Industry Vertical: Investigative

Agencies cannot afford to discount information because of its format, source, or language. Data may be internally-generated documents or databases, captured hard-disk images, open-source information, data from other departments, broadcast media, CCTV footage, or audio intercept. The challenge faced is how to deal with these large volumes of information to find the needle in the haystack.

Autonomy's Intelligent Data Operating Layer (IDOL) technology provides a single platform solution that can process data across 1,000+ data formats, including multimedia sources, and provide powerful and consistent analytical functions across all data. With over 400 connectors, IDOL can access the data wherever it is stored, without the need to migrate data into a central document management system for analysis. This means that operational systems can continue to operate as designed, but their data can be accessed, enhanced and made available for analysis, which would not otherwise be possible.

Autonomy's IDOL technology uses patented algorithms to form a conceptual understanding of the data that it ingests. This conceptual understanding can provide unique methods of finding key information. Clustering of large volumes of data into self-similar groups, based on the concepts they contain, can be used to identify areas of information for further analysis or to potentially discount large volumes of irrelevant information. This also enables new terms or mis-used terms to be detected and discoverable. Language can change when criminal activity is taking place in order to create a “code” or divert from the activity. With Autonomy, the subtleties, links and patterns in language can be identified.

Visualization tools allow analysts to easily see how information has developed over time. These types of tools allow an analyst to identify new lines of enquiry and discover the “unknown unknowns”. Additionally, the concepts in a document can also be used to automatically link information together that would otherwise not be linked. This allows speedy location of other relevant information once one piece of the jigsaw has been identified.

Additionally, Autonomy can analyze the usage patterns of information to build an implicit profile of analysts' areas of interest through an understanding of the information that they consume or produce. Through this implicit profile users can be easily located, allowing less experienced staff to be directed toward the experts for advice. Experts can also build explicit profiles to capture key knowledge areas that can be shared with other users as live updated feeds.

Investigative agencies utilize vast amounts of structured data either held in internal databases or received from other departments. In addition to IDOL's powerful capabilities across unstructured information, Autonomy's IDOL Structured Probabilistic Engine (SPE) platform works seamlessly with structured data held in databases, or xml. In the majority of cases this structured data contains some larger unstructured elements and this is where IDOL SPE offers huge benefits over traditional structured databases, combining its powerful conceptual capabilities with more traditional structured searches. This provides a more flexible alternative to SQL search by allowing probabilistic fuzziness across the data that can assist when an SQL query would provide zero results. This capability can be used to explore dirty data that includes deliberate misspellings, typos, and transliteration errors that reduce the effectiveness of SQL matching. This allows intelligence agencies to spot patterns and matches in information that would otherwise have been overlooked.

In most investigative applications, reference to a location is vital. The location may provide the key to an investigation, or could be used as a cue to route information directly to an operational unit on the ground. This is especially true of data that did not originate from within the organization or has not been geo-referenced. IDOL Eduction can extract place names or addresses and use a gazetteer to geo-reference the information. Once geo-referenced, IDOL provides powerful geospatial search capabilities, such as bounding box, radius, and polygon search that can be used alone or in combination with structured and unstructured search or automated workflows.

In the advent of the 9/11 terror attacks and subsequent events around the world, government departments and intelligence agencies are under increasing pressure to share information. However, they are sometimes constrained in what they can share because of legal and regulatory requirements, particularly when dealing with personal information, and because of data and source sensitivity issues.

In addition, information often resides within operational systems or document management repositories with limited access by other organizations. Even if access were provided, analysts or investigating officers would need to go to each individual system looking for relevant information.

Autonomy's Intellectual Asset-protection System (IAS) respects the native security of the repository where the information is stored. In addition, it is often necessary to support additional access control requirements, from role-based access through to government defined access restrictions based on classification, code words and nationality. Autonomy mapped security ensures that security is respected within the architecture, in a scalable manner that places minimal additional load on the existing infrastructure. Autonomy software is trusted to run on some of the most secure and sensitive networks across the globe.

Autonomy's technology can access and index data from operational systems and document repositories in multiple departments, and because it respects the native access controls of the source repository, it can provide a trusted single point of access across all data with powerful and consistent discovery and analytical tools. To enable information sharing while protecting operational sensitivities and legal requirements, the system can be architected to provide tiered discovery. Within the bounds of the access control restrictions, non-sensitive data can be accessed directly. For medium sensitivity information analysts can be provided with key metadata or summaries and a point of contact to seek access. For highly sensitive information, where even its existence should not be publicized, the custodian of that information can be alerted that someone is searching for it, and can then make a judgment call on whether access should be granted.

