Chart-based Reasoning: Transferring Capabilities from LLMs to VLMs
CoRR(2024)
摘要
Vision-language models (VLMs) are achieving increasingly strong performance
on multimodal tasks. However, reasoning capabilities remain limited
particularly for smaller VLMs, while those of large-language models (LLMs) have
seen numerous improvements. We propose a technique to transfer capabilities
from LLMs to VLMs. On the recently introduced ChartQA, our method obtains
state-of-the-art performance when applied on the PaLI3-5B VLM by
, while also enabling much better performance on PlotQA
and FigureQA.
We first improve the chart representation by continuing the pre-training
stage using an improved version of the chart-to-table translation task by
. We then propose constructing a 20x larger dataset than
the original training set. To improve general reasoning capabilities and
improve numerical operations, we synthesize reasoning traces using the table
representation of charts. Lastly, our model is fine-tuned using the multitask
loss introduced by .
Our variant ChartPaLI-5B outperforms even 10x larger models such as PaLIX-55B
without using an upstream OCR system, while keeping inference time constant
compared to the PaLI3-5B baseline. When rationales are further refined with a
simple program-of-thought prompt , our model outperforms
the recently introduced Gemini Ultra and GPT-4V.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要