HiddenLayer

HiddenLayer is a developer of a cybersecurity product designed to protect machine learning and artificial intelligence algorithms, models, and underlying data. The company's platform uses machine learning to analyze model interactions and to identify malicious activity without requiring access to the machine learning model or sensitive training data. HiddenLayer's turnkey artificial intelligence and machine learning security models are designed to not add unnecessary complexity to models and do not require access to raw data and algorithms. The company was founded in 2022 by CEO Christopher Sestito, Tanner Burns, and James Ballard and is headquartered in Austin, Texas.

HiddenLayer's platform is a machine learning security (MLSec) platform that works to provide security against attacks on artificial intelligence and machine learning models. These tend to include ML attacks, such as inference, data poisoning, extraction, and evasion. Inference attacks are the process of using the input and output of a model to learn how the model makes its decision and then allow attackers to tamper with the model. Data poisoning is the process of interfering with the data used for learning to corrupt the way the model works. Extraction is an advanced inference attack in which private data or a full copy of the model is stolen. And evasion is a form of inference attack in which the attacker learns how to bypass the intended use of a model.

The MLSec platform includes the HiddenLayer MLDR, ModelScanner, and Security Audit Reporting, with a machine learning-based approach to analyze events in real-time and identify malicious activity without requiring access to sensitive training data. The HiddenLayer MLDR offers real-time defense with flexible response operations that include alerting, isolation, profiling, and misleading and offers users configurable settings to allow companies to fine-tune their company's needs. The ModelScanner scans a machine learning model to identify vulnerabilities to ensure the model has not been compromised and detects malicious code injections. And the Security Audit Reporting works to offer a comprehensive view of AI and ML assets security and validate the ML model security across an enterprise. This includes an on-demand dashboard and distributed reporting with prioritization for vulnerabilities.