TYPES OF AI
We’ve been hearing about AI in pharma with increasing frequency for several years now, but what exactly do we mean when we discuss artificial intelligence? AI encompasses a number of different ways that automated algorithms are used to conduct tasks traditionally performed by humans.
These types of AI include, but are not limited to:
Machine Learning (ML): Algorithms teach computers to analyze data, identify patterns, classify, and predict outcomes.
- Natural Language Processing (NLP): Machines understand and generate natural language in the form of unstructured speech or text.
- Speech to text and text to speech: Computers transform oral language into text or vice versa.
- Computer Vision: Categorization of content in images, objects, scenes, and activities. Includes facial recognition.
- Expert systems: Computer systems that emulate the decision-making capabilities of humans.
- Planning: Technology that realizes strategy, action sequences, and execution by intelligent devices. Includes robots and unmanned vehicles.
- Robotics: Software that interprets applications to enable transaction processing, data manipulation, and communication across multiple systems.
PHARMA AND AI
Pharma and biotech companies on the leading edge have discovered the value in AI, using it to drive innovation and output in everything from drug discovery to manufacturing. Specifically, they’ve applied machine learning (ML) and natural language processing (NLP) to their processes and seen outstanding results, results which are only continuing to improve, since AI becomes stronger and “smarter” the more data it processes. Advantages seen from pharma and medtech’s use of ML and NLP include:
Improved efficiencies across the spectrum of pharma activities, resulting in reduced time to insight, faster time to market, and ultimately, improved human healthcare for recipients of life-saving or life-improving drugs or treatments.
Drug discovery improvements. As a 2019 editorial in Future Drug Discovery stated, “Clearly, AI is already helping drug discovery – it can help identify drug targets, find good molecules from data libraries, suggest chemical modifications, identify candidates for repurposing and so on.”1
Superior disease diagnosis, monitoring, and prevention. The use of enhanced image analysis via AI applications translates into earlier, more accurate diagnosis, while ongoing monitoring tasks are enhanced through AI applications.
Reduced risk in clinical trials. According to the report “Clinical Development Success Rates 2006-2015” by Biomedtracker, Biotechnology Innovation Organization (BIO), and Amplion, “…programs entering clinical development in Phase I were found to have only a one in ten (9.6%) chance of advancing all the way to FDA approval.”2 Using the data generated from AI, pharma companies are designing trials that succeed, sidestepping costly errors in determining trial feasibility and optimizing outcomes, leading to faster approvals and time to market.
Manufacturing optimization. AI is being used to improve fluidity within the supply chain, reduce waste in the manufacturing process, increase quality control efficiencies, and more.
Increased customer understanding. AI is taking the guesswork out of marketing activities, allowing pharma companies to zero in on the most effective tactics from past campaigns and know what is influencing the customer at each stage of their consumer journey.
MOVING AHEAD USING ML AND NLP
Informa Pharma Intelligence is one of the companies fuelling the power of both machine learning (ML) and natural language processing (NLP) to deliver vitally important data to its subscribers.
Via Pharma Intelligence, the vast, reliable stores of data in Pharma Intelligence products like Biomedtracker are prepped for optimal use in ML and NLP applications. ML and NLP work hand in hand: NL is used to take text-heavy and highly categorical clinical trials data and transform it into the data used in ML models so that the computer algorithm is able to apply patterns to that data and produce insights. Clinical trials data is enriched and structured, allowing the analysis and visualization of the data for its inclusion in successful plans and strategies across clinical trial design, manufacturing, marketing, and more. The end result: faster time to insight and increased success in business outcomes.
DRIVING FUTURE SUCCESS WITH NLP AND ML
The results that come from utilizing AI applications are only as strong as the data itself, an axiom that is especially true of machine learning applications. The use of high-quality, expansive data, pharma and biotech industry data found in Pharma Intelligence, combined with advanced analytics and AI applications in the form of the Pharma Intelligence offering, has helped customers with high-value products solve some of their most challenging key issues in target prioritization, modality innovation, competitive benchmarking, clinical trials design and deployment, and more.
Although the concepts behind AI may be complex, the bottom line is simple: superior data yields superior results. To achieve optimum outcomes, use appropriate, quality data in the right format. As Christian Ehl, co-author of the 2017 book AI&U states, “The key for your AI success is the quality and the quantity of your data. This is your biggest responsibility and this is your biggest opportunity.”3
1 - Lake, Francesca. “Artificial intelligence in drug discovery: what is new, and what is next?” Future Drug Discovery, vol.1, no. 1; Published Online:14 Oct 2019 - Link
2 - ”Clinical Development Success Rates 2006-2015”, BIO Industry Analysis, p.22 - PDF
3 - Ehl, Christian. “Data — the Fuel for Artificial Intelligence” Medium; Published Online: 14 January 2018. - Link