From Recognition to Cognition: Visual Commonsense Reasoning
This work addresses the need for higher-order cognition in vision systems, enabling better understanding of actions and mental states in images, which is crucial for applications like AI assistants and robotics.
The paper tackles the problem of visual commonsense reasoning, requiring machines to answer questions about images and provide rationales, and introduces the VCR dataset with 290k QA problems where humans achieve over 90% accuracy but state-of-the-art models only reach about 45%. The proposed R2C model improves performance to around 65%, narrowing the gap but not solving the challenge.
Visual understanding goes well beyond object recognition. With one glance at an image, we can effortlessly imagine the world beyond the pixels: for instance, we can infer people's actions, goals, and mental states. While this task is easy for humans, it is tremendously difficult for today's vision systems, requiring higher-order cognition and commonsense reasoning about the world. We formalize this task as Visual Commonsense Reasoning. Given a challenging question about an image, a machine must answer correctly and then provide a rationale justifying its answer. Next, we introduce a new dataset, VCR, consisting of 290k multiple choice QA problems derived from 110k movie scenes. The key recipe for generating non-trivial and high-quality problems at scale is Adversarial Matching, a new approach to transform rich annotations into multiple choice questions with minimal bias. Experimental results show that while humans find VCR easy (over 90% accuracy), state-of-the-art vision models struggle (~45%). To move towards cognition-level understanding, we present a new reasoning engine, Recognition to Cognition Networks (R2C), that models the necessary layered inferences for grounding, contextualization, and reasoning. R2C helps narrow the gap between humans and machines (~65%); still, the challenge is far from solved, and we provide analysis that suggests avenues for future work.