Retrospective on the SENSORIUM 2022 competition

Konstantin F. Willeke, Paul G. Fahey, Mohammad Bashiri, Laura Hansel, Christoph Blessing, Konstantin-Klemens Lurz, Max F. Burg, Santiago A. Cadena, Zhiwei Ding, Kayla Ponder, Taliah Muhammad, Saumil S. Patel, Kaiwen Deng, Yuanfang Guan, Yiqin Zhu, Kaiwen Xiao, Xiao Han, Simone Azeglio, Ulisse Ferrari, Peter Neri, Olivier Marre, Adrian Hoffmann, Kirill Fedyanin, Kirill Vishniakov, Maxim Panov, Subash Prakash, Kishan Naik, Kantharaju Narayanappa, Alexander S. Ecker, Andreas S. Tolias, Fabian H. Sinz
Proceedings of the NeurIPS 2022 Competitions Track, PMLR 220:314-333, 2022.

Abstract

The neural underpinning of the biological visual system is challenging to study experimentally, in particular as neuronal activity becomes increasingly nonlinear with respect to visual input. Artificial neural networks (ANNs) can serve a variety of goals for improving our understanding of this complex system, not only serving as predictive digital twins of sensory cortex for novel hypothesis generation in silico, but also incorporating bio-inspired architectural motifs to progressively bridge the gap between biological and machine vision. The mouse has recently emerged as a popular model system to study visual information processing, but no standardized large-scale benchmark to identify state-of-the-art models of the mouse visual system has been established. To fill this gap, we proposed the SENSORIUM benchmark competition. We collected a large-scale dataset from mouse primary visual cortex containing the responses of more than 28,000 neurons across seven mice stimulated with thousands of natural images, together with simultaneous behavioral measurements that include running speed, pupil dilation, and eye movements. The benchmark challenge ranked models based on predictive performance for neuronal responses on a held-out test set, and included two tracks for model input limited to either stimulus only (SENSORIUM) or stimulus plus behavior (SENSORIUM+). As a part of the NeurIPS 2022 competition track, we received 172 model submissions from 26 teams, with the winning teams improving our previous state-of-the-art model by more than 15 percent. Dataset access and infrastructure for evaluation of model predictions will remain online as an ongoing benchmark. We would like to see this as a starting point for regular challenges and data releases, and as a standard tool for measuring progress in large-scale neural system identification models of the mouse visual system and beyond.

