IMLD: A Python-Based Interactive Machine Learning Demonstration

The related fields of machine learning and pattern recognition have enjoyed significant success in recent years due to the impact of deep learning algorithms [1]. Pattern recognition is the automatic recognition of regularities or trends in data. Machine learning, a closely related field that has evolved considerably in the past two decades, refers to the ability of a machine to learn and adapt to data, improving a system's ability to detect patterns and perform inference. These methods are ubiquitous in engineering today impacting diverse fields including signal and image processing, human language technology, bioinformatics, and finance. The ISIP Machine Learning Demo (IMLD) is a tool used to introduce the basics of machine learning using a highly interactive environment in which users can easily visualize the performance of an algorithm. IMLD was first developed as a Java applet in the late-1990's when Java applets were envisioned as the future of interactive computing [2] and there was an emphasis on web-based educational tools [3]. The Institute for Signal and Information Processing, then located at Mississippi State University, developed a suite of interactive demos to teach important concepts in signal processing [4].