The personal website of Prof. Gregory J. Stein (mirrored from https://cs.gmu.edu/~gjstein )
This living guide is a collection of best practices surrounding paper preparation, which grows every so often as I realize that I’ve given the same advice multiple times. All my students are required to read through this guide before distributing drafts, even internally. Addressing any issues incrementally avoids a last-minute scramble to fix them later. Note: This guide is a work in progress.…
Imagine a service robot deployed in a household setting, assigned a sequence of tasks to complete—one at a time—within an environment that persists throughout the sequence. However, the robot is not told subsequent tasks in advance, mimicing a realistic scenario when the robot is assigned tasks as the user decides they need to be done.
As an advisor to a team of PhD students as well as a smattering of undergraduate researchers and a PI on a few federal grants, meetings to mentor and manage make up a large portion of my time. Making sure those meetings are maximally productive for everyone involved is critical. That objective is hard to measure: what does it mean for a meeting to be maximally productive? How often should my…
In my work as research faculty and a PhD supervisor, I constantly try to streamline the research process. Much of that work involves automating running (or re-running) of experiments so that we can quickly test new hypotheses, change parameters, or even just change colors in a figure.
We focus on the task of goal-directed navigation in a partially-mapped environment, in which a robot is expected to reach an unseen goal in minimum expected time. To perform well, a robot must understand how parts of the environment the robot cannot currently see (i.e., non-locally available information) to inform where it should go next, a challenging problem for many existing planning strategies…
Consider a robot that is tasked to navigate to an unseen goal in a maze in which a green path on the floor signals the route to the goal. The robot’s behavior is guided by a policy that determines what action the robot should take at any point during navigation. In the figure below, the robot comes across a fork while trying to find the goal: At a fork while trying to find the goal that exists…
Last month, my lab had one of the best group meetings we’ve ever had.
Summary: Mentoring others can be incredibly valuable experience and so opportunities to mentor should be sought out.
What is this guide? This guide aims to outline my strategies to encourage idea generation, essential for long-term research progress, through a habit of regular writing.
I often tell students in my lab that they need to put more of an emphasis on testing their research code. We have automated tests on quite a lot of our code to verify that it works as expected, which frequently helps prevent issues before they arise and makes refactoring code faster, easier, and more reliable. Even research code benefits greatly from automated testing. In discussions on testing…