Cortical Learning via Prediction

Christos H. Papadimitriou, Santosh S. Vempala
; Proceedings of The 28th Conference on Learning Theory, PMLR 40:1402-1422, 2015.

Abstract

What is the mechanism of learning in the brain? Despite breathtaking advances in neuroscience, and in machine learning, we do not seem close to an answer. Using Valiant’s neuronal model as a foundation, we introduce PJOIN (for “predictive join"), a primitive that combines association and prediction. We show that PJOIN can be implemented naturally in Valiant’s conservative, formal model of cortical computation. Using PJOIN — and almost nothing else — we give a simple algorithm for unsupervised learning of arbitrary ensembles of binary patterns (solving an open problem in Valiant’s work). This algorithm relies crucially on prediction, and entails significant downward traffic (“feedback") while parsing stimuli. Prediction and feedback are well-known features of neural cognition and, as far as we know, this is the first theoretical prediction of their essential role in learning.

Cite this Paper


BibTeX
@InProceedings{pmlr-v40-Papadimitriou15, title = {Cortical Learning via Prediction}, author = {Christos H. Papadimitriou and Santosh S. Vempala}, booktitle = {Proceedings of The 28th Conference on Learning Theory}, pages = {1402--1422}, year = {2015}, editor = {Peter Grünwald and Elad Hazan and Satyen Kale}, volume = {40}, series = {Proceedings of Machine Learning Research}, address = {Paris, France}, month = {03--06 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v40/Papadimitriou15.pdf}, url = {http://proceedings.mlr.press/v40/Papadimitriou15.html}, abstract = {What is the mechanism of learning in the brain? Despite breathtaking advances in neuroscience, and in machine learning, we do not seem close to an answer. Using Valiant’s neuronal model as a foundation, we introduce PJOIN (for “predictive join"), a primitive that combines association and prediction. We show that PJOIN can be implemented naturally in Valiant’s conservative, formal model of cortical computation. Using PJOIN — and almost nothing else — we give a simple algorithm for unsupervised learning of arbitrary ensembles of binary patterns (solving an open problem in Valiant’s work). This algorithm relies crucially on prediction, and entails significant downward traffic (“feedback") while parsing stimuli. Prediction and feedback are well-known features of neural cognition and, as far as we know, this is the first theoretical prediction of their essential role in learning.} }
Endnote
%0 Conference Paper %T Cortical Learning via Prediction %A Christos H. Papadimitriou %A Santosh S. Vempala %B Proceedings of The 28th Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2015 %E Peter Grünwald %E Elad Hazan %E Satyen Kale %F pmlr-v40-Papadimitriou15 %I PMLR %J Proceedings of Machine Learning Research %P 1402--1422 %U http://proceedings.mlr.press %V 40 %W PMLR %X What is the mechanism of learning in the brain? Despite breathtaking advances in neuroscience, and in machine learning, we do not seem close to an answer. Using Valiant’s neuronal model as a foundation, we introduce PJOIN (for “predictive join"), a primitive that combines association and prediction. We show that PJOIN can be implemented naturally in Valiant’s conservative, formal model of cortical computation. Using PJOIN — and almost nothing else — we give a simple algorithm for unsupervised learning of arbitrary ensembles of binary patterns (solving an open problem in Valiant’s work). This algorithm relies crucially on prediction, and entails significant downward traffic (“feedback") while parsing stimuli. Prediction and feedback are well-known features of neural cognition and, as far as we know, this is the first theoretical prediction of their essential role in learning.
RIS
TY - CPAPER TI - Cortical Learning via Prediction AU - Christos H. Papadimitriou AU - Santosh S. Vempala BT - Proceedings of The 28th Conference on Learning Theory PY - 2015/06/26 DA - 2015/06/26 ED - Peter Grünwald ED - Elad Hazan ED - Satyen Kale ID - pmlr-v40-Papadimitriou15 PB - PMLR SP - 1402 DP - PMLR EP - 1422 L1 - http://proceedings.mlr.press/v40/Papadimitriou15.pdf UR - http://proceedings.mlr.press/v40/Papadimitriou15.html AB - What is the mechanism of learning in the brain? Despite breathtaking advances in neuroscience, and in machine learning, we do not seem close to an answer. Using Valiant’s neuronal model as a foundation, we introduce PJOIN (for “predictive join"), a primitive that combines association and prediction. We show that PJOIN can be implemented naturally in Valiant’s conservative, formal model of cortical computation. Using PJOIN — and almost nothing else — we give a simple algorithm for unsupervised learning of arbitrary ensembles of binary patterns (solving an open problem in Valiant’s work). This algorithm relies crucially on prediction, and entails significant downward traffic (“feedback") while parsing stimuli. Prediction and feedback are well-known features of neural cognition and, as far as we know, this is the first theoretical prediction of their essential role in learning. ER -
APA
Papadimitriou, C.H. & Vempala, S.S.. (2015). Cortical Learning via Prediction. Proceedings of The 28th Conference on Learning Theory, in PMLR 40:1402-1422

Related Material