Reference library
At the Edge of Chaos: Real-time Computations and Self-Organized Criticality in Recurrent Neural Networks
Paper summary
Random synchronous threshold networks are analyzed using common-input Hamming-distance growth and bit-flip sensitivity. Delayed parity and random Boolean-function readouts perform well near a critical line. A local sensitivity-based scaling rule approaches that line and improves measured capability. Readouts are trained anew with recurrent weights frozen at each evaluation; this is not unrestricted simultaneous online learning.
Relevance to the project
Assistant interpretation: suggests measuring useful input separation after discounting dependence on arbitrary initial state. It offers a concrete comparison for regime regulation but does not select a critical target or gain for continuous adaptive F/H networks.
Binary threshold units, random graph/input statistics and finite task families. Page 4 and the separately cited mean-field supplement were not read. This three-author conference paper is distinct from the two-author 2004 Neural Computation paper, and its bit-flip criterion is not a smooth-state Lyapunov exponent.
Source record
- Reference ID
- edge-chaos-threshold-2004
- Archived version
- NeurIPS17 conference2004 paper,8-page proceedings PDF; author order from actual title page
- Summary provenance
- reference-library/papers/edge-chaos-threshold-2004/metadata.json · summary
- Review status
- source-scoped reading; not replication
- Reading scope
- PDF pages 1–3,5–8 read in extracted text; page 4 and separately linked technical supplement not read. No implementation or replication.