showing only pocket [see all]
Case Study: A world class image classifier for dogs and cats (err.., anything)
It is amazing how far computer vision has come in the last couple of years. Problems that are insanely intractable for classical machine learning methods are a piece of cake for the emerging field of…
deep-learning  convolutions  convolutional-neural-networks  neural-networks  differential-learning-rates  learning-rate  kaggle  fast.ai  transfer-learning  from pocket
february 2018
Decoding the ResNet architecture // teleported.in
A blog where I share my intuitions about artificial intelligence, machine learning, deep learning.
resnet  shortcut-connections  network-architecture  convolutional-neural-networks  cnns  deep-learning  fast.ai  from pocket
february 2018
Yet Another ResNet Tutorial (or not) – Apil Tamang – Medium
The purpose of this article is to expose the most fundamental concept driving the design and success of ResNet architectures. Many blogs and articles go on and on describing how this architecture is…
ResNet  neural-networks  network-architecture  from pocket
february 2018
Improving the way we work with learning rate. – techburst
Most optimization algorithms(such as SGD, RMSprop, Adam) require setting the learning rate — the most important hyper-parameter for training deep neural networks. Naive method for choosing learning…
deep-learning  learning-rate  cyclical-learning-rate  fast.ai  learning-rate-annealing  from pocket
february 2018
The Cyclical Learning Rate technique // teleported.in
Learning rate (LR) is one of the most important hyperparameters to be tuned and holds key to faster and effective training of neural networks. Simply put, LR decides how much of the loss gradient is to be applied to our current weights to move them in the direction of lower loss.
Cyclical-Learning-Rate  Learning-Rate  fast.ai  SGDR  from pocket
february 2018
Batch normalization in Neural Networks
This article explains batch normalization in a simple way. I wrote this article after what I learned from Fast.ai and deeplearning.ai. I will start with why we need it, how it works, then how to…
batch-normalization  transfer-learning  deep-learning  neural-networks  fast.ai  from pocket
february 2018
Designing great data products - O'Reilly Media
The Drivetrain Approach for optimization.
Step 1: Define the objective (and metric)
Step 2: Identify levers (what inputs can we control)
Step 3: Collect data
Step 4: Model (how the levers influence the objective)
Step 5: Simulate (see how the levers affect the distribution of the objective)
Step 6: Optimize (choose the possible outcome that best meets the objective)
data-science  mental-models  Drivetrain-Approach  from pocket
february 2018
absolute-error adam arxiv batch-normalization batch-size big-data bilingual-word-embeddings bitcoin blockchain blogs boostrapping career class-imbalance classification cloud-computing cnns confidence-intervals convolutional-neural-networks convolutions correlation cryptocurrencies cryptonetworks cuda cyclical-learning-rate data-augmentation data-science datascience decentralization decentralized-networks decision-trees deep-learning design-principles differential-learning-rates distributed-computing double-exponential-smoothing drivetrain-approach dropout ds eda embeddings ensembles error exploratory-data-analysis exponential-smoothing exponentials fast.ai feature-importance finance forecasting gartner-hype-cycles generalization gini-importance gradient-boosting holt-winters-forecasting human-biases human-computer-interaction hyperparameter-optimization hypothesis-testing incentives individual-conditional-expectation instapaper interpretability interpretable interpretation jobs kaggle learning learning-rate learning-rate-annealing learning-rate-finder libraries linear-regression logarithm logs long-short-term-memory_networks machine-learning math matplotlib memex mental-models model-interpretation modular-network network-architecture neural-networks nlp normal-distribution oob-error ordinary-least-squares organization oversampling packages pandas partial-dependence prioritization probability programming python random-forest rectified-linear-unit recurrent-neural-networks reinforcement-learning resnet rnns s-curve scikit-learn seasonality sgd sgdr shared-representation shortcut-connections simple-exponential-smoothing simple-moving-average sklearn smote squared-error statistics t-sne t-tests test-time-augmentation time-series-data transfer-learning tree-interpreter triple-exponential-smoothing undersampling validation visualization waterfall-charts word-embeddings

Copy this bookmark: