Picture for Shiye Lei

Shiye Lei

EarthSynth: Generating Informative Earth Observation with Diffusion Models

Add code
May 17, 2025
Viaarxiv icon

Offline Behavior Distillation

Add code
Oct 30, 2024
Viaarxiv icon

Image Captions are Natural Prompts for Text-to-Image Models

Add code
Jul 17, 2023
Viaarxiv icon

A Comprehensive Survey to Dataset Distillation

Add code
Jan 13, 2023
Viaarxiv icon

Understanding deep learning via decision boundary

Add code
Jun 03, 2022
Figure 1 for Understanding deep learning via decision boundary
Figure 2 for Understanding deep learning via decision boundary
Figure 3 for Understanding deep learning via decision boundary
Figure 4 for Understanding deep learning via decision boundary
Viaarxiv icon

Spatial-Temporal-Fusion BNN: Variational Bayesian Feature Layer

Add code
Dec 12, 2021
Figure 1 for Spatial-Temporal-Fusion BNN: Variational Bayesian Feature Layer
Figure 2 for Spatial-Temporal-Fusion BNN: Variational Bayesian Feature Layer
Figure 3 for Spatial-Temporal-Fusion BNN: Variational Bayesian Feature Layer
Figure 4 for Spatial-Temporal-Fusion BNN: Variational Bayesian Feature Layer
Viaarxiv icon

Spectral Complexity-scaled Generalization Bound of Complex-valued Neural Networks

Add code
Dec 07, 2021
Figure 1 for Spectral Complexity-scaled Generalization Bound of Complex-valued Neural Networks
Figure 2 for Spectral Complexity-scaled Generalization Bound of Complex-valued Neural Networks
Viaarxiv icon

Neural networks behave as hash encoders: An empirical study

Add code
Jan 14, 2021
Figure 1 for Neural networks behave as hash encoders: An empirical study
Figure 2 for Neural networks behave as hash encoders: An empirical study
Figure 3 for Neural networks behave as hash encoders: An empirical study
Figure 4 for Neural networks behave as hash encoders: An empirical study
Viaarxiv icon

A high performance computing method for accelerating temporal action proposal generation

Add code
Aug 02, 2019
Figure 1 for A high performance computing method for accelerating temporal action proposal generation
Figure 2 for A high performance computing method for accelerating temporal action proposal generation
Figure 3 for A high performance computing method for accelerating temporal action proposal generation
Figure 4 for A high performance computing method for accelerating temporal action proposal generation
Viaarxiv icon