Read about how knowledge graphs such as Ontotext’s GraphDB provide the context that enables many Artificial Intelligence applications.
Developing innovative products in the dynamic landscape of today’s world is challenging. The growing and ever-increasing number of scientific papers, patents and regulations, combined with the ever-evolving needs of customers, makes developing new products not only an R&D challenge, but also a knowledge management challenge. Together with the depth and richness of knowledge that R&D departments must constantly generate and maintain, grows the need for efficient data discovery and analysis of information. Yet, this information is often incomplete, scattered across platforms or locked in documents that are hard to automatically process.
To collect, classify and manage research data and related information, researchers need technology solutions that allow knowledge generation through the help of machine readable information.
Exactly the solutions knowledge graph technology is sophisticated and mature enough to power.
In today’s landscape of fast-changing customer demands and new research coming out everyday, agile product research and development is pivotal to achieving competitive advantage by bringing new products to market fast. Such agility can hardly be achieved without the help of the proper knowledge and data management approach.
Take for example the cleaning products industry.
Behind each and every detergent there lies a universe of chemical compounds, ingredients and their relationships to human and environmental impact. And, inevitably, to other data and information in the corresponding field. What is the impact of a certain chemical compound? Are the ingredients human and environment safe? What does the latest research in the particular domain say about formulas and their potential use? What are consumer’s preferences and needs? Are there related toxic by-products?
When it comes to research and development of cleaning products, all of the above questions need an answer – and they need it fast.
Translated into the world of data and knowledge management, being able to find connections and relationships is about having an appropriate system. Such a system should enable researchers to keep track of thousands of new researches from the field, to monitor and examine compounds, mixtures and ingredients. In addition, the system should also be capable of representing the complex relationships of chemical compounds to skin irritation, human and environmental impact.
Case in point, a cleaning products manufacturer and the processes they use to implement smart data to link domain-specific knowledge.
Taking an ecosystemic approach when washing our hands or cleaning the car is not by far the first thing we are included to do. Yet, for cleaning product manufactures this is a must.
When developing a new cleaning product, researchers start looking into the scientific literature about reported cases of skin irritation, rashes or other issues in connection to the use of a specific surfactant or surfactants.
Surfactants (surface-active agents) are molecules that, very simply put, bond with each other to form sealed bubbles. Most cleaning products contain surfactants as they are the chemical compounds that create foam to break down and remove dirt from countertops, clothing and skin.
Key to researching and developing surfactants are questions like: “In what way surfactants interact with human skin? “, “Which surfactants are related to certain skin irritations?”, “How surfactants and their consumption impact the environment?”, “What are the latest scientific findings about certain surfactants”, “What are the newest regulations related to surfactants?”, etc.
And when it comes to looking for answers in datasets, knowledge graph technology is an efficient and future-proof solution.
To encompass all that information and make it available for quick navigation and exploration, a comprehensive live knowledge graph was built. Using the power of linked data, it captured the knowledge in the domain of surfactants and the relationship of surfactants to all skin sensitivity issues.
The resulting knowledge graph is now a key asset for the company and powers smart solutions such as semantic search, intuitive data navigation and an efficient interface for knowledge exploration. With it researchers can search for surfactants, their impact on the skin and other related information. This system also provides ways for domain-specific knowledge (usually staying tacit across documents and in the heads of individual researchers) to be shared.
The building of this comprehensive system took the major steps needed for building a knowledge graph. In the specific case of the company, these general steps can be summarized as follows.
First, various substances, compounds and mixtures were described and classified as different types of surfactants in a complete and reliable taxonomy, which by then wasn’t present in the domain of chemical components. Next, a set of PubMed scientific articles about surfactants and reported skin reactions was processed applying text analysis techniques. Last, the extracted information was used to create a domain-specific knowledge graph.
As a result, the newly built knowledge graph allows researchers to uncover correlations between surfactants and reported issues in specific documents and also to navigate and explore data in a more effective way.
At the heart of the creation of any knowledge graph is a business problem and the data related to solving it. In the case of R&D data, the major challenges include incomplete, inaccurate and unreliable terminologies, complex and ambiguous domain-specific information, valuable knowledge locked in scientific literature.
With the example of the cleaning products manufacturer, we showed you how an enterprise knowledge graph proved effective to serve as an intelligence tool for conducting quick and efficient product research. The graph solved an important problem at a very early stage of product development – the efficient execution of thorough research on the relationship between surfactants and skin problems.
And this is only one of the examples of how R&D potential and know-how can be combined with the potential of knowledge graphs to unlock meaning and create a smart future-proof living repository of scientific data and its relationships.