Ensemble of MRR and NDCG models for Visual Dialog
This work addresses the problem of robust evaluation for visual dialog systems, which is crucial for researchers and developers in AI and human-computer interaction, though it is incremental as it builds on existing metrics and models.
The paper tackles the challenge of evaluating AI agents in visual dialog by addressing the limitations of existing metrics like MRR and NDCG, which favor different aspects of answer quality, and proposes a two-step non-parametric ranking approach that merges strong models for both metrics, achieving near state-of-the-art performance (70.41% MRR vs. 71.24% and 72.16% NDCG vs. 75.35%) and winning the Visual Dialog 2020 challenge.
Assessing an AI agent that can converse in human language and understand visual content is challenging. Generation metrics, such as BLEU scores favor correct syntax over semantics. Hence a discriminative approach is often used, where an agent ranks a set of candidate options. The mean reciprocal rank (MRR) metric evaluates the model performance by taking into account the rank of a single human-derived answer. This approach, however, raises a new challenge: the ambiguity and synonymy of answers, for instance, semantic equivalence (e.g., `yeah' and `yes'). To address this, the normalized discounted cumulative gain (NDCG) metric has been used to capture the relevance of all the correct answers via dense annotations. However, the NDCG metric favors the usually applicable uncertain answers such as `I don't know. Crafting a model that excels on both MRR and NDCG metrics is challenging. Ideally, an AI agent should answer a human-like reply and validate the correctness of any answer. To address this issue, we describe a two-step non-parametric ranking approach that can merge strong MRR and NDCG models. Using our approach, we manage to keep most MRR state-of-the-art performance (70.41% vs. 71.24%) and the NDCG state-of-the-art performance (72.16% vs. 75.35%). Moreover, our approach won the recent Visual Dialog 2020 challenge. Source code is available at https://github.com/idansc/mrr-ndcg.