wichmann-lab

bds

Bayesian difference scaling. Statistical models for analyzing difference scaling experiments.

error-consistency

Error consistency: a black-box analysis for comparing errors between decision makers (NeurIPS 2020)

psignifit

Toolbox for Bayesian inference for psychometric functions

psychopy-pixx

Psychophysical experiments with high luminance resolution in Python, using Psychopy and VPixx devices

python-psignifit

Python implementation of psignifit, for psychometric function estimation

robustness-development

scaling-dimension

Data and code from "Estimating the perceived dimension of psychophysical stimuli using triplet accuracy and hypothesis testing"

spatial-vision-model

Model as described in Schuett & Wichmann 2017

teamproject-photography

In the team project summer term 2021, students are working on automating image manipulation using professional photography software.

trivial-or-impossible

Code for "Trivial or impossible—dichotomous data difficulty masks model differences (on ImageNet and beyond)" (ICLR 2022)

wichmann-lab/psignifit

Toolbox for Bayesian inference for psychometric functions

wichmann-lab/python-psignifit

Python implementation of psignifit, for psychometric function estimation

wichmann-lab/error-consistency

Error consistency: a black-box analysis for comparing errors between decision makers (NeurIPS 2020)

wichmann-lab/trivial-or-impossible

Code for "Trivial or impossible—dichotomous data difficulty masks model differences (on ImageNet and beyond)" (ICLR 2022)

wichmann-lab/robustness-development

wichmann-lab/psychopy-pixx

Psychophysical experiments with high luminance resolution in Python, using Psychopy and VPixx devices

wichmann-lab/spatial-vision-model

Model as described in Schuett & Wichmann 2017

wichmann-lab/scaling-dimension

Data and code from "Estimating the perceived dimension of psychophysical stimuli using triplet accuracy and hypothesis testing"

wichmann-lab/teamproject-photography

In the team project summer term 2021, students are working on automating image manipulation using professional photography software.

wichmann-lab/bds

Bayesian difference scaling. Statistical models for analyzing difference scaling experiments.

wichmann-lab/Stylized-ImageNet

Code to create Stylized-ImageNet, a stylized version of standard ImageNet (ICLR 2019 Oral)

wichmann-lab/supervised-learning-dynamics

wichmann-lab/posters

Here we share our posters and preprints, for example, for easy access on conferences.

wichmann-lab/causal-understanding-via-visualizations

Official repository for the paper "How Well do Feature Visualizations Support Causal Understanding of CNN Activations?".

wichmann-lab/cblearn

For reference, a fork of our ordinal embedding toolbox. For contributions, use the original repository.

wichmann-lab/CLIP-imagenet-evaluation

Run CLIP inference on the ImageNet dataset and use these inferences as labels to train other models and again evaluate the trained model on Imagenet validation dataset using original labels or CLIP labels

wichmann-lab/db_sandbox

wichmann-lab/generalisation-humans-DNNs

Data, code & materials from the paper "Generalisation in humans and deep neural networks" (NeurIPS 2018)

wichmann-lab/imagecorruptions

Python package to corrupt arbitrary images.

wichmann-lab/model-vs-human

Benchmark your model on out-of-distribution datasets with carefully collected human comparison data

wichmann-lab/object-recognition

Data and materials from the paper "Comparing deep neural networks against humans: object recognition when the signal gets weaker" (arXiv 2017)

wichmann-lab/robust-detection-benchmark

Code, data and benchmark from the paper "Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming" (arXiv 2019)

wichmann-lab/shortcut-perspective

Figures & code from the paper "Shortcut Learning in Deep Neural Networks" (arXiv 2020)

wichmann-lab/stylize-datasets

A script that applies the AdaIN style transfer method to arbitrary datasets

wichmann-lab/testing_visualizations

wichmann-lab/texture-vs-shape

Pre-trained models, data, code & materials from the paper "ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness" (ICLR 2019 Oral)