ChartNet: MIT/IBM’s Million‑Scale Synthetic Chart Dataset with 1.5M Diverse Samples

MIT and IBM researchers introduce ChartNet, the largest code‑guided synthetic chart dataset containing 1.5 million multimodal samples across 24 chart types and six libraries, and demonstrate that fine‑tuning visual‑language models on it yields consistent, significant gains on chart reconstruction, data extraction, summarization, and reasoning tasks, outperforming much larger off‑the‑shelf models including GPT‑4o.

AI researchChartNetchart understanding

0 likes · 13 min read

ChartNet: MIT/IBM’s Million‑Scale Synthetic Chart Dataset with 1.5M Diverse Samples

Data Party THU

Sep 21, 2025 · Artificial Intelligence

How the New ECD Dataset Supercharges Multimodal LLM Chart Understanding

The paper introduces the Effective Chart Dataset (ECD), a large, high‑quality, diverse synthetic chart collection and the ECDBench benchmark, detailing a five‑stage modular synthesis pipeline, extensive QA generation, and experiments that show consistent performance gains for open‑source multimodal large language models on chart‑understanding tasks.

AIMLLMbenchmark

0 likes · 9 min read

How the New ECD Dataset Supercharges Multimodal LLM Chart Understanding

AI Frontier Lectures

Apr 3, 2025 · Artificial Intelligence

How ChartMoE Uses Sparse MoE to Master Chart Understanding and Preserve General Knowledge

ChartMoE, an oral paper at ICLR 2025, introduces a multi‑stage alignment training pipeline and a diversified MoE Connector that dramatically improves chart comprehension while maintaining performance on general multimodal tasks, backed by extensive data construction, training recipes, and thorough evaluations.

ChartMoEMixture of ExpertsSparse Modeling

0 likes · 10 min read

How ChartMoE Uses Sparse MoE to Master Chart Understanding and Preserve General Knowledge