Diversify, Rationalize, and Combine: Ensembling Multiple QA Strategies for Zero-shot Knowledge-based VQA
This work addresses the challenge of insufficient single strategies for K-VQA, offering a domain-specific solution for visual question answering tasks.
The paper tackles the problem of knowledge-based visual question answering (K-VQA) by proposing DietCoke, an ensemble method that combines multiple question-answering strategies with rationales, resulting in performance improvements of 2.8% on OK-VOA and 4.7% on A-OKVQA over state-of-the-art baselines.
Knowledge-based Visual Question-answering (K-VQA) often requires the use of background knowledge beyond the image. However, we discover that a single knowledge generation strategy is often insufficient for all K-VQA questions. To this end, we propose Diversification, Evidence Truncation, and Combination for Knowledge-based Elucidation (DietCoke), which utilizes a bundle of complementary question-answering tactics and aggregates their answers using textual rationales. DietCoke comprises of three stages: diversification, rationalization, and ensemble. The diversification stage generates three distinctive decision contexts, each leading to its own answer candidate. The rationalization stage generates two rationales, the automatic rationale and the mechanistic rationale, for each answer candidate using decorrelated techniques. Finally, in the ensemble stage, an LLM informed by the rationales selects one answer from the three candidates. Experiments show that DietCoke significantly outperforms state-of-the-art LLM-based baselines by 2.8% on OK-VOA and 4.7% on A-OKVOA and that the strategies in the ensembles are highly complementary. Code is available at: https://github.com/limiaoyu/DietCoke