EASY-TO-DEPLOY AI MODELS
PRIMER ENGINES
Choose from an array of ready-to-use NLP models. Tuned to work on domain-specific data for use in intelligence and defense operations, Engines are ready to deploy via API to quickly integrate with your unique applications and workflows, both in the cloud and on-premises.
Primer Engines
-
Performant
Deploy performant, fast, and scalable industrial-grade NLP to your business.
-
Domain-specific
Get access to groundbreaking engines trained and tuned for specific domains and data.
-
Interoperable
Orchestrate engines in robust pipelines for high-scale document processing.
-
Customizable
Rapidly customize models for your unique data and workflows.
Primer Engines
Primer’s leading data scientists are constantly updating and developing new state-of-the-art Engines to support unique and evolving mission needs.
-
Named Entity Recognition
Identify people, locations, and organizations within a body of text.
-
Coreference Resolution
Identify who the words “he,” “she,” and “they” are referring to.
-
Relation Extraction
Identify relationships between entities.
-
Classifiers
Categorize any kind of text.
-
Sentiment Analysis
Classify text as positive, negative, or neutral, returned as a classifier.
-
Topic Modeling
Identify the topic that the text is talking about without the need for training or ontology.
-
Summarization
Summarize a document to meet your needs and identify the key information.
-
Key Phrase Extraction
Automatically identify and extract the key information within a document.
-
Explainability
See what parts of the text the models are using for your predictions.
-
Synthetic Text Detection
Classify text as likely to have been written by a human or a machine.
-
Claims Detection and Dispute Resolution
Identify claims within any text and determine if these claims are supported or refute other claims.
-
Jargon Detection and Explanation
Identify jargon and industry acronyms within the text and link it to an explanation of the terms.
-
Quotes Attribution
Extract a quote and its author from a body of text.
-
Title Generator
Generate a human-quality title or headline for a body of text.
-
Question Answering
Answer specific questions about a body of text.
-
DIMEFIL
Classify a piece of text as being related to Diplomacy, Information Operations, Military, International Economics, Illicit Finance, Intelligence, and Law Enforcement.
-
Location Extraction
Identify all the locations in a piece of text and connect them to a map.
-
Named Entity Linking
Link extracted entities with your knowledge graph or Primer’s knowledge base.
-
Custom Entity Detection
Move beyond people, places, and organizations and identify other entities.
-
Difference Engine
Determine the difference between two pieces of text or two versions.
-
Event Detection
Identify the real world events within any piece of text.
-
Event Linking
Connect events to form a timeline across your set of documents.
Primer Engines
Frequently Asked Questions
-
What are Primer Engines?
Engines are pre-trained, proprietary models built with Primer’s state-of-the-art NLP (Natural Language Processing), NLU (Natural Language Understanding), and NLG (Natural Language Generation) technology. Engines enable humans to automate their reading and writing tasks at scale with human-level precision and discover insights that were previously impossible to find.
-
Who is it for?
Engines support a wide range of applications. They are used by decision-makers, data scientists, data analysts, subject-matter experts, and operators across government and commercial verticals. What these users have in common is that they need to rapidly derive meaningful insights from massive amounts of text-based data to improve their situational awareness, business processes and products, and operational decision-making.
-
How are Primer Engines deployed?
Engines can be deployed in the cloud or on-premises, depending on the organization’s mission needs and the scale of its data landscape. Primer offers managed cloud hosting to public sector customers, or these customers can choose to host Engines on their own GovCloud or on-premises infrastructure. Following deployment, customers can use Primer’s APIs to quickly integrate the engines into their preferred workflows.
-
What is the underlying technology that makes this possible?
Using its internal team of 60+ Machine Learning engineers, Primer has developed a suite of Engines, built on state-of-the-art transformer model technologies. Primer Engines exceed the performance of competitor models while enabling new tasks and use cases, thanks to our custom modeling approaches and our significant investment in high quality, labeled data for specific use cases.
-
How do you ensure fairness and transparency with your engines?
Primer Engines are trained on diverse data sources by design to minimize bias. Information on the training datasets is available to you for review, and you can reach out to our sales team to learn more.
-
Are Engines interoperable?
Yes, Primer Engines are designed to work seamlessly with each other. For instance, Primer‘s Sentiment Analysis engine can efficiently triage and classify a massive cache of documents into those showing positive and negative sentiments. You can deploy the Primer Summarization engine on top of the extracted content to derive summaries of each category.
-
What security measures do Primer Engines have?
Primer consistently leverages state-of-the-art security tools and platforms to ensure your hosted data remains safe. To ensure data privacy, Primer only stores data transiently until it completes the request.
-
How do I integrate Primer Engines with my existing systems and applications?
Developers can access our API to quickly integrate Engines into your enterprise. The API can be used with a single cURL statement and from any programming language. Primer API does not require any special libraries. Developers can make standard HTTP calls to combine multiple engines.
-
What skills do I need in my team to deploy Engines?
Any software developer should be able to help integrate Engines into the company’s workflow with Primer’s API documentation. Once you make a call to the API, the service handles the machine learning required to analyze the text.
-
What should I think about when integrating AI & NLP into my enterprise?
For starters, you will need to address these basic questions:
- Workflow automation: prioritize the areas for automation where Engines can alleviate workflows and improve processes.
- Input systems: identify the Engines’ data sources and where and how they can be accessed and prepared.
- Output systems: determine what applications the Primer output will integrate with, such as products, dashboards, BI tools, knowledge bases and more.
-
What does the output look like and how do I use it?
The Primer Engine API consists of HTTP(s) requests and JSON responses. Requests to Engines are standard HTTP calls that can be made from the command line or using the tools available in your chosen programming language or development framework. The JSON response contains the keys and values that the engine derived for the provided input. For example, if you make a call to the Summarization Engine, the response is a JSON object that contains the key “summary” and the resulting summarization.
-
Do you have an on-premises solution?
Yes, we offer an on-premises solution for Engines. If you select this option, our Solutions Architect will work closely with your representative to design and implement the most appropriate solution for your environment.