Research map / Time series / Foundation models for time series
Bootstrapped Exploration with Causal Reasoning: A Training Paradigm for Adaptive Forecasting Agent
Qingwen Zeng, Dajun Guo, Zhaoge Bi, lining chen, Jushang Qiu, Yitian Yang, Carl Yang, Huaming Chen, Ling Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add lesson-guided planning (ablation) | BECRA · ETTm1 | MSE: 1.197 -> 0.380 | Improved |
| add lesson-guided planning (ablation) | BECRA · ETTm2 | MSE: 1.034 -> 0.274 | Improved |
| add lesson-guided planning (ablation) | BECRA · ETTh1 | MSE: 0.563 -> 0.441 | Improved |
| add lesson-guided planning (ablation) | BECRA · Weather | MSE: 0.258 -> 0.240 | Improved |
| use contrast-aware UCB sampling instead of greedy (ablation) | BECRA · ETTh2 | MSE: 0.412 -> 0.367 | Improved |
| use contrast-aware UCB sampling instead of greedy (ablation) | BECRA · Electricity | MSE: 0.191 -> 0.171 | Improved |
| add lesson verification (ablation) | BECRA · ETTh2 | MSE: 0.432 -> 0.367 | Improved |
| add lesson verification (ablation) | BECRA · ETTm1 | MSE: 0.384 -> 0.380 | No clear effect |
| use in-context lessons instead of supervised fine-tuning (LLaMA) (ablation) | BECRA · Electricity | MSE: 0.184 -> 0.171 | Improved |
| use in-context lessons instead of fine-tuning (ChatGLM) (ablation) | BECRA · ETTh2 | MSE: 0.395 -> 0.367 | Improved |
| use in-context lessons instead of fine-tuning (LLaMA) (ablation) | BECRA · ETTh1 | MSE: 0.443 -> 0.441 | No clear effect |
| replace ChatGPT-4 reasoner with Llama-3.1 70B (ablation) | BECRA · Weather | MSE: 0.240 -> 0.254 | Hurt |
Closest papers by meaning
- ForecastCompass: Guiding Agentic Forecasting with Adaptive Factor Memory (NeurIPS 2026 · similarity 0.89)
- TimeClaw: A Time-Series AI Agent with Exploratory Execution Learning (NeurIPS 2026 · similarity 0.88)
- LangTime: A Language-Guided Unified Model for Time Series Forecasting with Proximal Policy Optimization (ICML 2025 · similarity 0.87)
- Rethinking Post-Training Recipes for Multimodal Time-Series Forecasting (NeurIPS 2026 · similarity 0.87)
- TimeXL: Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop (NeurIPS 2025 · similarity 0.86)
- TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot Forecaster (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
