Knowledge Engineering

The application of knowledge engineering has exceeded specialized expert and decision support systems (DSS), with advancements in knowledge representation and reasoning (KRR) leading to general-purpose knowledge bases and powerful inference engines – all so computers can answer unforeseen questions with human knowledge.

Requirements

Knowledge engineers need to integrate data which spans multiple domains, including from different lines of business, transform it into information defined by categories, names and relations, and create rules that reason over knowledge via logic.

Drug Discovery

Find new drugs faster by examining the relationships between diseases, pathways, genes, proteins and drugs to accelerate target identification, drug repurposing and disease gene prioritization.

Risk Management

Identify, assess, control and mitigate business risks (e.g., financial or reputational) by analyzing any and all connections between a business and potential threats, direct or indirect.

Supply Chain Management (SCM)

Improve the speed and efficiency of supply chains, and alleviate or prevent disruptions, by optimizing the connections between stores, warehouses, factories, suppliers and more.

Data Fabric

Create a unified layer that connects all of a business's data, often sourced from disparate databases, so employees can discover and access knowledge which encompasses the entire business.

Challenges

The biggest hurdle to effective knowledge engineering is connecting otherwise independent datasets in a way that makes it easy for others to find previously unknown information derived from new connections – whether direct, indirect or inferred.

Ongoing learning

The problem with constant knowledge acquisition is the disruption caused by having to continuously update the database schema in order to add knowledge requiring new types of data.

Endless connections

It's hard to grow a complete domain of knowledge because multilayered, multidimensional and highly interconnected data is exceedingly difficult to model and query.

Complex labyrinths

Once domain knowledge has been mapped, navigating the data and finding relevant paths among an unknown number of complex permutations is virtually impossible to manually.

Diverse datasets

It's common for the ideal knowledge base to include datasets from multiple sources, public and/or private, but traditional databases are designed to link disparate data.

TypeDB Solution

TypeDB provides cybersecurity software and teams with a database capable of storing every detail and connection that exists within their domain, both digital and physical, and applying logic and reason to help extrapolate critical information from it.

Extensible type system

Add new types of infrastructure, threats, attacks and users on demand, upon discovery and without having to modify existing queries.

Inheritance and polymorphism

Query on abstract types (e.g., all injection attacks) rather than having to know and specify all possible types (e.g., SQL injection) with joins.

Relations as a core type

Identify the connections between attackers and assets, from assets reachable by a compromised server to the next steps in a complex attack.

Built-in reasoning

Let the reasoning engine traverse relations to find relevant information (e.g., assets a user has indirect access to) without specifying them.

Further Learning

Download sample projects to get up and running in minutes, and check out the latest blogs from our research engineers.