Dialog without Dialog Data: Learning Visual Dialog Agents from VQA Data
This addresses the challenge of efficient adaptation for visually grounded dialog agents, reducing the need for expensive data collection, though it appears incremental as it builds on existing VQA data and methods.
The paper tackles the problem of developing visually grounded dialog agents that can adapt to new tasks without forgetting how to talk to people, by factorizing intention and language to minimize linguistic drift, and shows through qualitative results, automated metrics, and human studies that the model maintains language quality while baselines fail or become unintelligible.
Can we develop visually grounded dialog agents that can efficiently adapt to new tasks without forgetting how to talk to people? Such agents could leverage a larger variety of existing data to generalize to new tasks, minimizing expensive data collection and annotation. In this work, we study a setting we call "Dialog without Dialog", which requires agents to develop visually grounded dialog models that can adapt to new tasks without language level supervision. By factorizing intention and language, our model minimizes linguistic drift after fine-tuning for new tasks. We present qualitative results, automated metrics, and human studies that all show our model can adapt to new tasks and maintain language quality. Baselines either fail to perform well at new tasks or experience language drift, becoming unintelligible to humans. Code has been made available at https://github.com/mcogswell/dialog_without_dialog