Sherin Mathews
Principal AI Research Scientist
Bio
Sherin Mathews
Dr. Sherin Mathews currently serves as a Principal AI research scientist within the Chief Digital Innovation Team at U.S Bank, where she spearheads the development of cutting-edge solutions at the nexus of AI security & financial services. As a visionary leader in AI, Dr. Sherin Mathews is pioneering innovative solutions at the intersection of artificial intelligence, security and financial services, transforming the future of banking with her groundbreaking AI capabilities.
Dr. Mathews' current research interests center on ethical implications, the potential misuse of generative AI, and the development of transformative technologies to detect and mitigate hallucinations in large language models. She is deeply committed to developing Responsible AI frameworks, with a focus on advancing explainable and fair AI models.
Previously, Dr. Mathews served as Senior Data Scientist at Intel Security/ McAfee, where she has been instrumental in developing next-generation security solutions using computer vision and deep learning techniques to improve and increase the effectiveness of cybersecurity products. Within the Office of the CTO for Intel Security/ McAfee, Sherin pioneered the concept of Explainability and developed solutions for detecting deepfakes ,misinformation detection tools, ransomware attacks & steganography attacks
She is an acclaimed speaker, presenting her research on signal processing, computer vision, and AI at major industry conferences. Before joining McAfee, Dr. Mathews held research positions at Canon and Intel, applying her signal processing and machine learning expertise. She holds a BSEE with honors from the University of Mumbai, an MS in Electrical and Computer Engineering from the State University of New York, and both an MSEE and a Ph.D. in Machine Learning and Signal Processing from the University of Delaware.
Dr. Mathews is the recipient of numerous honors, including the University of Delaware Professional Development Award and 9th place in the prestigious IEEE GRSS Data Fusion Contest. Her impressive track record includes being a finalist for the Executive Women's Forum "Women to Watch" award in 2022, recognizing her as an exemplary leader in technology. Her achievements don't stop there - she was also named one of the Top 100+ Women Advancing AI in 2023, a testament to her trailblazing impact in this rapidly evolving field.
Projects
Deepfake
Deepfakes are falsified videos made by means of deep learning & GANs. Currently working on developing Deepfake Detection frameworks and generated content (GANs) detection. Media Articles on Mcafee Deepfake : GovTech, TechTarget, RSA2020 (Youtube), MobileWire, BusinessWire, NewsOrigin, VentureBeat,Interview on Deepfake, Panel Talk on Deeepfake Policy, Podcasts, RSA2019 and Cheddar.
Explainable Machine Learning (XAI) produces more explainable models while maintaining the high level of accuracy.Enables human users to understand, appropriately trust ML models.
Three novel dictionary learning and a deep learning framework were developed for activity recognition related to remote health monitoring systems.This PhD dissertation presents deep learning algorithms to solve the wearable sensor-based physical activity classification problem
Automatically identifying and predicting siblings from pairs of facial images with high confidence remains a challenge in computer vision applications. This work detects siblings from a pair of images, using defined similarity metric and meta heuristic genetic algorithms.
Publications
To learn and read more about my publications, please check my googlescholar, researchgate,semantic scholar, Microsoft academia and dblp profiles.
Published as a Book Chapter at Elsevier Intelligent Computing - Proceedings of the Computing Conference. This work was presented at Computing Conference London 2019
Here are some highlights of my recent work and Industry presentations
EWF 2021
Presentation on "Designing Trustworthy Equitable AI Solutions" at EWF 2021
​
RSA 2021
CyberSec & AI Prague 2020
Presenting a talk at prestigious CyberSec & AI Prague. Acceptance rate ~36% for industrial workshops
DataX Argyle 2020
Part of opening Keynote Panel at Argyle Forum's #DATAx event. Shared virtual stage with Dr. Nels Lindahl & Tracey Smith to discuss "Future of Data". Presented a XAI Usecase talk under Data Science Strategy & Leadership. DataX is one of top 22 conferences in machine learning and is considered as one of top AI conferences for industrial AI technology developments.
