Michael Poli

papers

952

total citations

papers (15)

Neural Hybrid Automata: Learning Dynamics With Multiple Modes and Stochastic Transitions

NEURIPS 2021arXiv

citations

STAR: Synthesis of Tailored Architectures

ICLR 2025arXiv

citations

Learning Efficient Surrogate Dynamic Models with Graph Spline Networks

NEURIPS 2023arXiv

citations

Exploring Diffusion Transformer Designs via Grafting

NEURIPS 2025arXiv

citations

Self-Similarity Priors: Neural Collages as Differentiable Fractal Representations

NEURIPS 2022arXiv

citations

Quantifying Memory Utilization with Effective State-Size

ICML 2025arXiv

citations

State-Free Inference of State-Space Models: The Transfer Function Approach

ICML 2024

citations

Michael Poli

papers (15)

HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution

Dissecting Neural ODEs

Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture

Mechanistic Design and Scaling of Hybrid Architectures

Hypersolvers: Toward Fast Continuous-Depth Models

Laughing Hyena Distillery: Extracting Compact Recurrences From Convolutions

Transform Once: Efficient Operator Learning in Frequency Domain

Differentiable Multiple Shooting Layers

Neural Hybrid Automata: Learning Dynamics With Multiple Modes and Stochastic Transitions

STAR: Synthesis of Tailored Architectures

Learning Efficient Surrogate Dynamic Models with Graph Spline Networks

Exploring Diffusion Transformer Designs via Grafting

Self-Similarity Priors: Neural Collages as Differentiable Fractal Representations

Quantifying Memory Utilization with Effective State-Size

State-Free Inference of State-Space Models: The *Transfer Function* Approach

papers (15)

HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution

Dissecting Neural ODEs

Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture

Mechanistic Design and Scaling of Hybrid Architectures

Hypersolvers: Toward Fast Continuous-Depth Models

Laughing Hyena Distillery: Extracting Compact Recurrences From Convolutions

Transform Once: Efficient Operator Learning in Frequency Domain

Differentiable Multiple Shooting Layers

Neural Hybrid Automata: Learning Dynamics With Multiple Modes and Stochastic Transitions

STAR: Synthesis of Tailored Architectures

Learning Efficient Surrogate Dynamic Models with Graph Spline Networks

Exploring Diffusion Transformer Designs via Grafting

Self-Similarity Priors: Neural Collages as Differentiable Fractal Representations

Quantifying Memory Utilization with Effective State-Size

State-Free Inference of State-Space Models: The *Transfer Function* Approach

State-Free Inference of State-Space Models: The Transfer Function Approach

State-Free Inference of State-Space Models: The Transfer Function Approach