Surveillance of individuals as part of law enforcement or intelligence operations is labor intensive. When information is gathered electronically it is generally still not in an easily processed form – typically video and audio files. Autonomy solutions reduce the workload in processing this data, freeing the analysts to concentrate on the higher value analysis rather than having to review many hours of 'empty' footage.

When reviewing CCTV footage, operator concentration is difficult to maintain and in recent research it was identified that after 15 minutes around 85% of on-screen information is missed. In unsophisticated systems, typically all of the footage has to be reviewed, or if in a quiet location, simple techniques, such as video motion detection (VMD), can be used to reduce the footage.

Autonomy's Virage Security and Surveillance division provides powerful analysis capabilities that can capture, record, and analyze CCTV footage. The Intelligent Scene Analysis Software (ISAS) can be trained to spot particular behaviors of objects in the video and automatically generate events. These events can provide real-time alerts for immediate response, or can be used to drastically reduce the amount of footage that has to be reviewed, even in busy environments where simple techniques such as VMD would be useless. It also provides a full review workflow for dealing with the events generated, including the capture of additional operator analysis and comments. This workflow capability can define workflows through a drag and drop visual designer so that any workflow required by the organization can be supported.

Open source information is often a key data source for investigative agencies. It offers challenges for collection, partially because of data volumes involved, but also because in some cases, knowledge that the agency is interested in the information may be sensitive.

Autonomy connectors can index information from all published internet sources, including websites, wikis, blogs, and RSS feeds. It can also collect data from social media sites which can provide unique insight into individual's social networks, and can be invaluable as part of any investigation. The connectors can also generally be configured to work in conjunction with measures designed to conceal the identity of the requestor. Once indexed, Autonomy offers all of the enhanced analytical capabilities across the information gathered, which is vastly in excess of the capabilities of internet search engines that are tuned to return matching sites by popularity rather than their relevance to the query.

Broadcast media can be used to provide key insight into events and the point of view from other nations. Television and radio channels in many languages can be monitored, recorded, and analyzed simultaneously. Autonomy Virage Rich Media solutions provide deep video and audio indexing to automatically enhance the captured media with items such as scene change detection, OCR of on-screen text, facial recognition, and time-synchronized transcript. This additional information enables users to navigate around the multimedia files to precise locations where key information is discussed without having to review the entire broadcast. In addition, because of Autonomy's unique conceptual understanding of the audio, the whole suite of analytical functions available can be used to exploit the multimedia asset. Any search or alerting operations only return a key segment of the multimedia file ensuring that analysts are taken directly to the relevant information without having to watch long and irrelevant video clips.

Hull is one of the UK's liveliest cities, with a working age population of over a quarter of a million, and a total population of 550,000 when combined with the neighboring riding of East Yorkshire. A few years ago, growing crime figures and disorder were causing increasing problems for Hull City Council. Anti-social, violent and drunken behavior was becoming the bane of people's lives, and policing strategies were proving difficult to enforce without concrete evidence.

Having been set a target of reducing crime and anti-social behavior by 40%, Hull City Council invested £6 million in what has become the largest CCTV installation outside London. Hull City Council chose Autonomy Virage to provide a vast network of Digital Video Recording (DVR) software and integrate over 280 cameras throughout the city.

Autonomy was chosen as the only vendor able to accommodate the vast amounts of data created every day and offer suitable archiving solutions. Users are able to access and retrieve evidence from multiple incidents simultaneously through a single, intuitive interface. Autonomy Virage's digital recording technology also offers unrivalled imaging, delivering 'directed surveillance' for multiple enforcement agencies and ensuring prosecutions through high-quality ID recognition shots.

Since the installation Hull city centre has seen a dramatic reduction in the amount of crime and anti-social behavior. Evidence from CCTV has helped various law enforcement organizations make more than 6,500 arrests over the past three years and operators in the control centre are being called upon up to 60 times per day to provide footage as evidence. CCTV footage provides concrete evidence in cases that otherwise may have gone unresolved, and officers estimate that footage will provide vital evidence in 8,700 cases this year.