How connectivity structure shapes rich and lazy learning in neural circuits

YH Liu, A Baratin, J Cornford, S Mihalas… - arXiv preprint arXiv …, 2023 - arxiv.org
In theoretical neuroscience, recent work leverages deep learning tools to explore how some
network attributes critically influence its learning dynamics. Notably, initial weight …

Evolutionary algorithms as an alternative to backpropagation for supervised training of Biophysical Neural Networks and Neural ODEs

J Hazelden, YH Liu, E Shlizerman… - arXiv preprint arXiv …, 2023 - arxiv.org
Training networks consisting of biophysically accurate neuron models could allow for new
insights into how brain circuits can organize and solve tasks. We begin by analyzing the …

[HTML][HTML] Transition to chaos separates learning regimes and relates to measure of consciousness in recurrent neural networks

D Mastrovito, YH Liu, L Kusmierz, E Shea-Brown… - bioRxiv, 2024 - ncbi.nlm.nih.gov
Recurrent neural networks exhibit chaotic dynamics when the variance in their connection
strengths exceed a critical value. Recent work indicates connection variance also modulates …

How Initial Connectivity Shapes Biologically Plausible Learning in Recurrent Neural Networks

W Liu, X Zhang, YH Liu - arXiv preprint arXiv:2410.11164, 2024 - arxiv.org
The impact of initial connectivity on learning has been extensively studied in the context of
backpropagation-based gradient descent, but it remains largely underexplored in …

Does online gradient descent (and variants) still work with biased gradient and variance?

A Al-Tawaha, M Jin - 2024 American Control Conference (ACC), 2024 - ieeexplore.ieee.org
Deterministic bias and stochastic unbiased noise in gradients can affect the performance of
online learning algorithms. While existing studies provide bounds for dynamic regret under …

Manifold Regularization for Memory-Efficient Training of Deep Neural Networks

S Sartipi, EA Bernal - arXiv preprint arXiv:2305.17119, 2023 - arxiv.org
One of the prevailing trends in the machine-and deep-learning community is to gravitate
towards the use of increasingly larger models in order to keep pushing the state-of-the-art …

Deep learning frameworks for modeling how neural circuits learn

YH Liu - 2024 - search.proquest.com
The brain's prowess in learning and adapting remains an enigma, particularly in its
approach to the'temporal credit assignment'problem. How do neural circuits determine …