Knowledge, Rules and Their Embeddings: Two Paths towards Neuro-Symbolic JEPA
Yongchao Huang, Hassan Raza · 2026
Read Knowledge, Rules and Their Embeddings: Two Paths towards Neuro-Symbolic JEPA by Yongchao Huang, Hassan Raza (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-13265.txt.
Description
Yongchao Huang and Hassan Raza outline two paths toward neuro-symbolic JEPA models. Predictive models learn rich statistical patterns but lack verifiable human logic, so shortcuts can mislead them. Rule systems give rigorous logic yet face discrete bounds and combinatorial explosion. Their RiJEPA framework first injects structured bias into JEPA training through Energy-Based Constraints. A dual-encoder design reshapes the representation space into logical basins. The second path relaxes rules into continuous logic and uses gradient-guided Langevin diffusion to find new rules.
Catalog record
| CREATOR | Yongchao Huang, Hassan Raza |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 94395 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.13265