Pulvinar Pathways as Skip Connections in Deep Neural Networks for Vision

A research document describing experiments with a pulvinar-inspired long-range skip pathway in a hierarchical convolutional vision model. The model was evaluated on CIFAR-10 categorization and a near-threshold contrast-detection task with noisy backgrounds. The document was authored by Narmin Zarinabadi and last updated on 2026-05-14.

Use Cases

• Investigating the role of transthalamic pathways in neural network models based on the described pulvinar-inspired connectivity.
• Analyzing the effects of gain-controlled gating on model stability and contrast-dependent response scaling as described in the CIFAR-10 experiment.
• Studying decision-criterion shifts under uncertainty using the described near-threshold detection task with noisy backgrounds.

Strengths

• The research is clearly described with specific experimental tasks (CIFAR-10 categorization and a near-threshold detection task).
• The document is licensed under CC-BY-4.0, permitting reuse with attribution.

Limitations

• The dataset is a 20.9 KB DOCX file, indicating a very small document rather than a large-scale data collection.
• Row count and column-level documentation are absent; the actual data content must be inspected after download.
• The description is a research summary; the availability of underlying model weights or raw experimental data is unknown.

Provenance

Source

figshare, authored by Narmin Zarinabadi.

Collection Method

Research involving a minimal pulvinar-inspired skip pathway implemented in a convolutional vision model.

Freshness

Last updated 2026-05-14 04:25:08; freshness should be verified.