Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2504.04711 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916676163862528 |
|---|---|
| author | Alhajir, Alfath Daryl Dodgson, Jennifer Lim, Joseph Phi, Truong Ma Peh, Julian Pattirane, Akira Rafhael Janson Poovaragan, Lokesh |
| author_facet | Alhajir, Alfath Daryl Dodgson, Jennifer Lim, Joseph Phi, Truong Ma Peh, Julian Pattirane, Akira Rafhael Janson Poovaragan, Lokesh |
| contents | Current large language models (LLMs) are constrained by human-derived training data and limited by a single level of abstraction that impedes definitive truth judgments. This paper introduces a novel framework in which AI models autonomously generate and validate new knowledge through direct interaction with their environment. Central to this approach is an unbounded, ungamable numeric reward - such as annexed disk space or follower count - that guides learning without requiring human benchmarks. AI agents iteratively generate strategies and executable code to maximize this metric, with successful outcomes forming the basis for self-retraining and incremental generalisation. To mitigate model collapse and the warm start problem, the framework emphasizes empirical validation over textual similarity and supports fine-tuning via GRPO. The system architecture employs modular agents for environment analysis, strategy generation, and code synthesis, enabling scalable experimentation. This work outlines a pathway toward self-improving AI systems capable of advancing beyond human-imposed constraints toward autonomous general intelligence. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_04711 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Generalising from Self-Produced Data: Model Training Beyond Human Constraints Alhajir, Alfath Daryl Dodgson, Jennifer Lim, Joseph Phi, Truong Ma Peh, Julian Pattirane, Akira Rafhael Janson Poovaragan, Lokesh Artificial Intelligence Current large language models (LLMs) are constrained by human-derived training data and limited by a single level of abstraction that impedes definitive truth judgments. This paper introduces a novel framework in which AI models autonomously generate and validate new knowledge through direct interaction with their environment. Central to this approach is an unbounded, ungamable numeric reward - such as annexed disk space or follower count - that guides learning without requiring human benchmarks. AI agents iteratively generate strategies and executable code to maximize this metric, with successful outcomes forming the basis for self-retraining and incremental generalisation. To mitigate model collapse and the warm start problem, the framework emphasizes empirical validation over textual similarity and supports fine-tuning via GRPO. The system architecture employs modular agents for environment analysis, strategy generation, and code synthesis, enabling scalable experimentation. This work outlines a pathway toward self-improving AI systems capable of advancing beyond human-imposed constraints toward autonomous general intelligence. |
| title | Generalising from Self-Produced Data: Model Training Beyond Human Constraints |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2504.04711 |