AI RESEARCH

BEDTime: A Unified Benchmark for Automatically Describing Time Series

arXiv CS.LG

ArXi:2509.05215v3 Announce Type: replace-cross Recent works propose complex multi-modal models that handle both time series and language, ultimately claiming high performance on complex tasks like time series reasoning and cross-modal question answering. However, they skip foundational evaluations that such complex models should have mastered.