English
Statistical learning in biological neural networks
Multi angle
Auteurs : Schmidt-Hieber, Johannes (Auteur de la Conférence)
CIRM (Editeur )
62J05 - Linear regression
62L20 - Stochastic approximation
CIRM (Editeur )
Loading the player...
|
Résumé :
Compared to artificial neural networks (ANNs), the brain seems to learn faster, generalize better to new situations and consumes much less energy. ANNs are motivated by the functioning of the brain but differ in several crucial aspects. While ANNs are deterministic, biological neural networks (BNNs) are stochastic. Moreover, it is biologically implausible that the learning of the brain is based on gradient descent. In the past years, statistical theory for artificial neural networks has been developed. The idea now is to extend this to biological neural networks, as the future of AI is likely to draw even more inspiration from biology. In this lecture series we will survey the challenges and present some first statistical risk bounds for different biologically inspired learning rules.
Keywords : biological learning; neural networks; statistical estimation rates; supervised learning; zeroth-order methods
Codes MSC : Keywords : biological learning; neural networks; statistical estimation rates; supervised learning; zeroth-order methods
62J05 - Linear regression
62L20 - Stochastic approximation
|
Informations sur la Vidéo
Réalisateur : Recanzone, LucaLangue : Anglais Date de publication : 14/01/2025 Date de captation : 17/12/2024 Sous collection : Research School arXiv category : Machine Learning ; Statistics Theory Domaine : Probability & Statistics Format : MP4 (.mp4) - HD Durée : 00:57:30 Audience : Researchers ; Graduate Students ; Doctoral Students, Post-Doctoral Students Download : https://videos.cirm-math.fr/2024-12-17_Schmidt-Hieber.mp4 
|
Informations sur la Rencontre
Nom de la rencontre : New challenges in high-dimensional statistics / Statistique mathématiqueOrganisateurs de la rencontre : Klopp, Olga ; Pouet, Christophe ; Rakhlin, Alexander Dates : 16/12/2024 - 20/12/2024 Année de la rencontre : 2024 URL Congrès : https://conferences.cirm-math.fr/3055.html
Données de citation
DOI : 10.24350/CIRM.V.20279603Citer cette vidéo: Schmidt-Hieber, Johannes (2024). Statistical learning in biological neural networks. CIRM. Audiovisual resource. doi:10.24350/CIRM.V.20279603 URI : http://dx.doi.org/10.24350/CIRM.V.20279603 |
Voir aussi
- [Multi angle] Are adaptive robust confidence intervals possible? / Auteur de la Conférence Gao, Chao.
- [Multi angle] Attention layers provably solve single-location regression / Auteur de la Conférence Boyer, Claire.
- [Multi angle] An introduction to conformal prediction and distribution-free inference / Auteur de la Conférence Barber, Rina Foygel.
- [Multi angle] From robust tests to robust Bayes-like posterior distribution / Auteur de la Conférence Baraud, Yannick.
Bibliographie
- SCHMIDT-HIEBER, Johannes. Interpreting learning in biological neural networks as zero-order optimization method. arXiv preprint arXiv:2301.11777, 2023. - https://doi.org/10.48550/arXiv.2301.11777
- BOS, Thijs et SCHMIDT-HIEBER, Johannes. Convergence guarantees for forward gradient descent in the linear regression model. Journal of Statistical Planning and Inference, 2024, vol. 233, p. 106174. - https://doi.org/10.1016/j.jspi.2024.106174
- SCHMIDT-HIEBER, Johannes et KOOLEN, Wouter M. Hebbian learning inspired estimation of the linear regression parameters from queries. arXiv preprint arXiv:2311.03483, 2023. - https://doi.org/10.48550/arXiv.2311.03483
- DEXHEIMER, Niklas et SCHMIDT-HIEBER, Johannes. Improving the Convergence Rates of Forward Gradient Descent with Repeated Sampling. arXiv preprint arXiv:2411.17567, 2024. - https://doi.org/10.48550/arXiv.2411.17567
Imagette Video
