Jeremy Jordan

The attention mechanism allows us to merge a variable-length sequence of vectors into a fixed-size context vector. What if we could use this mechanism to entirely replace recurrence for sequential modeling? This blog post covers the Transformer architecture which explores such an approach.

In this blog post, we'll discuss a key innovation in sequence-to-sequence model architectures: the attention mechanism. This architecture innovation dramatically improved model performance for sequence-to-sequence tasks such as machine translation and text summarization. Moreover, the success of this attention mechanism led to the seminal

Let's say you want to deploy a recommender system at your company. A typical architecture might include a set of inference servers to run your embedding and ranking models, an approximate nearest neighbor index to select a set of candidate items that match your query, a database to retrieve features

This page contains a quick reference for writing Terraform configuration.

This blog post aims to provide a simple, open-source solution for monitoring ML systems. We'll discuss industry-standard monitoring tools and practices for software systems and how they can be adapted to monitor ML systems.

This blog post will provide an introduction to Kubernetes so that you can understand the motivation behind the tool, what it is, and how you can use it. In a follow-up post, I'll discuss how we can leverage Kubernetes to power data science workloads using more concrete (data science) examples.

Previously, I wrote about organizing machine learning projects where I presented the framework that I use for building and deploying models. However, that framework operates on the implicit assumption that you already know generally what your model should do.

In this post, I'll discuss a third type of neural networks, recurrent neural networks, for learning from sequential data. For some classes of data, the order in which we receive observations is important. As an example, consider the two following sentences:

In this blog post, I'll cover a couple of techniques used for approximate nearest neighbors search. This post will not cover approximate nearest neighbors methods exhaustively, but hopefully you'll be able to understand how people generally approach this problem and how to apply these techniques

The goal of this document is to provide a common framework for approaching machine learning projects that can be referenced by practitioners. If you build ML models, this post is for you.

In this post, I'll discuss an overview of deep learning techniques for object detection using convolutional neural networks. Object detection is useful for understanding what's in an image, describing both what is in an image and where those objects are found.

When evaluating a standard machine learning model, we usually classify our predictions into four categories: true positives, false positives, true negatives, and false negatives. However, for the dense prediction task of image segmentation, it's not immediately clear what counts as a "true positive" and,

In this post, I'll discuss how to use convolutional neural networks for the task of semantic image segmentation. Image segmentation is a computer vision task in which we label specific regions of an image according to what's being shown.

In this post, I'll discuss commonly used architectures for convolutional networks. As you'll see, almost all CNN architectures follow the same general design principles of successively applying convolutional layers to the input, periodically downsampling the spatial dimensions while increasing the number of feature maps.

A variational autoencoder (VAE) provides a probabilistic manner for describing an observation in latent space. Thus, rather than building an encoder which outputs a single value to describe each latent state attribute, we'll formulate our encoder to describe a probability distribution

Autoencoders are an unsupervised learning technique in which we leverage neural networks for the task of representation learning. Specifically, we'll design a neural network architecture such that we impose a bottleneck in the network which forces a compressed knowledge representation of the original input.

In previous posts, I've discussed how we can train neural networks using backpropagation with gradient descent. One of the key hyperparameters to set in order to train a neural network is the learning rate for gradient descent.

In this blog post, I'll discuss a number of considerations and techniques for dealing with imbalanced data when training a machine learning model. The blog post will rely heavily on a sklearn contributor package called imbalanced-learn to implement the discussed techniques.

In this post, I'll discuss considerations for normalizing your data - with a specific focus on neural networks. In order to understand the concepts discussed, it's important to have an understanding of gradient descent.

When creating a machine learning model, you'll be presented with design choices as to how to define your model architecture. Often times, we don't immediately know what the optimal model architecture should be for a given model, and thus we'd like to be able