Philip A. Chou has longstanding interests in data compression, signal processing, machine learning, communication theory, and information theory and their applications to processing media such as holograms, video, images, audio, speech, and documents. He did the first work on multiple reference frame video coding (used in all modern video codecs), and he originated rate-distortion optimization for codecs (used in all modern video codecs and still driving further advances there). He performed the seminal work on client-driven network-adaptive streaming media on demand, including fast start (used in all modern video streaming), leading up to Microsoft IIS Smooth Streaming and the MPEG DASH standard (used by video streaming providers such as Netflix). He is one of the inventors of practical network coding using random codes, and one of the inventors of wireless network coding. He is the first or among the first to use entropy as a splitting criterion for decision tree design, and is the first to develop a method for decision DAG design. He worked on prosody in the Speech Plus CallText 5010 text-to-speech system (used by Stephen Hawking). He is among the first to define a probabilistic language grammar over images (used in Xerox and Google OCR systems) and the first to use it for optical character recognition of mathematical notation.