Picture for Jared Tanner

Jared Tanner

Mind the Gap: a Spectral Analysis of Rank Collapse and Signal Propagation in Transformers

Add code
Oct 10, 2024
Viaarxiv icon

Deep Neural Network Initialization with Sparsity Inducing Activations

Add code
Feb 25, 2024
Viaarxiv icon

Beyond IID weights: sparse and low-rank deep Neural Networks are also Gaussian Processes

Add code
Oct 25, 2023
Viaarxiv icon

Dynamic Sparse No Training: Training-Free Fine-tuning for Sparse LLMs

Add code
Oct 17, 2023
Viaarxiv icon

On the Initialisation of Wide Low-Rank Feedforward Neural Networks

Add code
Jan 31, 2023
Viaarxiv icon

Optimal Approximation Complexity of High-Dimensional Functions with Neural Networks

Add code
Jan 30, 2023
Figure 1 for Optimal Approximation Complexity of High-Dimensional Functions with Neural Networks
Viaarxiv icon

Improved Projection Learning for Lower Dimensional Feature Maps

Add code
Oct 27, 2022
Viaarxiv icon

Tuning-free multi-coil compressed sensing MRI with Parallel Variable Density Approximate Message Passing

Add code
Mar 08, 2022
Figure 1 for Tuning-free multi-coil compressed sensing MRI with Parallel Variable Density Approximate Message Passing
Figure 2 for Tuning-free multi-coil compressed sensing MRI with Parallel Variable Density Approximate Message Passing
Figure 3 for Tuning-free multi-coil compressed sensing MRI with Parallel Variable Density Approximate Message Passing
Figure 4 for Tuning-free multi-coil compressed sensing MRI with Parallel Variable Density Approximate Message Passing
Viaarxiv icon

Activation function design for deep networks: linearity and effective initialisation

Add code
May 17, 2021
Figure 1 for Activation function design for deep networks: linearity and effective initialisation
Figure 2 for Activation function design for deep networks: linearity and effective initialisation
Figure 3 for Activation function design for deep networks: linearity and effective initialisation
Figure 4 for Activation function design for deep networks: linearity and effective initialisation
Viaarxiv icon

Dense for the Price of Sparse: Improved Performance of Sparsely Initialized Networks via a Subspace Offset

Add code
Feb 12, 2021
Figure 1 for Dense for the Price of Sparse: Improved Performance of Sparsely Initialized Networks via a Subspace Offset
Figure 2 for Dense for the Price of Sparse: Improved Performance of Sparsely Initialized Networks via a Subspace Offset
Figure 3 for Dense for the Price of Sparse: Improved Performance of Sparsely Initialized Networks via a Subspace Offset
Figure 4 for Dense for the Price of Sparse: Improved Performance of Sparsely Initialized Networks via a Subspace Offset
Viaarxiv icon