No ‘Fake News’ Here
How Sherin Mathews Uses AI/ML to Make Social Media More Authentic
As a child, Sherin’s natural curiosity drove her to ask question after question; always asking “What’s this?” and “What’s that?” One PhD later, Sherin Mary-Mathews is refining artificial intelligence and machine-learning so computers can ask the same questions and identify false information.
Sherin credits two STEM-loving parents and her own innate math skills for steering her towards a career in AI. After completing her undergraduate degree in electronics and telecommunications at the University of Mumbai, Sherin pursued a master’s degree in electrical and computer engineering (with a specialty in wireless communication and signals processing) from the University of Buffalo A heavily math-based field, signals processing allowed Sherin to use her innate talent with numbers, while the ever-changing nature of the field satisfied Sherin’s innate curiosity and passion for learning. It was this passion that also drove her to complete a PhD in Machine Learning, Computer Vision, and Artificial Intelligence at the University of Delaware.
“It is important to be a learner throughout life. One must always- have the inclination to know and learn more.”
While earning her PhD, Sherin received an internship at Intel Security (now McAfee), where she currently works. One of her recent projects involved “explainable AI.” This concept allows customers to understand the underlying technology used in different situations. For example, if AI was used to reject a loan, the customer should be able to know AI was used in the decision process and subsequently, and why the loan was rejected. Data privacy is a critical concern for the security industry. Sherin strives to make artificial intelligence transparent, reliable, and ethical.
“I am really happy to develop break-through technologies…In any work I do, I want to give back to society in some way.”
Sherin’s current project with McAfee is analyzing “deepfake” videos, the most deceptive form of “fake news.” Deepfakes are fake videos that look real but are the product of advanced AI video and audio editing tools, similar to Photoshop. For example, a deepfake video can be created of Sherin belting the latest Mariah Carey hit onstage; or on a more serious note, a video of a politician making wild accusations in a speech that never occurred. Almost impossible to detect with the human eye, computer vision and deep learning technologies allows Sherin and her team to detect and stop deepfakes before they spread.
“The advent of artificial intelligence is considered a major milestone in technological development. AI is steadily flooding our world in a number of industries but the A.I. platforms can also be mis-used if in wrong hands. This is more of an arms race scenario and as security professionals, we need to step up to detect attacks or mishaps before they happen.”
With an amazing support system behind her, Sherin is making the digital world transparent and reliable. Throughout her career, Sherin found a number of personal advocates who helped her build confidence by identifying her unique skills and strengths. Surrounded by an army of advocates, Sherin wants to continue to pay it forward.
“If you are in a position to help someone, especially another woman; whether that be getting her through that leadership door, mentoring, or anything else…do it!”
This story was written by Laurel Kremer, Wogrammer Journalism Fellow. Connect with her on LinkedIn
