The Archive

Success is a difficult thing to measure and foster. Good advisors help all of their students succeed.

Org mode is a fantastic tool for managing references. Here's a description of how I use it, and some additional packages, to manage my annotated bibliography.

Detailed notes are needed to accomplish individual projects or tasks but not the big picture. Task lists and calendars focus on the big picture at the expense of detail. An effective organizational system requires both. This post gives a high-level overview of my Getting Things Done inspired organizational workflow.

Particularly for safety-critical applications or the automation of tasks that can directly impact quality of life, we must be careful to avoid the valley of AI trust—the dip in overall safety caused by premature adoption of automation.

One communication pitfall I often see is that many researchers will take figures from their papers and paste them into their slides. Here, I provide some tips for tailoring your figures to talks.

My thoughts on the past year of progress in Robotics and Machine Learning.

Surrounded by brilliant people, I see my friends and colleagues produce surprising insights seemingly from thin air. Only through dedication to a craft can one gain the depth of understanding necessary to demonstrate this level of mastery.

I rarely find these /Myth List/ articles compelling, yet many such articles could benefit from a more clearly presented statement of the author's viewpoint.

Recent ire from the media has focused on the high-power consumption of artificial neural nets (ANNs), yet popular discussion frequently conflates training and testing. Here, I aim to clarify the ways in which conversations involving the relative efficiency of ANNs and the human brain often miss the mark.

Machine learning must always balance flexibility and prior assumptions about the data. In neural networks, the network architecture codifies these prior assumptions, yet the precise relationship between them is opaque. Deep learning solutions are therefore difficult to build without a lot of trial and error, and neural nets are far from an out-of-the-box solution for most applications.

One of the key components to good technical communication is the right amount of context, and only experience yields such knowledge. Technical communication is understandably hard for newcomers.

Many machine learning systems are optimized using metrics that don't perfectly match the stated goals of the system. These so-called "proxy metrics" are incredibly useful, but must be used with caution.

I've discovered over time is that many of my skills and passions have developed only through little side-projects that never see the light of day. The combination of the satisfaction I get from freely exploring a new idea and the occasional long-term reward make me feel as if I am constantly growing.

Big, publically available datasets are great. Yet many practitioners who seek to use models pretrained on this data need to ask themselves how informative the data is likely to be for their purposes. Dataset bias and task specificity are important factors to keep in mind.

Using DeepMind's AlphaZero AI to solve real problems will require a change in the way computers represent and think about the world. In this post, we discuss how abstract models of the world can be used for better AI decision making and discuss recent work of ours that proposes such a model for the task of navigation.

An overview of the significance of simulation tools in the field of robotics and the promise and limitations of photorealistic simulators.

In my last lab, cosmic radiation was one of the many inescapable factors that limited how precisely we could measure the output of our laser; only by understanding what limits performance can one hope to have a reasonable estimate of what is possible.

The potential advantages of AI are many, and using machine learning to accelerate your business may outweigh potential pitfalls. If you are looking to use machine learning tools, here are a few guidelines you should keep in mind.

Translation is about expressing the same underlying information in different ways, and modern machine learning is making incredibly rapid progress in this space.

Bias is a pervasive problem in AI. Only by discouraging machine learning systems from exploiting a certain bias can we expect such a system to avoid doing so.

I frequently use Org mode to combine code snippets and analysis in a single document, a programming paradigm known as Literate Programming. Here are a few example showing how powerful this setup can be.

Having too many specific tasks is limiting and can stymie the creative impulses that give rise to fantastic employees.

Here are five deep learning papers I felt rose above the rest in 2017.

Clarity is key. Here are some tips for improving your website or blog.

The tech culture is broken, and Uber is just the tip of an iceberg.

Most email clients are a pain. Emacs & mu4e are less of a pain.

This time of year can be tough, so reach out to a friend to check in.

After using Emacs for 2 years, I decided to give "evil-mode" (Vim keybindings) a tentative try and I'm not looking back.

I use Emacs' "org-mode" to organize my life. Here's a snippet of how I get it all working.

More complicated ≠ better. In many circumstances, using the most sophisticated approaches to Machine Learning (like deep learning) may not be worthwhile.

Maxwell's Equations govern the study of electromagnetism, one of the fundamental forces of nature. Here, I attempt to motivate these beautiful equations and present some of their more interesting consequences.

Programming matters. As more of the country recognizes the importance of the skill, an increasing number of methods for teaching programming concepts have popped up. However, the emphasis should lean towards how to apply programming across disciplines.

Ubuntu's 802.11n Wi-Fi driver is notoriously broken; here's how to turn it off.

Jump-start emacs with use-package and never manually install another package again

Here's how I built this blog, which relies on Django/Apache on Amazon EC2

Handle large-project C++ completion with Irony & GNU GLOBAL

Koding’s free cloud-based linux dev environment has had me playing for a while. Here’s how to use it with Django.

Obligatory first post; welcome to our blog.