Alert button
Picture for Lech Szymanski

Lech Szymanski

Alert button

Conceptual capacity and effective complexity of neural networks

Add code
Bookmark button
Alert button
Mar 13, 2021
Lech Szymanski, Brendan McCane, Craig Atkinson

Figure 1 for Conceptual capacity and effective complexity of neural networks
Figure 2 for Conceptual capacity and effective complexity of neural networks
Figure 3 for Conceptual capacity and effective complexity of neural networks
Figure 4 for Conceptual capacity and effective complexity of neural networks
Viaarxiv icon

MIME: Mutual Information Minimisation Exploration

Add code
Bookmark button
Alert button
Jan 16, 2020
Haitao Xu, Brendan McCane, Lech Szymanski, Craig Atkinson

Figure 1 for MIME: Mutual Information Minimisation Exploration
Figure 2 for MIME: Mutual Information Minimisation Exploration
Figure 3 for MIME: Mutual Information Minimisation Exploration
Figure 4 for MIME: Mutual Information Minimisation Exploration
Viaarxiv icon

GRIm-RePR: Prioritising Generating Important Features for Pseudo-Rehearsal

Add code
Bookmark button
Alert button
Nov 27, 2019
Craig Atkinson, Brendan McCane, Lech Szymanski, Anthony Robins

Figure 1 for GRIm-RePR: Prioritising Generating Important Features for Pseudo-Rehearsal
Figure 2 for GRIm-RePR: Prioritising Generating Important Features for Pseudo-Rehearsal
Figure 3 for GRIm-RePR: Prioritising Generating Important Features for Pseudo-Rehearsal
Viaarxiv icon

VASE: Variational Assorted Surprise Exploration for Reinforcement Learning

Add code
Bookmark button
Alert button
Oct 31, 2019
Haitao Xu, Brendan McCane, Lech Szymanski

Figure 1 for VASE: Variational Assorted Surprise Exploration for Reinforcement Learning
Figure 2 for VASE: Variational Assorted Surprise Exploration for Reinforcement Learning
Figure 3 for VASE: Variational Assorted Surprise Exploration for Reinforcement Learning
Figure 4 for VASE: Variational Assorted Surprise Exploration for Reinforcement Learning
Viaarxiv icon

Switched linear projections and inactive state sensitivity for deep neural network interpretability

Add code
Bookmark button
Alert button
Sep 25, 2019
Lech Szymanski, Brendan McCane, Craig Atkinson

Figure 1 for Switched linear projections and inactive state sensitivity for deep neural network interpretability
Figure 2 for Switched linear projections and inactive state sensitivity for deep neural network interpretability
Figure 3 for Switched linear projections and inactive state sensitivity for deep neural network interpretability
Figure 4 for Switched linear projections and inactive state sensitivity for deep neural network interpretability
Viaarxiv icon

Pseudo-Rehearsal: Achieving Deep Reinforcement Learning without Catastrophic Forgetting

Add code
Bookmark button
Alert button
Dec 06, 2018
Craig Atkinson, Brendan McCane, Lech Szymanski, Anthony Robins

Figure 1 for Pseudo-Rehearsal: Achieving Deep Reinforcement Learning without Catastrophic Forgetting
Figure 2 for Pseudo-Rehearsal: Achieving Deep Reinforcement Learning without Catastrophic Forgetting
Figure 3 for Pseudo-Rehearsal: Achieving Deep Reinforcement Learning without Catastrophic Forgetting
Figure 4 for Pseudo-Rehearsal: Achieving Deep Reinforcement Learning without Catastrophic Forgetting
Viaarxiv icon

The effect of the choice of neural network depth and breadth on the size of its hypothesis space

Add code
Bookmark button
Alert button
Jun 06, 2018
Lech Szymanski, Brendan McCane, Michael Albert

Figure 1 for The effect of the choice of neural network depth and breadth on the size of its hypothesis space
Figure 2 for The effect of the choice of neural network depth and breadth on the size of its hypothesis space
Figure 3 for The effect of the choice of neural network depth and breadth on the size of its hypothesis space
Figure 4 for The effect of the choice of neural network depth and breadth on the size of its hypothesis space
Viaarxiv icon

Pseudo-Recursal: Solving the Catastrophic Forgetting Problem in Deep Neural Networks

Add code
Bookmark button
Alert button
May 07, 2018
Craig Atkinson, Brendan McCane, Lech Szymanski, Anthony Robins

Figure 1 for Pseudo-Recursal: Solving the Catastrophic Forgetting Problem in Deep Neural Networks
Figure 2 for Pseudo-Recursal: Solving the Catastrophic Forgetting Problem in Deep Neural Networks
Figure 3 for Pseudo-Recursal: Solving the Catastrophic Forgetting Problem in Deep Neural Networks
Figure 4 for Pseudo-Recursal: Solving the Catastrophic Forgetting Problem in Deep Neural Networks
Viaarxiv icon

Some Approximation Bounds for Deep Networks

Add code
Bookmark button
Alert button
Mar 08, 2018
Brendan McCane, Lech Szymanski

Figure 1 for Some Approximation Bounds for Deep Networks
Figure 2 for Some Approximation Bounds for Deep Networks
Figure 3 for Some Approximation Bounds for Deep Networks
Figure 4 for Some Approximation Bounds for Deep Networks
Viaarxiv icon

Effects of the optimisation of the margin distribution on generalisation in deep architectures

Add code
Bookmark button
Alert button
Apr 19, 2017
Lech Szymanski, Brendan McCane, Wei Gao, Zhi-Hua Zhou

Figure 1 for Effects of the optimisation of the margin distribution on generalisation in deep architectures
Figure 2 for Effects of the optimisation of the margin distribution on generalisation in deep architectures
Figure 3 for Effects of the optimisation of the margin distribution on generalisation in deep architectures
Figure 4 for Effects of the optimisation of the margin distribution on generalisation in deep architectures
Viaarxiv icon