Knowledge-Based Visual Question Answering (KB-VQA) requires grounding visual queries to external knowledge beyond directly observable content in images.While recent multi modal large language models (MLLMs) show strong perceptual abilities, they struggle on KB-VQA tasks requiring groundings from both fine-grained entity and evidence levels.Most existing multi-modal retrieval augmented generation (MM-RAG) methods tightly couple entity discrimination and section-level evidence ranking into a single re-ranking stage, leading to high cost and limited generalization.In this work, we revisit existing MM-RAG solutions from a workflow perspective and argue both entity-level and fact-level groundings are key bottlenecks.We observe that although MLLMs often fail under open-ended entity naming, they can better identify the correct entity when selecting from a small set of candidate names.Based on this insight, we propose a simple and training-free identify-before-answer IBA framework that decouples entity identification from section-level re-ranking.Our approach prompts an MLLM to select high-confidence entities using only candidate names, followed by an off-the-shelf textual re-ranker for evidence selection.Experiments on Encyclopedic-VQA and InfoSeek show that our method consistently outperforms fine-tuned multi-modal re-ranking baselines while reducing training and inference complexity.Additional analyses reveal that the improvements arise not only from better entity identification, but also from selecting more informative evidence once correct entity is fixed.Our implementation is made public to ease reproducibility

We present an architecture to boost the precision of existing information extraction systems. This is achieved by augmenting the existing parser, which may be constraint-based or hybrid statistical, with a character-level neural network. Our architecture combines the ability of constraint-based or hybrid extraction systems to easily incorporate domain knowledge with the ability of deep neural networks to leverage large amounts of data to learn complex features. The network is trained using a measure of consistency between extracted data and existing databases as a form of cheap, noisy supervision. Our architecture does not require large scale manual annotation or a system rewrite. It has led to large precision improvements over an existing, highly-tuned production information extraction system used at Bloomberg LP for financial language text.