VTBench: A Multimodal Framework for Time-Series Classification with Chart-Based Representations

Madhumitha Venkatesan, Xuyang Chen, Dongyu Liu · 2026

Read VTBench: A Multimodal Framework for Time-Series Classification with Chart-Based Representations by Madhumitha Venkatesan, Xuyang Chen, Dongyu Liu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-27259.txt.

Description

Venkatesan, Chen, and Liu introduce VTBench, a multi-mode framework for time-series sorting using chart-based representations alongside raw sequences. Most deep models use raw numbers alone, overlooking alternative encodings. Texture encodings such as Gramian Angular Fields and Recurrence Plots turn series into images but need heavy prep work and read less clearly. Chart visualizations read more naturally and show promise in specific domains, yet lacked thorough evaluation across chart types, encodings, and datasets, which VTBench supplies.

Catalog record

CREATORMadhumitha Venkatesan, Xuyang Chen, Dongyu Liu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE83153
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.27259

Browse the full library · About the Tower of Babel