Cite this Paper

BibTeX
@InProceedings{pmlr-v220-willeke23a, title = {Retrospective on the SENSORIUM 2022 competition}, author = {Willeke, Konstantin F. and Fahey, Paul G. and Bashiri, Mohammad and Hansel, Laura and Blessing, Christoph and Lurz, Konstantin-Klemens and Burg, Max F. and Cadena, Santiago A. and Ding, Zhiwei and Ponder, Kayla and Muhammad, Taliah and Patel, Saumil S. and Deng, Kaiwen and Guan, Yuanfang and Zhu, Yiqin and Xiao, Kaiwen and Han, Xiao and Azeglio, Simone and Ferrari, Ulisse and Neri, Peter and Marre, Olivier and Hoffmann, Adrian and Fedyanin, Kirill and Vishniakov, Kirill and Panov, Maxim and Prakash, Subash and Naik, Kishan and Narayanappa, Kantharaju and Ecker, Alexander S. and Tolias, Andreas S. and Sinz, Fabian H.}, booktitle = {Proceedings of the NeurIPS 2022 Competitions Track}, pages = {314--333}, year = {2022}, editor = {Ciccone, Marco and Stolovitzky, Gustavo and Albrecht, Jacob}, volume = {220}, series = {Proceedings of Machine Learning Research}, month = {28 Nov--09 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v220/willeke23a/willeke23a.pdf}, url = {https://proceedings.mlr.press/v220/willeke23a.html}, abstract = {The neural underpinning of the biological visual system is challenging to study experimentally, in particular as neuronal activity becomes increasingly nonlinear with respect to visual input. Artificial neural networks (ANNs) can serve a variety of goals for improving our understanding of this complex system, not only serving as predictive digital twins of sensory cortex for novel hypothesis generation in silico, but also incorporating bio-inspired architectural motifs to progressively bridge the gap between biological and machine vision. The mouse has recently emerged as a popular model system to study visual information processing, but no standardized large-scale benchmark to identify state-of-the-art models of the mouse visual system has been established. To fill this gap, we proposed the SENSORIUM benchmark competition. We collected a large-scale dataset from mouse primary visual cortex containing the responses of more than 28,000 neurons across seven mice stimulated with thousands of natural images, together with simultaneous behavioral measurements that include running speed, pupil dilation, and eye movements. The benchmark challenge ranked models based on predictive performance for neuronal responses on a held-out test set, and included two tracks for model input limited to either stimulus only (SENSORIUM) or stimulus plus behavior (SENSORIUM+). As a part of the NeurIPS 2022 competition track, we received 172 model submissions from 26 teams, with the winning teams improving our previous state-of-the-art model by more than 15 percent. Dataset access and infrastructure for evaluation of model predictions will remain online as an ongoing benchmark. We would like to see this as a starting point for regular challenges and data releases, and as a standard tool for measuring progress in large-scale neural system identification models of the mouse visual system and beyond.} }
Endnote
%0 Conference Paper %T Retrospective on the SENSORIUM 2022 competition %A Konstantin F. Willeke %A Paul G. Fahey %A Mohammad Bashiri %A Laura Hansel %A Christoph Blessing %A Konstantin-Klemens Lurz %A Max F. Burg %A Santiago A. Cadena %A Zhiwei Ding %A Kayla Ponder %A Taliah Muhammad %A Saumil S. Patel %A Kaiwen Deng %A Yuanfang Guan %A Yiqin Zhu %A Kaiwen Xiao %A Xiao Han %A Simone Azeglio %A Ulisse Ferrari %A Peter Neri %A Olivier Marre %A Adrian Hoffmann %A Kirill Fedyanin %A Kirill Vishniakov %A Maxim Panov %A Subash Prakash %A Kishan Naik %A Kantharaju Narayanappa %A Alexander S. Ecker %A Andreas S. Tolias %A Fabian H. Sinz %B Proceedings of the NeurIPS 2022 Competitions Track %C Proceedings of Machine Learning Research %D 2022 %E Marco Ciccone %E Gustavo Stolovitzky %E Jacob Albrecht %F pmlr-v220-willeke23a %I PMLR %P 314--333 %U https://proceedings.mlr.press/v220/willeke23a.html %V 220 %X The neural underpinning of the biological visual system is challenging to study experimentally, in particular as neuronal activity becomes increasingly nonlinear with respect to visual input. Artificial neural networks (ANNs) can serve a variety of goals for improving our understanding of this complex system, not only serving as predictive digital twins of sensory cortex for novel hypothesis generation in silico, but also incorporating bio-inspired architectural motifs to progressively bridge the gap between biological and machine vision. The mouse has recently emerged as a popular model system to study visual information processing, but no standardized large-scale benchmark to identify state-of-the-art models of the mouse visual system has been established. To fill this gap, we proposed the SENSORIUM benchmark competition. We collected a large-scale dataset from mouse primary visual cortex containing the responses of more than 28,000 neurons across seven mice stimulated with thousands of natural images, together with simultaneous behavioral measurements that include running speed, pupil dilation, and eye movements. The benchmark challenge ranked models based on predictive performance for neuronal responses on a held-out test set, and included two tracks for model input limited to either stimulus only (SENSORIUM) or stimulus plus behavior (SENSORIUM+). As a part of the NeurIPS 2022 competition track, we received 172 model submissions from 26 teams, with the winning teams improving our previous state-of-the-art model by more than 15 percent. Dataset access and infrastructure for evaluation of model predictions will remain online as an ongoing benchmark. We would like to see this as a starting point for regular challenges and data releases, and as a standard tool for measuring progress in large-scale neural system identification models of the mouse visual system and beyond.
APA
Willeke, K.F., Fahey, P.G., Bashiri, M., Hansel, L., Blessing, C., Lurz, K., Burg, M.F., Cadena, S.A., Ding, Z., Ponder, K., Muhammad, T., Patel, S.S., Deng, K., Guan, Y., Zhu, Y., Xiao, K., Han, X., Azeglio, S., Ferrari, U., Neri, P., Marre, O., Hoffmann, A., Fedyanin, K., Vishniakov, K., Panov, M., Prakash, S., Naik, K., Narayanappa, K., Ecker, A.S., Tolias, A.S. & Sinz, F.H.. (2022). Retrospective on the SENSORIUM 2022 competition. Proceedings of the NeurIPS 2022 Competitions Track, in Proceedings of Machine Learning Research 220:314-333 Available from https://proceedings.mlr.press/v220/willeke23a.html.

Related Material