Abstract Robustness becomes one of the most desired properties in machine learning (ML) models due to their increasing adoption in safety/security-sensitive settings. Most attempts to train robust methods…
Abstract We aim to build neural networks that are intrinsically robust against adversarial attacks. We focus on classifying images in real-world scenarios with complex backgrounds under unforeseen adversarial…
Abstract Although deep neural networks (DNNs) have achieved impressive performance in several applications, they also exhibit several well-known sensitivities and security concerns that can emerge for a variety…
Machine learning provides valuable methodologies for detecting and protecting against security attacks at scale. However, a machine-learning algorithm used for security is different from other domains because in…
Abstract Current machine learning models suffer from evasion attacks such as adversarial examples. This introduces security and safety concerns that lack any clear solution. Recently, the usage of…
Our research is about providing safe access to the Internet in places where network access is restricted or censored. Many people are limited in what they can say…
We plan to build a pipeline that leverages novel robust secure machine learning techniques to detect and defeat cybersecurity threats against computer systems. Modern cyber-threats to computer systems…
Large language models (LLMs) such as ChatGPT have greatly advanced coding tasks but often fail to generate secure code. Current approaches to improving code security, relying on fine-tuning,…
We will study how to harden machine learning classifiers against adversarial attack. We will explore general mechanisms for making deep-learning classifiers more robust against attack, with a special…
The Center for Long-Term Cybersecurity has had robust engagement with a wide range of supporters, including industry collaborators, since our founding in 2015. In 2023 we have scoped…
A review of current trends in technology innovation suggests that societal demands for increased functionality conflict with the imperatives of robust cybersecurity. Articulating the nature of this tradeoff is a useful first step (and is the focus of much of this talk), but an understanding of how to manage…
A report published by the Center for Long-Term Cybersecurity, Cyber Resilience and Social Equity: Twin Pillars of a Sustainable Energy Future, examines the importance of cybersecurity in ensuring equitable access to energy. In an era of worsening cybersecurity threats, the paper advocates for “sustainable energy delivery systems that ensure robust defenses without compromising the goals of reducing energy poverty and ensuring energy security.”
