Teaching Machines to Describe Images via Natural Language Feedback
This addresses the challenge of making robots more adaptable in household settings by allowing non-experts to provide effective guidance, though it is incremental as it builds on existing captioning methods.
The paper tackles the problem of enabling non-expert users to guide learning agents by using natural language feedback instead of numeric rewards, specifically in image captioning, and shows that their model performs better with descriptive feedback than with independently written human captions.
Robots will eventually be part of every household. It is thus critical to enable algorithms to learn from and be guided by non-expert users. In this paper, we bring a human in the loop, and enable a human teacher to give feedback to a learning agent in the form of natural language. We argue that a descriptive sentence can provide a much stronger learning signal than a numeric reward in that it can easily point to where the mistakes are and how to correct them. We focus on the problem of image captioning in which the quality of the output can easily be judged by non-experts. We propose a hierarchical phrase-based captioning model trained with policy gradients, and design a feedback network that provides reward to the learner by conditioning on the human-provided feedback. We show that by exploiting descriptive feedback our model learns to perform better than when given independently written human captions.