Research

Paper

AI LLM March 02, 2026

CHLU: The Causal Hamiltonian Learning Unit as a Symplectic Primitive for Deep Learning

Authors

Pratik Jawahar, Maurizio Pierini

Abstract

Current deep learning primitives dealing with temporal dynamics suffer from a fundamental dichotomy: they are either discrete and unstable (LSTMs) \citep{pascanu_difficulty_2013}, leading to exploding or vanishing gradients; or they are continuous and dissipative (Neural ODEs) \citep{dupont_augmented_2019}, which destroy information over time to ensure stability. We propose the \textbf{Causal Hamiltonian Learning Unit} (pronounced: \textit{clue}), a novel Physics-grounded computational learning primitive. By enforcing a Relativistic Hamiltonian structure and utilizing symplectic integration, a CHLU strictly conserves phase-space volume, as an attempt to solve the memory-stability trade-off. We show that the CHLU is designed for infinite-horizon stability, as well as controllable noise filtering. We then demonstrate a CHLU's generative ability using the MNIST dataset as a proof-of-principle.

Metadata

arXiv ID: 2603.01768
Provider: ARXIV
Primary Category: cs.LG
Published: 2026-03-02
Fetched: 2026-03-03 04:34

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.01768v1</id>\n    <title>CHLU: The Causal Hamiltonian Learning Unit as a Symplectic Primitive for Deep Learning</title>\n    <updated>2026-03-02T11:53:09Z</updated>\n    <link href='https://arxiv.org/abs/2603.01768v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.01768v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Current deep learning primitives dealing with temporal dynamics suffer from a fundamental dichotomy: they are either discrete and unstable (LSTMs) \\citep{pascanu_difficulty_2013}, leading to exploding or vanishing gradients; or they are continuous and dissipative (Neural ODEs) \\citep{dupont_augmented_2019}, which destroy information over time to ensure stability. We propose the \\textbf{Causal Hamiltonian Learning Unit} (pronounced: \\textit{clue}), a novel Physics-grounded computational learning primitive. By enforcing a Relativistic Hamiltonian structure and utilizing symplectic integration, a CHLU strictly conserves phase-space volume, as an attempt to solve the memory-stability trade-off. We show that the CHLU is designed for infinite-horizon stability, as well as controllable noise filtering. We then demonstrate a CHLU's generative ability using the MNIST dataset as a proof-of-principle.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='physics.app-ph'/>\n    <published>2026-03-02T11:53:09Z</published>\n    <arxiv:comment>Accepted as a short paper at ICLR 2026 (AI &amp; PDE)</arxiv:comment>\n    <arxiv:primary_category term='cs.LG'/>\n    <author>\n      <name>Pratik Jawahar</name>\n    </author>\n    <author>\n      <name>Maurizio Pierini</name>\n    </author>\n  </entry>"
}