Scores on benchmarks
Model rank shown below is with respect to all public models..127 |
average_vision
rank 366
81 benchmarks |
|
.254 |
behavior_vision
rank 152
43 benchmarks |
|
.191 |
Geirhos2021-error_consistency
[reference]
rank 143
17 benchmarks |
|
.300 |
Geirhos2021cueconflict-error_consistency
v1
[reference]
rank 60
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.175 |
Geirhos2021edge-error_consistency
v1
[reference]
rank 45
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.445 |
Geirhos2021eidolonI-error_consistency
v1
[reference]
rank 82
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.463 |
Geirhos2021eidolonII-error_consistency
v1
[reference]
rank 80
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.505 |
Geirhos2021falsecolour-error_consistency
v1
[reference]
rank 62
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.139 |
Geirhos2021highpass-error_consistency
v1
[reference]
rank 53
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.252 |
Geirhos2021lowpass-error_consistency
v1
[reference]
rank 81
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.195 |
Geirhos2021phasescrambling-error_consistency
v1
[reference]
rank 81
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.222 |
Geirhos2021powerequalisation-error_consistency
v1
[reference]
rank 74
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.548 |
Geirhos2021silhouette-error_consistency
v1
[reference]
rank 87
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.519 |
Baker2022
rank 48
3 benchmarks |
|
.524 |
Baker2022fragmented-accuracy_delta
v1
[reference]
rank 80
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.666 |
Baker2022frankenstein-accuracy_delta
v1
[reference]
rank 53
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.367 |
Baker2022inverted-accuracy_delta
v1
[reference]
rank 38
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.460 |
Maniquet2024
rank 122
2 benchmarks |
|
.262 |
Maniquet2024-confusion_similarity
v1
[reference]
rank 154
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.659 |
Maniquet2024-tasks_consistency
v1
[reference]
rank 78
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.291 |
Hebart2023-match
v1
rank 91
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.209 |
BMD2024
rank 57
4 benchmarks |
|
.125 |
BMD2024.dotted_1Behavioral-accuracy_distance
v1
rank 109
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.258 |
BMD2024.texture_1Behavioral-accuracy_distance
v1
rank 45
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.315 |
BMD2024.texture_2Behavioral-accuracy_distance
v1
rank 28
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.137 |
BMD2024.dotted_2Behavioral-accuracy_distance
v1
rank 94
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.360 |
Coggan2024_behavior-ConditionWiseAccuracySimilarity
v1
rank 78
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.342 |
engineering_vision
rank 154
25 benchmarks |
|
.732 |
ImageNet-top1
v1
[reference]
rank 82
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.467 |
ImageNet-C-top1
[reference]
rank 47
4 benchmarks |
|
.444 |
ImageNet-C-noise-top1
v2
[reference]
rank 54
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.357 |
ImageNet-C-blur-top1
v2
[reference]
rank 74
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.501 |
ImageNet-C-weather-top1
v2
[reference]
rank 54
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.566 |
ImageNet-C-digital-top1
v2
[reference]
rank 39
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.415 |
Geirhos2021-top1
[reference]
rank 231
17 benchmarks |
|
.228 |
Geirhos2021cueconflict-top1
v1
[reference]
rank 92
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.269 |
Geirhos2021edge-top1
v1
[reference]
rank 135
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.494 |
Geirhos2021eidolonI-top1
v1
[reference]
rank 144
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.533 |
Geirhos2021eidolonII-top1
v1
[reference]
rank 113
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.975 |
Geirhos2021falsecolour-top1
v1
[reference]
rank 39
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.503 |
Geirhos2021highpass-top1
v1
[reference]
rank 61
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.463 |
Geirhos2021lowpass-top1
v1
[reference]
rank 79
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.627 |
Geirhos2021phasescrambling-top1
v1
[reference]
rank 100
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.804 |
Geirhos2021powerequalisation-top1
v1
[reference]
rank 66
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.519 |
Geirhos2021silhouette-top1
v1
[reference]
rank 106
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.631 |
Geirhos2021sketch-top1
v1
[reference]
rank 93
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.450 |
Geirhos2021stylized-top1
v1
[reference]
rank 61
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.554 |
Geirhos2021uniformnoise-top1
v1
[reference]
rank 69
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
.094 |
Hermann2020
[reference]
rank 271
2 benchmarks |
|
.188 |
Hermann2020cueconflict-shape_match
v1
[reference]
rank 98
|
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
How to use
from brainscore_vision import load_model model = load_model("inception_v3_pytorch") model.start_task(...) model.start_recording(...) model.look_at(...)
