The Simplest Neural Model and a Hypothesis for Language @MITCBMM
The Simplest Neural Model and a Hypothesis for Language  @MITCBMM
Uploaded December 2023 | Updated November 2026, 14 hours ago
Daniel Mitropolsky, Columbia University

Abstract: How do neurons, in their collective action, beget cognition, as well as intelligence and reasoning? As Richard Axel recently put it, we do not have a logic for the transformation of neural activity into thought and action; discerning this logic as the most important future direction of neuroscience. I will present a mathematical neural model of brain computation called NEMO, whose key ingredients are spiking neurons, random synapses and weights, local inhibition, and Hebbian plasticity (no backpropagation). Concepts are represented by interconnected co-firing assemblies of neurons that emerge organically from the dynamical system of its equations. It turns out it is possible to carry out complex operations on these concept representations, such as copying, merging, completion from small subsets, and sequence memorization. NEMO is a neuromorphic computational system that, because of its simplifying assumptions, can be efficiently simulated on modern hardware. I will present how to use NEMO to implement an efficient parser of a small but non-trivial subset of English, and a more recent model of the language organ in the baby brain that learns the meaning of words, and basic syntax, from whole sentences with grounded input. In addition to constituting hypotheses as to the logic of the brain, we will discuss how principles from these brain-like models might be used to improve AI, which, despite astounding recent progress, still lags behind humans in several key dimensions such as creativity, hard constraints, energy consumption.

https://cbmm.mit.edu/news-events/events/cbmm-research-meeting-neural-hypothesis-language
The Simplest Neural Model and a Hypothesis for LanguageCell Type-Specific TranscriptomicsWhat are the computations underlying primate vs. machine vision?The Checkershadow IllusionHippocampal memory, cognition, and the role of sleep (part 2)Panel Discussion: Open Questions in Theory of LearningClustering and generalization of abstract structures in reinforcement learningA Theory of Appropriateness with Applications to Generative Artificial IntelligenceTowards Understanding the Implicit Regularization Effect of SGDPhotographic Image Priors in the Era of Machine LearningCBMM10 Panel: Research on Intelligence in the Age of AISprouting a STEM Community in Puerto Rico
MITCBMM |

The Simplest Neural Model and a Hypothesis for Language

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER