Picture for Behnoosh Zamanlooy

Behnoosh Zamanlooy

Locally Private Sampling with Public Data

Add code
Nov 13, 2024
Viaarxiv icon

Machine Learning-powered Course Allocation

Add code
Oct 03, 2022
Figure 1 for Machine Learning-powered Course Allocation
Figure 2 for Machine Learning-powered Course Allocation
Figure 3 for Machine Learning-powered Course Allocation
Figure 4 for Machine Learning-powered Course Allocation
Viaarxiv icon

Piecewise-Linear Activations or Analytic Activation Functions: Which Produce More Expressive Neural Networks?

Add code
Apr 24, 2022
Figure 1 for Piecewise-Linear Activations or Analytic Activation Functions: Which Produce More Expressive Neural Networks?
Figure 2 for Piecewise-Linear Activations or Analytic Activation Functions: Which Produce More Expressive Neural Networks?
Viaarxiv icon

Universal Approximation Under Constraints is Possible with Transformers

Add code
Oct 07, 2021
Figure 1 for Universal Approximation Under Constraints is Possible with Transformers
Figure 2 for Universal Approximation Under Constraints is Possible with Transformers
Figure 3 for Universal Approximation Under Constraints is Possible with Transformers
Figure 4 for Universal Approximation Under Constraints is Possible with Transformers
Viaarxiv icon

Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes

Add code
Oct 29, 2020
Figure 1 for Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes
Figure 2 for Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes
Figure 3 for Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes
Figure 4 for Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes
Viaarxiv icon

Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time

Add code
Jun 24, 2020
Figure 1 for Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time
Figure 2 for Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time
Figure 3 for Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time
Figure 4 for Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time
Viaarxiv icon