「"Blake Richards"」に一致 するユーザー プロフィール
Blake RichardsMila + McGill University |
Mixture-of-Depths: Dynamically allocating compute in transformer-based language models
Transformer-based language models spread FLOPs uniformly across input sequences. In
this work we demonstrate that transformers can instead learn to dynamically allocate FLOPs (…
this work we demonstrate that transformers can instead learn to dynamically allocate FLOPs (…
Sequential predictive learning is a unifying theory for hippocampal representation and replay
The mammalian hippocampus contains a cognitive map that represents an animal’s position
in the environment 1 and generates offline “replay” 2 , 3 for the purposes of recall 4 , …
in the environment 1 and generates offline “replay” 2 , 3 for the purposes of recall 4 , …
Excitability mediates allocation of pre-configured ensembles to a hippocampal engram supporting contextual conditioned threat in mice
AJ Mocle, AI Ramsaran, AD Jacob, AJ Rashid… - Neuron, 2024 - cell.com
Little is understood about how engrams, sparse groups of neurons that store memories, are
formed endogenously. Here, we combined calcium imaging, activity tagging, and …
formed endogenously. Here, we combined calcium imaging, activity tagging, and …
Formalizing locality for normative synaptic plasticity models
In recent years, many researchers have proposed new models for synaptic plasticity in the
brain based on principles of machine learning. The central motivation has been the …
brain based on principles of machine learning. The central motivation has been the …
Learning better with Dale's Law: A Spectral Perspective
Most recurrent neural networks (RNNs) do not include a fundamental constraint of real
neural circuits: Dale's Law, which implies that neurons must be excitatory (E) or inhibitory (I). …
neural circuits: Dale's Law, which implies that neurons must be excitatory (E) or inhibitory (I). …
Responses to pattern-violating visual stimuli evolve differently over days in somata and distal apical dendrites
Scientists have long conjectured that the neocortex learns patterns in sensory data to generate
top-down predictions of upcoming stimuli. In line with this conjecture, different responses …
top-down predictions of upcoming stimuli. In line with this conjecture, different responses …
A unified, scalable framework for neural population decoding
Our ability to use deep learning approaches to decipher neural activity would likely benefit
from greater scale, in terms of both the model size and the datasets. However, the integration …
from greater scale, in terms of both the model size and the datasets. However, the integration …
Engineering protein nanoparticles for drug delivery
BA Richards, AG Goncalves, MO Sullivan… - Current Opinion in …, 2024 - Elsevier
Highlights • Protein nanoparticles are capable of highly uniform multifunctionalization. •
Interior modifications of protein nanoparticles can improve loading capacity. • Exterior …
Interior modifications of protein nanoparticles can improve loading capacity. • Exterior …
Contrastive Retrospection: honing in on critical steps for rapid learning and generalization in RL
In real life, success is often contingent upon multiple critical steps that are distant in time from
each other and from the final reward. These critical steps are challenging to identify with …
each other and from the final reward. These critical steps are challenging to identify with …
Stimulus information guides the emergence of behavior-related signals in primary somatosensory cortex during learning
Neurons in the primary cortex carry sensory- and behavior-related information, but it remains
an open question how this information emerges and intersects together during learning. …
an open question how this information emerges and intersects together during learning. …