Dr. Subbalakshmi’s research in artificial intelligence and machine learning focuses on mental health diseases and disorders. Her research and inventions resulted in an inexpensive, unobtrusive AI-based tool which monitors speech to identify dementia, Alzheimer’s disease, and aphasia markers in a natural real-world setting for early detection and monitoring.

As a Founding Director of the Stevens Institute for Artificial Intelligence and a Professor in the Department of Electrical and Computer Engineering at Stevens Institute of Technology, Dr. K.P. “Suba” Subbalakshmi focuses her research on artificial intelligence and machine learning, emphasizing mental health. Unfortunately, current health systems have not yet adequately responded to the burden of mental disorders. Early identification and intervention can help bridge that gap.

Dr. Subbalakshmi recognized how often early signs of human cognitive diseases like Alzheimer’s, aphasia, and dementia are unsuspected or undiagnosed. Artificial intelligence (AI) has shown promising predictions, detection, and treatment solutions for mental health care.

Mental disorders are among the most common health conditions worldwide, affecting millions of individuals. According to data from the World Health Organization (WHO), approximately 50 million people have dementia; 280 million people are affected by depression, one of the leading causes of disability. Suicide is the second leading cause of death among 15-29-year-olds.

Dr. Subbalakshmi’s inventions feature an inexpensive, unobtrusive technology to monitor a patient’s mental health continuously, analyzing patients’ words and speech in a natural real-world setting (e.g., phone calls, conversations, social media interactions) to identify markers for dementia, Alzheimer’s disease, or aphasia. This machine learning-based algorithm of anomaly detection with linguistic biomarkers is a promising approach to diagnosing early-stage cognitive impairment with accuracy as high as 93 percent.

As this technology enables continual monitoring, unlike brain imaging, Dr. Subbalakshmi  envisions that this algorithm could be implemented in smartphones or smart home devices for unobtrusive mental health monitoring, foreshadowing the next generation of AI-based, affordable mental healthcare.

Dr. Subbalakshmi has also been recognized for her suite of AI-based tools that can detect deception online, identify gender from text content, and digital image steganography and steganalysis. She is the author or co-author of dozens of articles and publications in her areas of expertise. She is a Fellow of the National Academy of Inventors, a Jefferson Science Fellow, a Member of the National Academy of Sciences Engineering, and Medicine’s Intelligence Science and Technology Experts Group (ISTEG).

Authors: Elvis Kim, attorney and Tara Norgard, shareholder, and mentor.

To learn more about Carlson Caspers’ commitment to Diversity & Inclusion, click here.

To apply for the Carlson Caspers Diversity & Inclusion Scholarship and meet past recipients, click here.

To learn about Carlson Caspers’ engagements with organizations committed to diversity and inclusion in the legal profession, click here.

To learn about other featured inventors, click here.


  • US 10,593,336: Machine learning for authenticating voice
  • US 10,548,021: Systems and methods for wireless spectrum access in heterogeneous networks
  • US 10,516,701: Natural language processing artificial intelligence network and data security system
  • US 10,397,829: System apparatus and methods for cognitive cloud offloading in a multi-rat enabled wireless device
  • US 9,959,365: Method and apparatus to identify the source of information or misinformation in large-scale social media networks
  • US 9,292.493: Systems and methods for automatically detecting deception in human communications expressed in digital form
  • US 9,116,877: Psycho-linguistic statistical deception detection from text content
  • US 8,555,052: Data hiding based messages and advertisements
  • US 7,529,384: Robust hidden data extraction method for scaling attacks