Setting the terminal title in zsh

By default, zsh doesn’t set the terminal title, making it inefficient to know which application is running in each tab, or when a long-running application has exited, as each tab must be visited in turn.

Adopting Colemak

Last year, I reflected that my typing accuracy was a source of frustration, especially when programming. I’d picked up typing decades prior and had developed an idiosyncratic technique. I was happy with my speed, but not my consistency, and had to make regular corrections. Furthermore, I used my right hand to press keys on the left side of the keyboard, which meant my hands were frequently moving side to side, which was tiring during long sessions.

Visibility and monitoring in deployed ML systems

Modern machine learning has ushered in an era of unparalleled system capabilities, exemplified by self-driving cars and synthetic speech indistinguishable from human. However, these techniques bring with them the challenge of monitoring and understanding the behaviour of live ML systems.

Viewing Jupyter notebooks at the command line

The Jupyter notebook is a literate programming environment that has become ubiquitous in machine learning. While the standard tools for interacting with notebooks are web applications, it’s often useful to be able to view notebooks at the command line. This is convenient when you’re logged into a training workstation via SSH, and the process of configuring SSH to forward a port, starting a Jupyter server, and navigating to it in a web browser is a chore to view a notebook for a few seconds.

Representation learning for audio data

The application of classical machine learning methods on complex data formats, such as audio of human speech, typically necessitates extensive feature engineering. This requires significant domain knowledge to extract the key components of the data.

Deep learning can allow models to learn their data representations, obviating the need for feature engineering. However, as the quality of the learned representations strongly influences performance on downstream tasks, how can you ensure that these representations are appropriate?

Jupyter notebooks and collaboration

Git has been almost universally adopted for collaborating on code, as Jupyter notebooks have been for data exploration and interactive modelling. However, when bringing the two together, a problem arises: Git was designed to version plain text files, such as those containing source code, and not structured data like JSON documents or binary data such as images embedded in Jupyter notebooks.