ICML 2020
Presenting a breakout session at WiML and poster session at WiML Workshop at ICML
RSA 2020
Presenting a talk on Deepfake Detection at Emerging Threats Session at RSA 2020. RSALink, Full Recording & Key Takeaways
Executive Women's Forum 2019
Presented an invited Speaker Session on Explainable AI at Executive Women's Forum(EWF) 2019 (One of 20 speaker from 251 submissions. The final selection was determined by the votes of our Content Committee based on the submission Acceptance rate 7%). Link
Computing Conference London 2019
Explainable Artificial Intelligence was accepted as a presentation at Springer Computing Conference 2019,London. Link to Springer Book
Global Big Data Conference
Invited Panelist Speaker at Global AI Santa Clara,California.(January 2018)
IEEE IoT Vertical and Topical Summit
Invited Speaker on "Addressing Privacy attacks on encrypted IOT traffic logs" at University of Alaska, Anchorage - IEEE IoT Vertical and Topical Summit. ( June 2018)
Spotlight @ AISC
Presented Spotlight talk on XAI & DeepFake at AISC.
​
AISC is a global community of machine learning practitioners and researchers who have gathered around topics in AI research, engineering, and products. AISC has a repository of YouTube videos that showcase authors presenting their research, and practitioners demonstrating the use of emerging machine learning methods in the field. The AISC team YouTube channel has over 12K subscribers. For more information visit the web site, Twitter, Linkedin, Youtube recording
UC Berkeley Innovation Talk 2021
Invited Panel Talk on Deepfakes at UC Berkeley Innovation X Labs
DataX Argyle 2021
Panel Discussion on " Structuring High Performing Teams for the AI-Ready Enterprise" with Liora Guy David, PhD(SVP NLP & Data Science Group), Prasun Mishra(Tavant Vice President),
Abhishek Rajpurohit (PayPal Global BI & Analytics Lead)
Abstract
In this session, machine learning experts will weigh in to discuss how to properly structure teams that will embrace, enable and optimize AI solutions company wide.
You will learn:
How to structure teams to maximize efficiency and minimize bottlenecks
Analyze the resources and talent needed to scale AI to meet business needs
How data teams can properly structure data to improve quality and better implement ML models
​
How can you maximize the efficiency of machine learning models? Structuring a high performing team is a start. Business leaders are looking to AI as the solution to their problems and are increasing the pressure on data science to perform. A sizable amount of AI solutions fail, but this percentage could be decreased if business leaders and data leaders worked together to invest in high-performing data teams.
​
Decode-- UC Berkeley
To get the entire list of invited talks at industry and academia, please check "Timeline - External Presentations"
Recent Features
EWF2019
Presented a Session on Explainable Artificial Intelligence at Executive Womens Forum 2019 Session Details: https://ewf2019.sched.com/event/TL9P/leveraging-ml-and-ai-for-security-compliance
XAI at EWF 2019
Presented a Session on Explainable Artificial Intelligence at Executive Womens Forum 2019 Session Details: https://ewf2019.sched.com/event/TL9P/leveraging-ml-and-ai-for-security-compliance
Timeline- External Presentations
Industry Invited Talks
Speaker at Executive Womens Forum 2019
Invited Panelist Speaker at Global AI Santa Clara ,California.(January 2018)
Invited Speaker on "Addressing Privacy attacks on encrypted IOT traffic logs" at University of Alaska, Anchorage and IEEE IoT Vertical and Topical Summit( June 2018)
Session Presented at Grace Hopper Conference 2017 on "Security and Privacy in IoT"
Intel DSCOE "Dictionary and deep learning algorithms with applications to remote health monitoring systems" April 2017
Intel FIT "Dictionary and deep learning algorithms with applications to remote health monitoring systems" April 2017
Invited Talks at Academic conferences & Scientific Publications
Presented a talk on EXplainable Artificial Intelligence (XAI) at Computing Conference London 2019
IEEE CISS John Hopkins University "Centralized Class Specific Dictionary Pair Learning for Wearable Sensors based Activity Recognition." March 2016
ISVC Conference Oral Presentation "Maximum Correntropy Based Dictionary Learning Framework for Physical Activity Recognition Using Wearable Sensors " November 2016
IEEE CISS John Hopkins University "Am I your sibling?' Inferring kinship cues from facial image pairs." March 2015
Presented at Grace Hopper Conference 2015 on "A Deep Learning Framework for ECG classification" October 2015
Academic Services
Awards
Academic & Industry Awards
Professional Development Award
2016
Office of Graduate and Professional Education. University of Delaware, Newark, DE, 2016
CVPR Travel Award
2017
GHC Scholarship
2015,2016
Anita Borg Institute- 2016-Talk & 2015-Poster
Grad Cohort Travel Award
2015
Travel grant for attending the CRA-W Graduate Cohort Workshop
The Computer Research Association’s Committee, 2015
IEEEÂ CISÂ Outstanding Organization Award
2020
IEEE Computational Intelligence Society (CIS) Outstanding Organization Award was award to Mcafee based on FIAT Team Contribution