PhilSci Archive

Explaining Neural Transitions through Resource Constraints

Klein, Colin (2021) Explaining Neural Transitions through Resource Constraints. In: UNSPECIFIED.

[img]
Preview
Text
PSA2020 Paper 1De1.pdf

Download (213kB) | Preview

Abstract

One challenge in explaining neural evolution is the formal equivalence of a variety of different computational architectures. Well-known results show that various architectures, including neural networks with a single hidden layer and a nonlinear activation function, can be universal func- tion approximators (Hornik et al., 1989). Why change? The answer must involve the intense competition for resources—including time, space, and energy— under which brains operate (Sterling and Laughlin, 2015). I ar- gue that such explanations are ultimately an abstract species of resource explanation (Klein, 2018), which are distinct from but complementary to explanations in terms of mechanical parts. Resource explanations play an important role in computer science, one that is often under-appreciated by philosophers of neuroscience. As a case study, I show how the develop- ment of recurrence in neural networks can be favored when the increased complexity allows for more efficient use of existing resources. While re- source competition drives the change itself, the development of recurrence creates shifts in the landscape of what is evolvable. The resulting frame- work suggests a mechanisms by which major neural transitions can occur, and shows why organisms on either side of a transition boundary may have very similar cognitive capacities but very different potential for evolving new capacities.


Export/Citation: EndNote | BibTeX | Dublin Core | ASCII/Text Citation (Chicago) | HTML Citation | OpenURL
Social Networking:
Share |

Item Type: Conference or Workshop Item (UNSPECIFIED)
Creators:
CreatorsEmailORCID
Klein, Colincolin.klein@anu.edu.au0000-0002-7406-4010
Keywords: major transitions, resources, computation
Subjects: Specific Sciences > Computation/Information > Classical
Specific Sciences > Biology > Evolutionary Theory
Specific Sciences > Neuroscience > Cognitive Neuroscience
Depositing User: Prof Colin Klein
Date Deposited: 14 Nov 2021 01:47
Last Modified: 14 Nov 2021 01:47
Item ID: 19852
Subjects: Specific Sciences > Computation/Information > Classical
Specific Sciences > Biology > Evolutionary Theory
Specific Sciences > Neuroscience > Cognitive Neuroscience
Date: 13 November 2021
URI: https://philsci-archive.pitt.edu/id/eprint/19852

Monthly Views for the past 3 years

Monthly Downloads for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item