Digital Science solutions metaphacts and Dimensions are excited to announce the highly anticipated launch of the Dimensions Knowledge Graph, a large ready-made knowledge graph powering AI solutions in the pharmaceutical and life sciences industries.
The Dimensions Knowledge Graph, powered by metaphactory, is an all-in-one knowledge graph solution, ready-made for easy integration with customers’ data infrastructure and existing internal knowledge graphs. At its core is an explicitly defined and flexible semantic model that can be easily extended to include internal data (which can range from domain expert knowledge or data from internal documents) and harness the synergy of global research and internal knowledge to drive business decisions. Through a unified semantic layer, it connects data across all relevant sources and adds richness and context that simplifies and accelerates the generation of actionable insights grounded in evidence.
Companies in the pharmaceutical and life sciences fields can now gain unparalleled access to an abundance of interlinked global research data covering around 350 million records, such as patents, clinical study reports and publications as well as public datasets and ontologies, that they can use in conjunction with the wealth of internal data they already possess to drive business decisions and power AI applications.
“A unique component of the Dimensions Knowledge Graph is the symbolic AI layer it provides, as it introduces an enhanced level of transparency and trustworthiness to AI applications (e.g., LLMs and gen-AI), which is particularly crucial in pharma due to the sensitive and high-impact nature of its use cases. Companies pursuing independent development of AI within their own systems can benefit from the symbolic AI layer provided by the Dimensions Knowledge Graph, which serves as an essential building block that can enhance the efficiency and reliability of their AI initiatives,” said Dr Peter Haase, Founder & Chief Scientific Officer, metaphacts.
The underlying knowledge graph can be combined with LLMs and gen-AI to scale and support business decisions with machine-generated insights, augment explicit knowledge with AI algorithms and much more. This powerful combination of knowledge graphs and LLMs can also drive internal end-user interfaces for knowledge discovery.
Behind the scenes – What is the Dimensions Knowledge Graph?
The model for the Dimensions Knowledge Graph was built using metaphactory’s semantic modeling interface, which is based on open standards and supports data reusability across all use cases and domains within an organization, as well as extensibility to enable the capturing of internal data. End users can access the integrated knowledge through use-case-specific views for intuitive search, exploration and visualization. Alternatively, insights can be delivered via API and injected into already-established Business Intelligence or data visualization interfaces.
The built-in AI Assists across the solution can deliver trust and explainability to AI-generated insights and help scale business decisions, allowing customers to move from a human-driven model to a human-in-the-loop model.
The Dimensions Knowledge Graph data is built upon one of the world’s largest linked research databases – Dimensions – and includes data covering millions of publications, patents, grants, clinical trials, policy documents and technical reports. The Dimensions database not only enables search on the metadata of publications (e.g., title, abstract, authors), but it also offers search and discovery over full text.
The Dimensions Knowledge Graph enhances Dimensions data by offering integration with public datasets and ontologies, including data, metadata and relations for genomics, proteomics, metabolomics, molecular interactions, biological processes, and pharmacology and includes metadata and semantic annotations derived from 350 million research outputs. Additionally, customers can easily combine this data with internal data or any existing knowledge graphs they wish to integrate.
We believe the Dimensions Knowledge Graph is the largest interconnected set of semantically annotated knowledge, including data from:
- 143 million publications
- 160 million patents
- 30 million datasets
- 7 million grants
- 2 million policy documents
People & organizations
- 34 million researchers
- 129 thousand organizations
Semantic annotations
- 307 billion linked semantic annotations
- 35 million research integrity trust markers
Pharma ontologies/vocabularies
- 30 million concepts from 38 domain ontologies
With the Dimensions Knowledge Graph, internal data becomes contextualized and consumable, and enhanced with knowledge from public data and global research. Through the integration of the Dimensions database and open datasets, you can uncover insights that may have been previously inaccessible and could lead to remarkable solutions and discoveries, and support quick validation of hypotheses.