Benchmarks bibtex
@article{geirhos2021partial, title={Partial success in closing the gap between human and machine vision}, author={Geirhos, Robert and Narayanappa, Kantharaju and Mitzkus, Benjamin and Thieringer, Tizian and Bethge, Matthias and Wichmann, Felix A and Brendel, Wieland}, journal={Advances in Neural Information Processing Systems}, volume={34}, year={2021}, url={https://openreview.net/forum?id=QkljT4mrfs} } @article{BAKER2022104913, title = {Deep learning models fail to capture the configural nature of human shape perception}, journal = {iScience}, volume = {25}, number = {9}, pages = {104913}, year = {2022}, issn = {2589-0042}, doi = {https://doi.org/10.1016/j.isci.2022.104913}, url = {https://www.sciencedirect.com/science/article/pii/S2589004222011853}, author = {Nicholas Baker and James H. Elder}, keywords = {Biological sciences, Neuroscience, Sensory neuroscience}, abstract = {Summary A hallmark of human object perception is sensitivity to the holistic configuration of the local shape features of an object. Deep convolutional neural networks (DCNNs) are currently the dominant models for object recognition processing in the visual cortex, but do they capture this configural sensitivity? To answer this question, we employed a dataset of animal silhouettes and created a variant of this dataset that disrupts the configuration of each object while preserving local features. While human performance was impacted by this manipulation, DCNN performance was not, indicating insensitivity to object configuration. Modifications to training and architecture to make networks more brain-like did not lead to configural processing, and none of the networks were able to accurately predict trial-by-trial human object judgements. We speculate that to match human configural sensitivity, networks must be trained to solve a broader range of object tasks beyond category recognition.} } @article {Maniquet2024.04.02.587669, author = {Maniquet, Tim and de Beeck, Hans Op and Costantino, Andrea Ivan}, title = {Recurrent issues with deep neural network models of visual recognition}, elocation-id = {2024.04.02.587669}, year = {2024}, doi = {10.1101/2024.04.02.587669}, publisher = {Cold Spring Harbor Laboratory}, URL = {https://www.biorxiv.org/content/early/2024/04/10/2024.04.02.587669}, eprint = {https://www.biorxiv.org/content/early/2024/04/10/2024.04.02.587669.full.pdf}, journal = {bioRxiv} } @INPROCEEDINGS{5206848, author={J. {Deng} and W. {Dong} and R. {Socher} and L. {Li} and {Kai Li} and {Li Fei-Fei}}, booktitle={2009 IEEE Conference on Computer Vision and Pattern Recognition}, title={ImageNet: A large-scale hierarchical image database}, year={2009}, volume={}, number={}, pages={248-255}, } @ARTICLE{Hendrycks2019-di, title = "Benchmarking Neural Network Robustness to Common Corruptions and Perturbations", author = "Hendrycks, Dan and Dietterich, Thomas", abstract = "In this paper we establish rigorous benchmarks for image classifier robustness. Our first benchmark, ImageNet-C, standardizes and expands the corruption robustness topic, while showing which classifiers are preferable in safety-critical applications. Then we propose a new dataset called ImageNet-P which enables researchers to benchmark a classifier's robustness to common perturbations. Unlike recent robustness research, this benchmark evaluates performance on common corruptions and perturbations not worst-case adversarial perturbations. We find that there are negligible changes in relative corruption robustness from AlexNet classifiers to ResNet classifiers. Afterward we discover ways to enhance corruption and perturbation robustness. We even find that a bypassed adversarial defense provides substantial common perturbation robustness. Together our benchmarks may aid future work toward networks that robustly generalize.", month = mar, year = 2019, archivePrefix = "arXiv", primaryClass = "cs.LG", eprint = "1903.12261", url = "https://arxiv.org/abs/1903.12261" } @article{hermann2020origins, title={The origins and prevalence of texture bias in convolutional neural networks}, author={Hermann, Katherine and Chen, Ting and Kornblith, Simon}, journal={Advances in Neural Information Processing Systems}, volume={33}, pages={19000--19015}, year={2020}, url={https://proceedings.neurips.cc/paper/2020/hash/db5f9f42a7157abe65bb145000b5871a-Abstract.html} }