Boosting Text-to-Chart Retrieval through Training with Synthesized Semantic Insights

ArXi:2505.10043v4 Announce Type: replace-cross Text-to-chart retrieval, enabling users to find relevant charts via natural language queries, has gained significant attention. However, evaluating models in real-world business intelligence (BI) scenarios is challenging, as current benchmarks fail to simulate realistic user queries or test for deep semantic understanding with static chart images. To address this gap, we