9 papers
TimeSage-EV: A Live Benchmark for Agentic Time Series Analysis in Evolving Environments
Qingren Yao, Yaxuan Kong, Yuqi Nie +6
Time series analysis in high-stakes domains relies on recurring data releases, where new observations can alter the evidence base and the validity of later conclusions. Existing ti…
TimeSage-MT: A Multi-Turn Benchmark for Evaluating Agentic Time Series Reasoning
Yaxuan Kong, Qingren Yao, Yuqi Nie +7
Time series data inform critical decisions across many real-world domains. While large language model (LLM) agents can analyze data through natural language and tools, it remains u…
Evaluating LLMs in Finance Requires Explicit Bias Consideration
Yaxuan Kong, Hoyoung Lee, Yoontae Hwang +7
Large Language Models (LLMs) are increasingly integrated into financial workflows, but evaluation practice has not kept up. Finance-specific biases can inflate performance, contami…
Achieving Time Series Reasoning Requires Rethinking Model Design, Tasks Formulation, and Evaluation
Yaxuan Kong, Yiyuan Yang, Shiyu Wang +7
Understanding time series data is fundamental to many real-world applications. Recent work explores multimodal large language models (MLLMs) to enhance time series understanding wi…
Signature-Informed Transformer for Asset Allocation
Yoontae Hwang, Stefan Zohren
Modern deep learning for asset allocation typically separates forecasting from optimization. We argue this creates a fundamental mismatch where minimizing prediction errors fails t…
Fusing Narrative Semantics for Financial Volatility Forecasting
Yaxuan Kong, Yoontae Hwang, Marcus Kaiser +3
We introduce M2VN: Multi-Modal Volatility Network, a novel deep learning-based framework for financial volatility forecasting that unifies time series features with unstructured ne…