Paper: arXiv 2611.07464

Authors: Yupeng Cao, Xueqing Peng, Lingfei Qian, Haohang Li, Yangyang Yu, Anke Xu, Wenbo Cao, Yan Wang, Yuechen Jiang, Yuyan Wang, Zhuohan Xie, Yankai Chen, Zehan Li, Jimin Huang, Xue Liu, Jian-Yun Nie, Sophia Ananiadou

Abstract

Language agents increasingly work for users on the open Web, and finance is where trusting them is hardest. The data must come from the Web, the strategy lives in the user’s head, and correctness depends on time, so an agent can fail silently: a strategy the user never asked for, unverified data, or a backtest that uses the future yields a report as convincing as a correct one. Existing benchmarks close the world somewhere, fixing the task, the data, or the timeline, and thereby remove these failures instead of measuring them. We introduce FINCOSMOS, an open-world evaluation paradigm for agents doing financial work. A simulated user holds a hidden strategy sampled from a compositional grammar and reveals it only when asked. The agent receives no data, so it must acquire data from the Web, build its own backtest under a data cutoff, and deliver verifiable results. Grading needs no reference answer. Programs verify point-in-time correctness and recompute every reported metric, the delivered strategy is checked against the hidden specification, and a run is accepted only when the user is satisfied and none of 15 veto checks fails. Across four coding agents and a model without tools in 150 sessions, no session is accepted: every session that the simulated user signed off fails a program check, and agents raise only 54–64% of the details they were not told. Human experts attribute 88.33% of failures to the agents. With FINCOSMOS feedback, an agent with persistent memory learned all ten tasks.

Complexity vs Empirical Score

  • Math Complexity: 6.0/10
  • Empirical Rigor: 9.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: This paper introduces a highly novel and empirically rigorous evaluation paradigm for financial AI agents. It addresses critical issues of trust and silent failures in open-world financial tasks, demonstrating strong methodological clarity and a commitment to reproducibility.

Research Flowchart

  flowchart TD
    A[Research Goal: Evaluate Financial Agents in Open-World Scenarios] --> B{Key Methodology: FINCOSMOS Paradigm};
    B --> C[Inputs: Simulated User Hidden Strategy (Compositional Grammar), No Agent Data];
    C --> D[Computational Processes: Agent Acquires Web Data, Builds Backtest (Data Cutoff), Delivers Results];
    D --> E[Verification & Grading: Programmatic Checks (Point-in-Time Correctness, Metric Recomputation), Strategy vs. Hidden Spec, 15 Veto Checks];
    E --> F[Key Findings: No Agent Accepted (0/150 sessions), Failures Attributed to Agents (88.33%), Agents Miss Details (54-64%)];
    F --> G[Outcome: FINCOSMOS Feedback Enables Agent Learning (all 10 tasks for persistent memory agent)];