SCALER:Synthetic Scalable Adaptive Learning Environment for Reasoning
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910189047775232 |
|---|---|
| author | Xu, Caijun Xiao, Changyi Peng, Zhongyuan Wang, Xinrun Cao, Yixin |
| author_facet | Xu, Caijun Xiao, Changyi Peng, Zhongyuan Wang, Xinrun Cao, Yixin |
| contents | Reinforcement learning (RL) offers a principled way to enhance the reasoning capabilities of large language models, yet its effectiveness hinges on training signals that remain informative as models evolve. In practice, RL progress often slows when task difficulty becomes poorly aligned with model capability, or when training is dominated by a narrow set of recurring problem patterns. To jointly address these issues, we propose SCALER (Synthetic sCalable Adaptive Learning Environment for Reasoning), a framework that sustains effective learning signals through adaptive environment design. SCALER introduces a scalable synthesis pipeline that converts real-world programming problems into verifiable reasoning environments with controllable difficulty and unbounded instance generation, enabling RL training beyond finite datasets while preserving strong correctness guarantees. Building on this, SCALER further employs an adaptive multi-environment RL strategy that dynamically adjusts instance difficulty and curates the active set of environments to track the model's capability frontier and maintain distributional diversity. This co-adaptation prevents reward sparsity, mitigates overfitting to narrow task patterns, and supports sustained improvement throughout training. Extensive experiments show that SCALER consistently outperforms dataset-based RL baselines across diverse reasoning benchmarks and exhibits more stable, long-horizon training dynamics. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_04809 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SCALER:Synthetic Scalable Adaptive Learning Environment for Reasoning Xu, Caijun Xiao, Changyi Peng, Zhongyuan Wang, Xinrun Cao, Yixin Artificial Intelligence Reinforcement learning (RL) offers a principled way to enhance the reasoning capabilities of large language models, yet its effectiveness hinges on training signals that remain informative as models evolve. In practice, RL progress often slows when task difficulty becomes poorly aligned with model capability, or when training is dominated by a narrow set of recurring problem patterns. To jointly address these issues, we propose SCALER (Synthetic sCalable Adaptive Learning Environment for Reasoning), a framework that sustains effective learning signals through adaptive environment design. SCALER introduces a scalable synthesis pipeline that converts real-world programming problems into verifiable reasoning environments with controllable difficulty and unbounded instance generation, enabling RL training beyond finite datasets while preserving strong correctness guarantees. Building on this, SCALER further employs an adaptive multi-environment RL strategy that dynamically adjusts instance difficulty and curates the active set of environments to track the model's capability frontier and maintain distributional diversity. This co-adaptation prevents reward sparsity, mitigates overfitting to narrow task patterns, and supports sustained improvement throughout training. Extensive experiments show that SCALER consistently outperforms dataset-based RL baselines across diverse reasoning benchmarks and exhibits more stable, long-horizon training dynamics. |
| title | SCALER:Synthetic Scalable Adaptive Learning Environment for Reasoning |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2601.04809 |