Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.31273 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913172900806656 |
|---|---|
| author | Nguimatsia-Tiofack, Franki Schramm, Fabian Hellard, Théotime Le Carpentier, Justin |
| author_facet | Nguimatsia-Tiofack, Franki Schramm, Fabian Hellard, Théotime Le Carpentier, Justin |
| contents | While self-supervised Contrastive Reinforcement Learning (CRL) has shown remarkable depth-scaling capabilities, successfully using networks over 64 layers, scaled CRL still struggles with long-horizon goal-conditioned planning due to the uniformity-tolerance dilemma inherent in contrastive losses. We introduce Survival Reinforcement Learning (SRL), an online classification-based alternative that extends the survival value learning framework by maximizing the agent's dwell time at target goals. SRL bypasses the structural constraints of CRL and mitigates the "bang-bang" control solutions inherent to survival frameworks, which often induce undesirable behavior in complex dynamical systems. Evaluated across diverse robotic benchmarks, scaled SRL matches state-of-the-art CRL on manipulation tasks and outperforms it by 2x to 8x on stable, long-horizon locomotion tasks. Our results provide strong additional evidence that classification-based methods may serve as a key primitive in the broader effort to scale reinforcement learning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_31273 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Survival Reinforcement Learning: Toward Scalable Self-Supervised RL Nguimatsia-Tiofack, Franki Schramm, Fabian Hellard, Théotime Le Carpentier, Justin Machine Learning While self-supervised Contrastive Reinforcement Learning (CRL) has shown remarkable depth-scaling capabilities, successfully using networks over 64 layers, scaled CRL still struggles with long-horizon goal-conditioned planning due to the uniformity-tolerance dilemma inherent in contrastive losses. We introduce Survival Reinforcement Learning (SRL), an online classification-based alternative that extends the survival value learning framework by maximizing the agent's dwell time at target goals. SRL bypasses the structural constraints of CRL and mitigates the "bang-bang" control solutions inherent to survival frameworks, which often induce undesirable behavior in complex dynamical systems. Evaluated across diverse robotic benchmarks, scaled SRL matches state-of-the-art CRL on manipulation tasks and outperforms it by 2x to 8x on stable, long-horizon locomotion tasks. Our results provide strong additional evidence that classification-based methods may serve as a key primitive in the broader effort to scale reinforcement learning. |
| title | Survival Reinforcement Learning: Toward Scalable Self-Supervised RL |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2605.31273 |