Explorations in Parallel Distributed Processing: A Handbook of Models, Programs, and Exercises

Livro digital

Título:
Explorations in Parallel Distributed Processing: A Handbook of Models, Programs, and Exercises

Autor:
James L. McClelland

Categoria:
Tecnologia > IA

Pessoa doadora:
Raffaello D. N.

Sinopse:
Readers who want to understand neural networks by building them from the ground up will find a hands-on path here, beginning with a MATLAB tutorial and moving through interactive activation, constraint satisfaction, and pattern-associator models. The table of contents makes the progression explicit: each major model is paired with implementation notes and exercises rather than presented as abstract theory alone. The handbook develops learning with Hebbian and delta rules, back propagation, competitive learning, self-organizing maps, simple recurrent networks, recurrent back propagation, and temporal-difference learning. Its later sections connect the models to practical work through installation guidance, network construction, software usage, and detailed exercises, with appendices documenting PDPTool and standalone operation. This combination of conceptual explanation, executable model design, and guided experimentation makes the volume useful for readers who want to see how distributed-processing ideas behave in code. It offers a coherent framework for studying representations, learning dynamics, sequence processing, and reinforcement learning while retaining the concrete exercises needed to turn each chapter into a working investigation.