Draft:Rita Singh

Rita Singh is a computer scientist known for her work in the algorithmic dimensions of voice recognition technologies and the application of artificial intelligence in voice forensics. She holds a position as a Research Faculty in the Language Technologies Institute of the School of Computer Science of Carnegie Mellon University. She led global conversations at the World Economic Forum on topics related to voice technologies. Singh is a founder and technology director of Center for Voice Intelligence and Security (CVIS), an organization dedicated to pioneering developments in voice technology and its security implications.

Early Education and Career
Rita Singh completed her early academic pursuits in India, where she received a Bachelor of Science (Hons.) degree in physics and a Master of Science degree in Exploration Geophysics, both from Banaras Hindu University. Her academic journey in geophysics further extended to earning a PhD from the National Geophysical Research Institute of the Council of Scientific and Industrial Research, India, in 1996.

Postdoctoral Fellowship and Initial Research
After completing her PhD, Singh joined the Tata Institute of Fundamental Research in India as a postdoctoral fellow from March 1996 to November 1997. During her fellowship, she was part of the Condensed Matter Physics and Computer Systems and Communications Groups. Her research focused on nonlinear dynamical systems and signal processing, building upon her doctoral work in nonlinear geodynamics and chaos.

Transition to Carnegie Mellon University
In November 1997, Singh transitioned to Carnegie Mellon University (CMU) in Pittsburgh, PA, USA, where she became a member of the research faculty at Language Technologies Institute in the School of Computer Science. At CMU, she affiliated herself with the Robust Speech Recognition and SPHINX Groups. In 2020, she founded Center for Voice Intelligence and Security (CVIS) and currently serves as a technology director.

Research
Singh researches primarily in machine learning, deep learning for computer voice recognition, and artificial intelligence applied to voice forensics.

Contributions to Speech Recognition and Audio Processing
Singh's work in computer speech recognition and general audio processing began in 1997. Her research until 2014 encompassed a broad spectrum of topics within this domain. She developed algorithms that contributed to making speech processing systems language-agnostic, automated the discovery and learning of information from speech, and enabled speech processing with minimal reliance on external human-generated knowledge. Her objectives were to enhance automation, devise more effective search strategies, scale up learning algorithms for voice processing systems, and improve their accuracy in complex acoustic environments, including those with high levels of noise.

Human Profiling through Voice Analysis
In December 2014, Singh pioneered the development of the science of profiling humans from their voice. This innovative field involves deducing various human parameters based solely on voice analysis. Singh posits that the human voice, akin to DNA and fingerprints, is unique to each individual and contains an abundance of information about physical, physiological, medical, psychological, sociological, and behavioral aspects, among others. Her approach is grounded in the quantitative analysis of voice signals, leveraging the principles of physics and bio-mechanics of human voice production - leading to an estimation of 3D portrait of a person. A key feature of her methodology is its language-agnostic nature, focusing on the voice signal rather than the semantic or pragmatic content. This language-agnostic nature significantly enhances the application of vocal biomarker technology in medical diagnostics, including the detection of neuromuscular conditions affecting the upper-respiratory tract, as well as diseases such as COVID-19, Alzheimer's, Parkinson's, and coronary artery disease.

Current Endeavors and Future Aspirations
Singh's current research involves designing advanced AI systems to delve into the rich information harbored in the human voice. These systems are being developed for various purposes, including genetic discovery, biomarker identification, and exploring aspects of the human physical state and psyche, such as emotions and personality. Parallel to her work in human profiling, Singh is also engaged in developing core designs for universal speech and audio processing AI systems. Her vision is to create a system capable of replicating the brain's response to multi-sensory inputs. This ambitious project involves not only advanced computing but also integrating aspects of mobility. Singh is actively working on these dimensions to bring her vision to fruition.

Teaching
She teaches the CMU course 11-785 Introduction to Deep Learning and 11-860 Quantum Computing, Cryptography, and Machine Learning Lab.

Previously, she taught:


 * Artificial Intelligence in Digital Multimedia and Cyber Forensics at the University of Pittsburgh
 * 11-788 Concepts in Digital Multimedia and Cyber Forensics
 * 11-788Computational Forensics and AI
 * 11-788 Advanced Topics: Quantum Computing Lab
 * 11-860 Advnaced Topics: Quantum Computing Theory and Lab
 * Generative AI for Software Implementations in Quantum Computing and Machine Learning
 * 17-620 Quantum Machine Learning
 * 11-775 Large-Scale Multimedia Analysis
 * 11-688 Computational Forensics and Investigative Intelligence
 * 11-364 An Introduction to Knowledge based Deep Learning and Socratic Coaches
 * 11-756 Design and Implementation of Speech Recognition Systems

Books
Singh contributed one chapter to "Techniques for Noise Robustness in Automatic Speech Recognition" (2012) by Wiley. In 2019, she wrote and published "Profiling Humans from their Voice" (2019) by Springer, Singapore