Charts as Epistemic Actions for Experience-Driven Agentic Data Analysis
Data agents built on ReAct-style loops now execute complex analytical workflows over structured data, yet their intermediate reasoning still runs almost entirely through text.
ProofPaper ↗
Key points
- Human analysts instead move between numbers, tables, and charts, reading visual evidence to expose trends and anomalies that text summaries drop.
- We ask whether charts can serve as internal epistemic actions inside a data agent's reasoning loop, and how to teach an agent to use them.
- Agents score as well from self-generated charts as from numeric output, yet plotting access alone does not translate into consistent use.
- Task slices show that charts lead on temporal, pattern, and anomaly reasoning, while text leads on precise reading and conditional verification.
Sources (1)
- [1]Charts as Epistemic Actions for Experience-Driven Agentic Data AnalysisarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 21, 05:26 PM
Data agents built on ReAct-style loops now execute complex analytical workflows over structured data, yet their intermediate reasoning still runs almost entirely through text.
Human analysts instead move between numbers, tables, and charts, reading visual evidence to expose trends and anomalies that text summaries drop.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 21, 2026Granite Speech 5.0: Transcribe, Translate, and Attribute more Accurately and Faster
- Oct 20, 2026ALTK-Evolve: How learned memory tops the τ2-bench leaderboard on customer service tasks without a frontier model
- Oct 20, 2026Voxtral 3
- Oct 20, 2026Cache Engineering for Agents
- Oct 19, 2026RealControlBench: A Benchmark for Interpretability-Informed Fine-Grained Control of LLMs and Agents
- Oct 19, 2026Evaluating Agents on Highly Subjective Tasks