A signature is a unique or distinguishing measurement, pattern or collection of data that identifies a phenomenon (object, action or behavior) of interest. Across multiple domains, signatures are used in a variety of ways:
- Biomarkers can be used to indicate the presence of disease or identify a drug resistance.
- Acoustic signals distinguish one maritime vessel from another.
- Explosives might be detected using results from mass spectrometry, Raman spectroscopy, or ion mobility spectrometry.
- Anomalous network traffic is often an indicator of a computer virus or malware.
The most widely understood signature is the human fingerprint.The Pacific Northwest National Laboratory has over 40 years of experience in the development of nuclear signatures, and has developed signature libraries in multiple domains, including the infrared spectral library, optical signatures of bacteria and signature discovery for nerve agent precursors, among others.Our goal is to develop and demonstrate new scientific approaches that include the implementation of algorithms, methods and tools in a reusable analytic framework. We are currently exploring their application in the areas of border security, bioforensics and nuclear non-proliferation to solve important real-world problems.
Current practice is typically accomplished by asking domain experts to characterize and/or model individual phenomena to identify what might compose a useful signature. What is lacking is an approach that can be applied across a broad spectrum to efficiently and robustly construct candidate signatures, validate their reliability, measure their quality and overcome the challenge of detection; despite dynamic conditions, measurement obfuscation and "noisy" data environments.
The solution is a systematic process to rapidly discover new signatures in any domain. PNNL's transformational research agenda is expected to yield new algorithms, methods, tools and techniques that will solve pressing national challenges, and produce an analytic framework where these tools can be tested and evaluated in the context of multi-domain application projects and integrated demonstrations. Researchers at the Pacific Northwest National Laboratory are developing a methodology that will enable decision-makers' ability to
- anticipate future events by detecting precursor signatures, such as combinations of line overloads that may lead to a cascading power failure, biomarkers for the onset of disease and mechanical and chemical hallmarks of material fatigue
- characterize current conditions by matching observations against known signatures, such as the characterization of chemical processes via comparisons against known emission spectra
- analyze past events by examining signatures left behind, such as the identity of cyber hackers whose techniques conform to known modi operandi. Such analyses can contribute to larger signature libraries which in turn serve as a resource for future anticipation and characterization.