50 lines
2.2 KiB
Markdown
50 lines
2.2 KiB
Markdown
# History
|
|
|
|
## Provenance
|
|
|
|
Synaptopus grows out of a much earlier line of work: a 1989 master's thesis project at The University of Texas at Arlington by Wesley Royce Elsberry on hybrid artificial neural network modelling.
|
|
|
|
That original system combined multiple architecture families in a single loop:
|
|
|
|
- a Hopfield-Tank style generator
|
|
- a backpropagation-based critic
|
|
- an ART-style novelty and category mechanism
|
|
- a rule-based instructor and acceptance policy around them
|
|
|
|
The important idea was not just that neural networks could be used for a task, but that unlike neural systems could be made to cooperate, constrain one another, and contribute different functional roles within a larger process.
|
|
|
|
## Why A Separate Repository
|
|
|
|
The thesis reconstruction and Python port made the historical system accessible again, but it also clarified that the deeper contribution was architectural rather than domain-bound. The composition project is one concrete application of a broader pattern:
|
|
|
|
- heterogeneous neural components
|
|
- explicit orchestration
|
|
- inspectable intermediate states
|
|
- sequential acceptance and rejection loops
|
|
- evaluation beyond raw fitting or classification
|
|
|
|
Synaptopus exists to make that broader pattern the primary subject.
|
|
|
|
## Relationship To TriuneCadence
|
|
|
|
TriuneCadence is the thesis-focused reconstruction: historically grounded, composition-centered, and intentionally close to the original hybrid system.
|
|
|
|
Synaptopus is the broader framework direction: a place where reusable architecture interfaces, generic implementations, educational tools, and new multi-architecture experiments can live without being tied to one historical task.
|
|
|
|
In short:
|
|
|
|
- TriuneCadence is one important exemplar
|
|
- Synaptopus is the larger lab
|
|
|
|
## Intended Future
|
|
|
|
Over time, Synaptopus may include:
|
|
|
|
- generic architecture families beyond the original three
|
|
- additional domains beyond music
|
|
- execution graphs and visual workbenches
|
|
- browser-based and pedagogical interfaces
|
|
- experiment tracing, timing, and information-theoretic analysis
|
|
|
|
The aim is to support both serious experimentation and explanation: a system that can be used to build artificial neural systems and to teach how they work together.
|