Back to first-pass browser

First-pass extracted concept

convolutional feedforward networks

Candidate: concept label1 source documents4 linked claims
Jump to explainersJump to evidenceJump to sourcesJump to claims
Live refresh every 5sNext refresh in 5s

Extracted Explainers

What the tool is doing

Convolutional feedforward networks are described as the dominant model class in computer vision. They are said to draw inspiration from the primate visual hierarchy.

Source 1DOIPubMed

What problem it solves

They support strong visual recognition performance in artificial systems.

Source 1DOIPubMed

What it does not solve

The abstract indicates these models are not yet designed primarily to model brain computations faithfully.

Source 1DOIPubMed

Alternatives

The abstract mentions recurrent computational models as an important future direction alongside feedforward models.

Source 1DOIPubMed

Evidence Snippets

Convolutional feedforward networks, which now dominate computer vision, take further inspiration from the architecture of the primate visual hierarchy.
Evidence 1Source 1DOIPubMedprovenance

Supporting Sources

Linked Claims

Claim 1architecture relationshipsupports2015Source 1DOIPubMed

Convolutional feedforward networks dominate computer vision and take inspiration from the architecture of the primate visual hierarchy.

Claim 2future directionsupports2015Source 1DOIPubMed

Biologically faithful feedforward and recurrent computational models of how biological brains perform high-level feats of intelligence, including vision, are becoming feasible.

Claim 3limitationsupports2015Source 1DOIPubMed

Current models are designed with engineering goals rather than to model brain computations.

Claim 4representational alignmentsupports2015Source 1DOIPubMed

Initial studies comparing internal representations between current neural network models and primate brains find surprisingly similar representational spaces.