Situated Reasoning
Situated reasoning focuses on developing AI systems that can understand and reason about information within its specific context, encompassing sensory inputs, past experiences, and ongoing interactions. Current research emphasizes creating large-scale, multimodal datasets and benchmarks that evaluate AI's ability to perform complex reasoning tasks in dynamic, real-world scenarios, often involving vision-language models and graph-based representations. This research is crucial for advancing embodied AI, improving human-computer interaction, and creating more robust and adaptable intelligent systems across various applications, such as robotics and assistive technologies.
Papers
SOK-Bench: A Situated Video Reasoning Benchmark with Aligned Open-World Knowledge
Andong Wang, Bo Wu, Sunli Chen, Zhenfang Chen, Haotian Guan, Wei-Ning Lee, Li Erran Li, Chuang Gan
STAR: A Benchmark for Situated Reasoning in Real-World Videos
Bo Wu, Shoubin Yu, Zhenfang Chen, Joshua B Tenenbaum, Chuang Gan