Entity Extraction Enables Real-Time Threat Detection and Tracking

Companies Need to Deal with an Unprecedented Threat Level

Companies need to track potential and ongoing threats to all aspects of their operations. These days the prime source of information about such threats is on-line. For example, political and economic events around the world often affect supply chains: a company may find a main international supplier knocked off-line for a while, have its operations severely curtailed, or its ability to export reduced. Other adverse events, such as violent conflicts, natural disasters, and cyberattacks, also need to be watched. Any of these could have very harmful effects on a company’s ability to operate.

To gain this information, companies have to monitor various data sources: the open web, dark web, social media, message boards, etc. Company names – particularly the Fortune 500 – pop up frequently in these places. Companies need to be aware of tweets such as “We are coming after XYZ Company,” and be able to quantify how prevalent they are – are they increasing or decreasing?  They also need to track adverse comments not just about the company as a whole, but about specific aspects of its business. Companies need to be able to distinguish between bad but harmless sentiments expressed regarding the company (“I hate their customer support”) and those that are more sinister (“XYZ’s CEO deserves to die”).

Other threats to a company include its products being sold on the Dark Web or leaks of its critical IP data popping up somewhere. Threats to its infrastructure – plants, offices, etc. – need to be monitored, particularly in the case of energy or utility companies. A company also has to look for mentions of its executive team and what threats might be being issued against them. Companies also need to monitor competitive threats, e.g., a pharmaceutical company needs to monitor its competitors’ clinical trials.

Entity Extraction for Detecting and Tracking Threats

It is an enormous challenge to analyze all relevant sources of threat information due to their sheer volume. It would be physically impossible for human analysts to wade through these sources looking for relevant information. Fortunately, there is a technology that can help.

Entity Extraction finds the threat information quickly within very large amounts of unstructured data and turns it into structured form so that the downstream applications can utilize it and present it to the user in real time. A popular form of visualization for such rapidly developing data is a dashboard, which can track multiple threat data in real time.

At its most basic level, Entity Extraction finds previously unknown named entities like the names of people, organizations, time expressions, locations, and monetary amounts. In conventional keyword search, to find a person name you have to enter that name. Entity Extraction will find instances of names previously unknown. It performs a sophisticated linguistic analysis of the context surrounding a name and uses distinctive clues that indicate the presence of a name.

An advanced capability called Event Extraction is also able to identify a great many relevant events — including political changes, conflicts, natural disasters, assassinations. It identifies the nature of the event, the participants, and the date and time if available. It will identify if political unrest is breaking out in a region where one of your key suppliers is based. It will analyze reports of natural disasters’ affecting factories in a given country and report the seriousness of the damage.

A special feature of Entity Extraction, geotagging, will allow the extracted events to be plotted on a map. News reports or social media typically don’t supply the latitudes/longitudes of where events have occurred. They usually only give a place name. This is unfortunate because place names are ambiguous (think “Springfield,” which is a common town name in the U.S.) and companies need to know precisely where the disruptive event is happening.

In sum, Entity Extraction is a valuable technology that will help companies safely navigate increasingly dangerous waters. It provides real-time, high-volume monitoring of all activities and events that may affect a company’s performance.