Reference library
A Modern Self-Referential Weight Matrix That Learns to Modify Itself
Paper summary
Introduces a self-referential weight matrix that generates signals used to modify its own matrix during execution. Outer-product and delta-rule operations make the construction finite and scalable; experiments evaluate few-shot learning and multitask reinforcement learning.
Relevance to the project
A concrete comparison for mutable per-neuron F and self-modification. Distinguish its supplied update primitives and external training from the project's fixed type-level updater and evolutionary goals.
Runtime self-modification does not imply unconstrained recursive improvement.
Source record
- Reference ID
- pmlr-irie22b
- Archived version
- Publisher/institution PDF advertised by article metadata
- Summary provenance
- reference-library/papers/pmlr-irie22b/metadata.json · summary
- Review status
- initial orientation summary; not a critical full-paper review
- Reading scope
- Selected abstract/opening and available closing discussion/conclusion passages in the preserved text extract; broader project notes provide context. No reproduction of experiments, exhaustive methods review, or equation/figure audit.