Programmatic discovery, retrieval, and analysis of communications to identify abnormal communication activity
Conventional email filtering services are not suitable for recognizing sophisticated malicious emails, and therefore may allow sophisticated malicious emails to reach inboxes by mistake. Introduced here are threat detection platforms designed to take an integrative approach to detecting security threats. For example, after receiving input indicative of an approval from an individual to access past email received by employees of an enterprise, a threat detection platform can download past emails to build a machine learning (ML) model that understands the norms of communication with internal contacts (e.g., other employees) and/or external contacts (e.g., vendors). By applying the ML model to incoming email, the threat detection platform can identify security threats in real time in a targeted manner.
1. A non-transitory computer-readable medium with instructions stored thereon that, when executed by a processor, cause the processor to perform operations comprising:
acquiring an incoming email addressed to an employee of an enterprise;
applying multiple machine learning (ML) models to the incoming email to identify linguistic features from which a determination of risk can be made,
wherein each linguistic feature is representative of one or more words that are indicative of content, sentiment, or tone of the incoming email,
wherein each ML model is designed to detect linguistic features of a different type, such that the multiple ML models are able to collectively surface information regarding a goal of the incoming email, and
wherein the multiple ML models include (i) a first ML model designed to identify financial terminology and (ii) a second ML model designed to identify request terminology;
determining whether the incoming email poses a risk to the enterprise based on the linguistic features identified by the multiple ML models; and
causing display of a visualization component that indicates how a determination regarding the risk was made on an interface.
2. The non-transitory computer-readable medium of claim 1 , wherein at least one ML model of the multiple ML models is unable to identify any instances of the corresponding type of linguistic feature.
3. The non-transitory computer-readable medium of claim 1 , wherein the visualization component is a graphic that specifies why the determination was made, and wherein the graphic is shown overlaying the incoming email on the interface.
4. The non-transitory computer-readable medium of claim 1 , wherein the visualization component includes the incoming email with digital elements embedded therein to visually highlight the linguistic features.
5. The non-transitory computer-readable medium of claim 4 , wherein each digital element includes (i) the one or more words that are representative of the linguistic feature and (ii) a tag indicating the type of linguistic feature.
6. The non-transitory computer-readable medium of claim 4 , wherein each type of linguistic feature is represented by digital elements in a different color.
7. A non-transitory computer-readable medium with instructions stored thereon that, when executed by a processor, cause the processor to perform operations comprising:
acquiring an incoming email addressed to an employee of an enterprise;
applying multiple machine learning (ML) models to the incoming email to identify linguistic features from which a determination of risk can be made,
wherein each linguistic feature is representative of one or more words that are indicative of content, sentiment, or tone of the incoming email,
wherein each ML model is designed to detect linguistic features of a different type, and
wherein the multiple ML models include (i) a first ML model designed to identify financial terminology and (ii) a second ML model designed to identify request terminology;
determining whether the incoming email poses a risk to the enterprise based on the linguistic features identified by the multiple ML models; and
causing display of a visualization component so as to indicate how a determination of the risk was made on an interface,
wherein the visualization component includes a representation of the incoming email in which each type of linguistic feature is represented by digital elements of a different color that are embedded within the representation.
8. The non-transitory computer-readable medium of claim 7 , wherein each linguistic feature is set within the bounds of the corresponding digital element.
9. The non-transitory computer-readable medium of claim 7 , wherein each digital element includes a tag indicating the type of the corresponding linguistic feature.
10. The non-transitory computer-readable medium of claim 7 , wherein the color-type relationship governing the color used for each type of linguistic feature is predetermined.
11. The non-transitory computer-readable medium of claim 7 , wherein the representation includes information regarding an identity of a sender of the incoming email.
12. The non-transitory computer-readable medium of claim 11 , wherein the information includes an email address with which the incoming email was sent.
13. The non-transitory computer-readable medium of claim 7 , wherein the representation includes (i) a subject of the incoming email, (ii) a body of the incoming email, and (iii) an indication of whether the incoming email included any links or attachments.