NetOwl Extractor
Event Extraction
Event Extraction is an advanced AI-based Natural Language Processing (NLP) technology that identifies “Who did What to Whom and Where and When” in text. NetOwl offers over 100 types of events applicable to a wide variety of domains.
Advanced Event Extraction Powered by NLP
In a fast-changing world, it is critical to utilize AI technologies to analyze large volumes of text data quickly and accurately. Named entity recognition (NER) is one of these technologies, but a more advanced and powerful technology called event extraction is needed for the in-depth analysis of what is happening in the world.
Event extraction recognizes events (e.g., arrest, attack, hire) as well as the participating entities (e.g., attacker, attack target, weapon, place, time). It plays a key role in many mission-critical solutions such as Intelligence Analysis, Adverse Media Monitoring, Geopolitical Monitoring, Business Intelligence, PEP, KYC, Due Diligence, and E-Discovery, among others.
While most entity extraction products offer only NER capabilities, NetOwl goes beyond NER to offer advanced NLP-based event extraction with an extensive event ontology pertaining to multiple domains (e.g., business, military, law enforcement, legal, financial, intelligence).
Key Product Features
Accurate
Provides state-of-the-art event extraction accuracy even on noisy text.
Extensive Ontology Coverage
Extracts to a semantic ontology of over 100 types of events along with the roles of participating entities.
Customizable
Creator Edition (CE) enables the customization of entities, relationships, and events.
Coreference Resolution
Resolves co-referring entities, whether they are names, pronouns, definite noun phrases, or events: “the technology company” → “Apple”.
Fast & Scalable
Extremely fast for real-time analysis. Highly scalable entity and event extraction software with Docker and Kubernetes support.
Easy Integration
Easy-to-integrate event extraction product with a REST API. Pre-integrated with popular search and analytics tools like Elasticsearch and Esri ArcGIS.
The Challenges of Event Extraction
NetOwl addresses the unique set of challenges that come with event extraction:
- Lexical Variation: Events can be expressed with multiple lexical choices:
- Chicago authorities arrested/detained/apprehended a suspect (arrest event)
- The intense shelling/attacks/strikes on the compound (attack event)
- Syntactic Variation: Events may be expressed with a variety of syntactic constructions:
- Susan was arrested (verb phrase) vs. Susan’s arrest (noun phrase)
- Conference Resolution: Event participants are often referred to with pronouns and noun phrases:
- In the past Danielle Martin had expressed disagreement with the board. Yet, her sudden resignation sparked concern among investors. (“her” refers to “Danielle Martin”)
- Event Merging: Events are often mentioned multiple times in the same document, with later mentions providing key information (e.g., participants, time, place). Identifying and merging those event mentions is a must in order to extract complete event information.
- An attack on a Kabul hospital has just been reported. The attack was allegedly perpetrated by Pakistan’s military at 3am local time.
Extensive Ontology for Event Extraction
NetOwl offers an extensive semantic event ontology out of the box that is applicable to many domains. An event typically contains one or more participants and often is associated with a particular time or place. NetOwl’s event extraction not only extracts the events themselves but also the event participants. This truly unique rich event extraction is supported by NetOwl’s entity extraction with its broad entity ontology.
NetOwl’s event ontology provides over 100 types of events. Sample events include transactions, conflicts, mergers and acquisitions, personnel events, etc.
Event Extraction and Event Participants
NetOwl not only extracts events but also identifies the event participants. For example, NetOwl can identify multiple event participants with their distinct roles for a given event – an “arrest” event may involve the person arrested, the entity that made the arrest, and the time and location of the arrest.
Event extraction with participant information allows users to gain unparalleled insight from a large collection of text data: companies acquired or acquiring other companies last year, places a given person traveled to in the last month, the number of attacks that happened in a particular city in the month of August, etc.
Connecting Event Participants
For successful event extraction, connecting non-name mentions (pronouns and definite noun phrases) to name mentions is crucial. NetOwl not only extracts named entities (e.g., “Mary Smith”) but also descriptive noun phrases (e.g., “the CEO”) for each entity ontology type.
Additionally, NetOwl connects anaphors such as pronouns (“she”) and definite noun phrases (“the company”) to corresponding named entities through its advanced coreference resolution capability. Together, NetOwl makes sense out of a sentence like “The company announced today that she will be become the next CEO,” extracting a personnel change event with a particular company and person as the event participants.
Event Extraction for Adverse Media Monitoring
Adverse Media Monitoring is the real-time or near real-time analysis of news and other media to identify negative information related to organizations or individuals (e.g., bankruptcies, lawsuits, arrests, corruption, product recalls). Adverse Media Monitoring is crucial for many applications such as Due Diligence, KYC, PEP, AML, Risk Management, etc.
NetOwl’s Event Extraction, with over 100 types of events organized in an ontology, extracts a wide variety of adverse events as well as the participants in these events (e.g., a “lawsuit” event along with the plaintiff and defendant).
By discovering adverse events and event participants automatically and at scale, NetOwl meets the requirements of Adverse Media Monitoring to quickly analyze massive volumes of news and other text sources.
Event Extraction for Cyber Security
NetOwl’s Cyber Security ontology adds Cyber Security-specific semantic entity and event types to NetOwl’s entity and event ontologies. It integrates critical cyber-related concepts from US-CERT, Department of Defense, and other leading cyber security organizations.
Sample cyber entities and events include malware and hacking tools along with denial of service attacks, phishing, website hijacking, etc. With the addition of the Cyber Security ontology, organizations can monitor and analyze ever evolving cybercrime incidents effectively.
Event Extraction Solutions
Adverse Media Monitoring
Intelligence Analysis
E-discovery
Enterprise Search
Remove tag on Small
AML, KYC, PEP
Deploy on Premise or in the Cloud
Trusted by leading global organizations
Featured Blog Posts
What is Event Extraction?
Many critical applications such as risk, link, and geospatial analysis depend on accessing complex information buried in staggering amounts of…
What Is Coreference Resolution?
And why is it hard? We walk you through a hallmark of Advanced Entity Extraction
Entity Extraction is a Critical Technology for Detecting and Monitoring Geopolitical Events
Ongoing and newly emerging conflicts underscore the need to monitor and be alerted to geopolitical events around the world in…
Discover what NetOwl can do for you!
Schedule a demo
Request a no-cost evaluation
Request our whitepaper