no code implementations • 23 Apr 2024 • Saeid Abbassi, Kamaledin Ghiasi-Shirazi, Ahad Harati
Capsule networks are a type of neural network that identify image parts and form the instantiation parameters of a whole hierarchically.
no code implementations • 20 Aug 2021 • Kamyar Nasiri, Kamaledin Ghiasi-Shirazi
One way to increase the capability of the convolution operator is by applying activation functions on the inner product operator.
no code implementations • 26 Feb 2021 • Ahmad Navid Ghanizadeh, Kamaledin Ghiasi-Shirazi, Reza Monsefi, Mohammadreza Qaraei
By this interpretation, we propose a Neural Generalization of Multiple Kernel Learning (NGMKL), which extends the conventional multiple kernel learning framework to a multi-layer neural network with nonlinear activation functions.
no code implementations • 20 Aug 2020 • Ramin Zarei Sabzevar, Kamaledin Ghiasi-Shirazi, Ahad Harati
We propose a novel training algorithm for the $\pm$ED-WTA network, which cleverly switches between updating the positive and negative prototypes and is essential to the emergence of interpretable prototypes.
no code implementations • 11 Mar 2020 • Naeem Paeedeh, Kamaledin Ghiasi-Shirazi
Many attempts took place to improve the adaptive filters that can also be useful to improve backpropagation (BP).
no code implementations • 8 Dec 2017 • Kamaledin Ghiasi-Shirazi
In this paper, we prove that 2D Gabor functions are translation-invariant positive-definite kernels and propose a novel formulation for the problem of image representation with Gabor functions based on infinite kernel learning regression.
no code implementations • 14 Jul 2017 • Kamaledin Ghiasi-Shirazi
The second one is the class of similarity measures defined based on a distance function.
no code implementations • 26 Jun 2016 • Amir Ahooye Atashin, Kamaledin Ghiasi-Shirazi, Ahad Harati
Experimental results on two gesture recognition tasks show that the proposed method outperforms LDCRFs, hidden Markov models, and conditional random fields.
no code implementations • 14 Jun 2016 • Amir Ahooye Atashin, Parsa Bagherzadeh, Kamaledin Ghiasi-Shirazi
In the proposed method, the denoising auto encoders are employed for learning robust features.