Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

Depth Perception and Spatial Vision

Depth Perception and Spatial Vision

Depth perception is the ability to perceive objects three-dimensionally. It relies on two types of cues: binocular and monocular. Binocular cues depend on the combination of images from both eyes and how the eyes work together. Since the eyes are in slightly different positions, each eye captures a slightly different image. This disparity between images, known as binocular disparity, helps the brain interpret depth. When the brain compares these images, it determines the distance to an object.

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Acoustic Analysis of Primary Care Patient-Clinician Conversations to Screen for Cognitive Impairment.

JAMA neurology·2026

Same author

THOI: An efficient and accessible library for computing higher-order interactions enhanced by batch-processing.

PloS one·2026

Same author

Evolving reservoir computers reveal bidirectional coupling between predictive power and emergent dynamics.

Patterns (New York, N.Y.)·2026

Same author

Synergistic Motifs in Gaussian Systems.

Physical review letters·2026

Same author

Competitive interactions shape mammalian brain network dynamics and computation.

Nature neuroscience·2026

Same author

SMARCA4-deficient non-small cell lung cancer and immunotherapy resistance: mechanisms and novel strategies-a mini-review.

Discover oncology·2026

Search research articles

Related Experiment Video

Updated: May 5, 2026

Methods to Explore the Influence of Top-down Visual Processes on Motor Behavior

Methods to Explore the Influence of Top-down Visual Processes on Motor Behavior

Published on: April 16, 2014

Directional Confusions Reveal Divergent Inductive Biases Through Rate-Distortion Geometry in Human and Machine

Leyla Roksan Caglar¹, Pedro A M Mediano², Baihan Lin³

¹Windreich Department of AI and Human Health, Icahn School of Medicine, at Mount Sinai, New York, NY, USA.

|May 4, 2026

Summary

Directional confusions in image classification reveal distinct inductive biases invisible to accuracy alone. Analyzing human and deep vision model errors shows differing asymmetry patterns, offering new insights into generalization geometry.

More Related Videos

Generating Strictly Controlled Stimuli for Figure Recognition Experiments

Generating Strictly Controlled Stimuli for Figure Recognition Experiments

Published on: March 18, 2019

Perceptual and Category Processing of the Uncanny Valley Hypothesis' Dimension of Human Likeness: Some Methodological Issues

Perceptual and Category Processing of the Uncanny Valley Hypothesis' Dimension of Human Likeness: Some Methodological Issues

Published on: June 3, 2013

Related Experiment Videos

Last Updated: May 5, 2026

Methods to Explore the Influence of Top-down Visual Processes on Motor Behavior

Methods to Explore the Influence of Top-down Visual Processes on Motor Behavior

Published on: April 16, 2014

Generating Strictly Controlled Stimuli for Figure Recognition Experiments

Generating Strictly Controlled Stimuli for Figure Recognition Experiments

Published on: March 18, 2019

Perceptual and Category Processing of the Uncanny Valley Hypothesis' Dimension of Human Likeness: Some Methodological Issues

Perceptual and Category Processing of the Uncanny Valley Hypothesis' Dimension of Human Likeness: Some Methodological Issues

Published on: June 3, 2013

Area of Science:

Computer Vision
Cognitive Science
Machine Learning

Background:

Humans and deep vision models achieve comparable image classification accuracy.
However, they make systematically different types of errors, differing in the direction and nature of confusions.
These directional confusions are overlooked when relying solely on overall accuracy metrics.

Purpose of the Study:

To investigate whether directional confusions in image classification reveal distinct inductive biases.
To compare the generalization geometry of human and deep vision model performance under various perturbations.
To understand how robustness training affects these inductive biases and their representation in Rate-Distortion geometry.

Main Methods:

Collected human and deep vision model responses on a natural-image categorization task across 12 perturbation types.
Quantified asymmetry in confusion matrices to identify directional error patterns.
Linked confusion matrix asymmetry to generalization geometry using a Rate-Distortion (RD) framework, analyzing slope (beta), curvature (kappa), and efficiency (AUC).

Main Results:

Humans exhibit broad but weak directional asymmetries, while deep vision models display sparser, stronger directional collapses.
Robustness training reduced global asymmetry but did not restore the human-like breadth-strength profile of graded similarity.
Different organizations of asymmetry shifted the RD frontier in opposite directions, even with matched performance.

Conclusions:

Directional confusions serve as interpretable signatures of inductive biases, particularly under distribution shifts.
Rate-Distortion geometry provides a framework to quantify and compare these biases between humans and artificial systems.
Understanding these differences is crucial for developing more human-like and robust artificial intelligence.