Home

At the cornerstone of every scientific experiment, of course, is the setup which is to be used to collect the data we need. While the MLFilms project intends to produce a working machine learning pipeline which can be adapted and used in a wide variety of experiments in multiple labs, we require a test case experiment on which the system can be ...

Although we are able to generate perfectly annotated simulation data, the simulation data doesn’t necessarily represent the data we see in actual experiments. This causes issues when trying to train a darkflow model with simulation data to detect defects in the experimental data. The model will not learn to deal with issues such as noise, low co...

In the previous post, [[link]], we talked about our realization of the XY-model, tilted, smectic liquid crystals. In this post, we’ll go into a little more detail about the XY-model, and how it ties in with machine learning. The XY model is a simple model to describe. First, picture a grid. a simple grid Now, at every point on the grid, we p...

Algorithms A great deal of research has been performed on applying deep learning to object detection tasks. The resulting methods have proved to be quicker and more accurate than traditional computer vision techniques, however, adoption of these algorithms for laboratory usage has lagged due to the difficulty in acquiring the large, labeled data...

We are a group of graduate and undergraduate students at the University of Colorado Boulder, investigating the use of machine learning in experimental physics. This is partly an experiment in applying machine learning to experimental physics, and partly an experiment in open academic collaboration. Follow our journey in real time! what? As pa...