ChartLens: Fine-grained Visual Attribution in Charts
This addresses the issue of unreliable chart interpretations for users in domains like finance and policy, though it is incremental as it builds on existing MLLM and segmentation techniques.
The paper tackles the problem of hallucinations in multimodal large language models (MLLMs) for chart understanding by introducing Post-Hoc Visual Attribution for Charts, which identifies fine-grained chart elements to validate responses, and shows that their method, ChartLens, improves fine-grained attributions by 26-66%.
The growing capabilities of multimodal large language models (MLLMs) have advanced tasks like chart understanding. However, these models often suffer from hallucinations, where generated text sequences conflict with the provided visual data. To address this, we introduce Post-Hoc Visual Attribution for Charts, which identifies fine-grained chart elements that validate a given chart-associated response. We propose ChartLens, a novel chart attribution algorithm that uses segmentation-based techniques to identify chart objects and employs set-of-marks prompting with MLLMs for fine-grained visual attribution. Additionally, we present ChartVA-Eval, a benchmark with synthetic and real-world charts from diverse domains like finance, policy, and economics, featuring fine-grained attribution annotations. Our evaluations show that ChartLens improves fine-grained attributions by 26-66%.