Is semantic technology the AI sweet spot for pharma?
Barley Laing, UK managing director at Melissa explains why a new type of artificial intelligence could help solve problems for pharma researchers.
Artificial intelligence (AI) has been trumpeted by many in the pharmaceutical industry as the magic bullet to the issues they face in drug discovery and other research. However, expectations need to be tempered because although it’s easy to get excited by the potential of AI it’s important not to get too carried away with the reality of what’s possible today – at least at low cost and at virtually no risk.
The standout value that AI can currently provide to the pharmaceutical industry is in removing the bottlenecks around data identification, extraction, quality, harmonisation and integration. These are all issues that have plagued the pharmaceutical industry for a long time. And in an age of digitisation where researchers are increasingly facing a proliferation in the types and variations of data, including drug doses and molecular structures in many forms, such as text, numbers and images; it’s AI that can play a vital role in helping to make sense of it all.
Semantic technology
A relatively new AI technology that has data quality, integration and analytics at its core called semantic technology, which is also called the semantic web and “semtech”, is well positioned to solve the issues that face researchers in pharma.
Where it really stands out is in its ability to associate words with meanings and understand the relationships between them. By building relationships between data semtech can deliver machine reasoning, allowing computers to understand data objects in meaningful ways; helping researchers to make connections, uncover hidden relationships and deliver greater learnings from their data. Even better, semantic technology can do this in real-time, accelerating research and discovery, for example in assessing drug efficacy, detecting toxicity during drug development, or testing hypotheses. As a result, it not only helps to drive efficiency during the research process, but also uncover hidden insight within the data, which cannot be delivered via any other method.
A further important plus for semtech is it can deliver universal data interoperability. This means data sets that are relevant to each other but exist in other formats or locations can be securely and reliably exchanged; something that’s very beneficial for research.
Additionally, when researchers upload new data into the system semantic technology can update and deliver learnings quickly and cost effectively. Looking at the bigger picture, this technology can redefine concepts and relationships as new information becomes available – increasing its value as a flexible and powerful tool that evolves in step with data.
It’s this evolution in AI and its ability to automate and streamline the process of data capture and integration, which will help researchers in the pharmaceutical industry to move away from time consuming manual work in these areas and focus their efforts on making learnings, quickly and cost effectively, during drug and clinical trials, and other important research.
Data accuracy and intuitive platform
It’s important for the pharmaceutical industry to bear in mind that for semantic technology to successfully deliver insight it requires access to accurate data. This can sometimes be hard to come by in a sector that has a reputation for having poor data because of its complexity, which can result in critical errors and inefficiencies.
To solve this problem the industry needs to prioritise investment in data quality operations. This means continually cleansing, standardising, and enhancing their data. Ideally those in pharma should source semantic technologies optimised for healthcare which include built-in, in-depth knowledge of drugs, genes, and diseases that recognise and validate drug names, variants, dosages, and spellings. This will help prevent confusion and mitigate errors generated by busy researchers and physicians in a very complex field as well as aid data enrichment.
In a world that can be made increasingly complicated by new technology, it’s vital the front-end interface to access semantic technology is easy to use by non-IT personnel. To this end it must enable straightforward visual exploration and data analysis across numerous internal and external databases – with no coding knowledge required – to uncover insight securely. This should include the ability to perform detailed semantic queries easily and interact with third-party applications for analysis.
Semantic technology hits the sweet spot for AI in pharma. It opens up a wealth of opportunity to improve efficiency in drug trials and medical research alike, by extracting the relevant information to deliver insights and outcomes, cost-effectively, and in real-time. It’s time for the industry investigate its wide-ranging benefits.