Papers by Azher Ahmed Efat
Beyond Single Plots: A Benchmark for Question Answering on Multi-Charts (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing research on chart understanding has been limited to single chart images. |
| Approach: | They propose a dataset specifically designed for question answering over multi-chart images. |
| Outcome: | The proposed method shows a 27.4% LLM-based accuracy drop on human-authored questions and a 5.39% gain in the human-generated questions. |
Assessing Y-Axis Influence: Bias in Multimodal Language Models on Chart-to-Table Translation (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing studies have not examined the y-axis-related biases in chart-to-table translation models. |
| Approach: | They propose a framework for analyzing y-axis-related bias on five state-of-the-art models. |
| Outcome: | The proposed framework analyzes y-axis-related biases on five state-of-the-art models. |