Robusto-1 Dataset: Comparing Humans and VLMs on real out-of-distribution Autonomous Driving VQA from Peru
This work addresses the cognitive alignment problem for deploying multimodal foundational models in self-driving cars, particularly in challenging environments, but it is incremental as it builds on existing VQA and RSA methods.
The authors created the Robusto-1 dataset using dashcam video from Peru to compare human and visual language model (VLM) responses in out-of-distribution autonomous driving scenarios via visual question answering, finding that alignment between humans and VLMs varies significantly based on question type.
As multimodal foundational models start being deployed experimentally in Self-Driving cars, a reasonable question we ask ourselves is how similar to humans do these systems respond in certain driving situations -- especially those that are out-of-distribution? To study this, we create the Robusto-1 dataset that uses dashcam video data from Peru, a country with one of the worst (aggressive) drivers in the world, a high traffic index, and a high ratio of bizarre to non-bizarre street objects likely never seen in training. In particular, to preliminarly test at a cognitive level how well Foundational Visual Language Models (VLMs) compare to Humans in Driving, we move away from bounding boxes, segmentation maps, occupancy maps or trajectory estimation to multi-modal Visual Question Answering (VQA) comparing both humans and machines through a popular method in systems neuroscience known as Representational Similarity Analysis (RSA). Depending on the type of questions we ask and the answers these systems give, we will show in what cases do VLMs and Humans converge or diverge allowing us to probe on their cognitive alignment. We find that the degree of alignment varies significantly depending on the type of questions asked to each type of system (Humans vs VLMs), highlighting a gap in their alignment.