Ansh Surapaneni
Class of 2025Danville, California
Project Portfolio
Publish an article: Kickstart your career as a scholar learning the neuroscience of how your brain interprets sound
Started July 3, 2023
Abstract or project description
The brain is incredible and complex and the source of much scientific research around the world. In this pod, you’ll focus on how our brains allow us to take unassuming sound waves and enable us to enjoy music, learn a new language, or recognize danger. Through this fascinating process, students will learn about the research process, the brain regions, how the brain regions cooperate to interpret sensory information, and how those individual actions culminate into the actions we take every day but may take for granted.
Project Portfolio
Benchmarking State of the Art Speech to Text systems on Accented Speech
Started July 6, 2023
Abstract or project description
Voice to text has been a crucial technology for accessibility, but current research has shown that many of the architectures that exist today perform poorly on accented speech. This is especially important when voice assistants like Alexa and Siri are being deployed around the world. We will be exploring how a variety of state of the art deep learning models perform on the Speech Accent Archive dataset. By doing so, we can hopefully understand what techniques were used by each model to guide a potential future model built for accented data. This will all be done via the HuggingFace API so we can quickly test and validate our findings!