions, and adapt when conditions change unexpectedly, using kitchen cleaning scenarios as examples.

• →Microsoft Research has developed AsgardBench, a new benchmark for embodied AI systems that can adapt to changing environments.
• →The benchmark tests AI's ability to perform visually grounded interactive planning in real-world scenarios.
• →Kitchen cleaning tasks serve as the primary testing environment for evaluating adaptive decision-making.
• →The system evaluates how AI handles unexpected situations like objects being in different states than anticipated.
• →This represents Microsoft's continued investment in advancing practical AI applications for robotics.