Deep learning
Neural networks from the ground up: layers, optimisers, CNNs, RNNs, GANs and training tricks.
Page 3 of 10
- How Autoencoders Work for Dimension Reduction9 min
An autoencoder is a neural network architecture that learns to compress data into a lower-dimensional representation and then reconstruct it. The network...
- Introducing Adaptive Neural Networks with Stochastic Synapses in Python12 min
Adaptive Neural Networks with Stochastic Synapses are an advanced form of artificial neural networks that incorporate randomness into their synaptic...
- Introducing Geometric Hypergraph Neural Networks in Python10 min
Geometric Hypergraph Neural Networks (GHNNs) are a novel class of neural networks designed to process complex relational data represented as hypergraphs....
- Introduction to Autoencoders and Variational Autoencoders in Python:10 min
Autoencoders are neural networks designed to learn efficient data representations in an unsupervised manner. They compress input data into a...
- Neural Networks for Solving Differential Equations in Python10 min
Neural networks have emerged as a powerful tool for solving differential equations, offering a data-driven approach that complements traditional numerical...
- Physics Informed Neural Networks with Python11 min
Physics Informed Neural Networks (PINNs) are a novel approach that combines the power of neural networks with the fundamental laws of physics. They aim to...
- Physics-Informed Neural Networks for Damped Oscillations10 min
Physics-Informed Neural Networks are a novel approach that combines the power of neural networks with physical laws. They are particularly useful for solving...
- Physics-Informed Neural Networks for Quantum Wave Functions14 min
Physics-Informed Neural Networks (PINNs) are a powerful tool that combines the flexibility of neural networks with the constraints of physical laws. They are...
- Poisson Variational Autoencoder Modeling Count Data with Python6 min
Introduction to Poisson Variational Autoencoder
- Popular Deep Learning Architectures for Recommender Systems14 min
The Wide and Deep architecture combines a wide linear model with a deep neural network to capture both memorization and generalization in recommender...
- Temporal Quantum-Inspired Neural Networks in Python10 min
Temporal Quantum-Inspired Neural Networks (TQINNs) are an innovative approach to neural network architecture that draws inspiration from quantum mechanics...
- Uncertainty Quantification for Neural Networks with Python and PyTorch Lightning5 min
Uncertainty quantification (UQ) is the process of quantifying and characterizing the uncertainties associated with computational models, input data, and...
- Unexpected Benefits of Self-Modeling in Neural Networks14 min
Self-modeling in neural systems refers to the ability of neural networks to create internal representations of their own structure and functionality. This...
- Variational Autoencoders A Python-Powered Generative Modeling Approach8 min
Variational Autoencoders (VAEs) are powerful generative models that combine ideas from deep learning and probabilistic inference. They learn to encode data...
- Variational Autoencoders Generative Models and Applications7 min
A Variational Autoencoder combines probabilistic inference with neural networks, creating a powerful generative model. The architecture consists of an...
- Variational Autoencoders Generative Models with Deep Learning9 min
A Variational Autoencoder combines probabilistic inference with neural networks, creating a powerful generative model. The architecture consists of an...
- Variational Lossy Autoencoder in Python:11 min
Variational Lossy Autoencoders (VLAEs) are an extension of traditional autoencoders, designed to learn efficient data representations while allowing for...
- Vector Quantized Variational Auto-Encoder (VQ-VAE) using Python10 min
Vector Quantized Variational Auto-Encoder (VQ-VAE) is a powerful neural network architecture for learning discrete representations of data. It combines the...
- 2-Stage Backpropagation in Python5 min
2-Stage Backpropagation is an advanced technique for training neural networks that aims to improve convergence and generalization. It involves splitting the...
- AdaGrad Optimization in Deep Learning with Python17 min
AdaGrad (Adaptive Gradient Algorithm) is an optimization algorithm used in deep learning to automatically adapt the learning rate for each parameter. It's...
- Adaptive Learning Rate Optimization Algorithms8 min
Adaptive learning rate optimization algorithms automatically adjust the learning rate parameter during training to achieve better convergence. These methods...
- AdEMAMix Optimizer! Combining Adam and EMA for Faster Deep Learning Convergence11 min
AdEMAMix is an optimization algorithm that combines the benefits of Adaptive Moment Estimation (Adam) and Exponential Moving Average (EMA). It aims to...
- Backpropagation Algorithm in Neural Networks10 min
Backpropagation is a fundamental algorithm used in training neural networks. It's an efficient method for calculating the gradient of the loss function with...
- Bias in Stochastic Gradient Descent for Neural Network Architectures11 min
Stochastic Gradient Descent is a fundamental optimization algorithm in machine learning. While it's highly effective, it can introduce bias in the training...
- Building Stochastic Gradient Descent from Scratch in Python8 min
Stochastic Gradient Descent is a fundamental optimization algorithm used in machine learning to minimize the loss function. It's an iterative method that...
- Chain Rule in Neural Network Backpropagation16 min
The chain rule is a fundamental concept in calculus that plays a crucial role in the backpropagation algorithm used to train neural networks. It allows us to...
- Efficient Optimizers for Neural Network Training11 min
AdamW and Adam are popular optimization algorithms for training neural networks. These algorithms adapt learning rates for each parameter, which can lead to...
- Explaining the Vanishing Gradient Problem in Neural Networks with Python11 min
The vanishing gradient problem is a significant challenge in training deep neural networks. It occurs when the gradients of the loss function approach zero...
- Exploring Learning Rate in Deep Learning with Python10 min
Learning rate is a crucial hyperparameter in deep learning that determines the step size at each iteration while moving toward a minimum of the loss...
- Exploring the Advantages of ADAM Optimizer for Deep Learning8 min
The ADAM optimizer combines momentum and RMSprop approaches to provide adaptive learning rates for each parameter. It maintains both first and second moment...