Enhancing Explainability for Exchange Rate Forecasting via Self-Reflective Reinforcement Learning
Bibliographic record
Abstract
Exchange rate forecasting faces the challenge of balancing accuracy and explainability. Meanwhile, large language models (LLMs) are mainly used for feature extraction and have few applications in generating explainable forecast results, and there is a lack of in-depth analysis of the reasoning process. In order to address this issue, we propose the Self-Reflective Reinforcement Learning (SRRL) framework, which integrates machine learning, LLMs and reinforcement learning. The framework’s methodology is as follows: firstly, it utilises the forecast results of the Space-Time and Times-Net fusion model (ST-fusion model) to automatically construct a dataset that includes “explanation-reflection”. Secondly, it uses this dataset to fine-tune an LLM via reinforcement learning. Finally, the fine-tuned LLM generates forecast signals and reasonable explanations. The experimental results demonstrate that the explanations generated by the fine-tuned LLM show significant improvements in logical coherence and relevance to input evidence, thereby effectively enhancing the quality of model explainability.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.015 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".