Neural Network Loss Visualization
Plotting its shape helps in understanding the properties and behaviour of a function. Unfortunately since we live in a 3D world, we can’t visualize functions […]
Wrapping a Python Application into a Web Service using mod_wsgi and gunicorn
So you developed a cool AI algorithm and want to show it off through a web service? You know a lot about AI algorithms and […]
Understanding Matrix Multiplication on a Weight-Stationary Systolic Architecture
If you follow the hardware for deep learning space, you may have heard of the term “systolic array”. A 2D systolic array forms the heart […]
Understanding Roofline Charts
The purpose of this post is to discuss my current understanding of roofline charts. Let me lay some background first. Before I got into machine […]
Data Augmentation in SSD (Single Shot Detector)
Over the past few days, I have been investigating how SSD (Single Shot Detector), an object detector introduced in the following paper in Dec 2016 […]
Efficiently Computing the Fisher Vector Product in TRPO
The purpose of this post is to provide math proofs and clarify some implementation details in the recently introduced reinforcement learning method called “Trust Region […]
Initializing Weights for the Convolutional and Fully Connected Layers
You may have noticed that weights for convolutional and fully connected layers in a deep neural network (DNN) are initialized in a specific way. For […]
Object Detection and Classification using R-CNNs
In this post, I’ll describe in detail how R-CNN (Regions with CNN features), a recently introduced deep learning based object detection and classification method works. R-CNN’s […]
Word2Vec Experiments
In this post, I’ll describe some experiments on word2vec, a technique invented by researchers at Google that aims to find compact vector representations for words […]
Understanding Data Parallelism in Machine Learning
In this post, I’ll show how to modify the spiral data set example presented in Karpathy’s post (http://cs231n.github.io/neural-networks-case-study/) to run in a data parallel mode. […]