Gaoyuan Zhang

Title: Gaoyuan Zhang: Innovator in Neural Network Security

Introduction

Gaoyuan Zhang is a prominent inventor based in Medford, MA (US). He has made significant contributions to the field of machine learning and neural networks, particularly in enhancing their security against adversarial attacks. With a total of 5 patents, Zhang's work is at the forefront of innovation in this critical area of technology.

Latest Patents

Zhang's latest patents include groundbreaking methods for protecting machine learning models. One of his notable inventions is a method for root cause analysis of the vulnerability of neural networks to adversarial examples. This method involves determining the concept-level interpretability of respective units within the model, assessing their sensitivity to adversarial attacks, and enhancing defenses by masking sensitive units. Another significant patent focuses on an interpretability-aware adversarial attack and defense method for deep learning systems. This invention supports convolutional neural networks (CNNs) by localizing class-specific discriminative image regions and measuring interpretability discrepancies to strengthen the CNN against adversarial attacks.

Career Highlights

Zhang is currently employed at International Business Machines Corporation (IBM), where he continues to push the boundaries of technology. His work is instrumental in developing more robust machine learning systems that can withstand adversarial challenges.

Collaborations

Zhang collaborates with talented individuals in his field, including Sijia Liu and Chuang Gan. Their combined expertise contributes to the advancement of innovative solutions in neural network security.

Conclusion

Gaoyuan Zhang's contributions to the field of machine learning and neural network security are invaluable. His innovative patents and ongoing work at IBM highlight his commitment to enhancing the safety and reliability of artificial intelligence systems.