MACHINE LEARNING – so foreign and yet so familiar. We hear it everywhere, on LinkedIn, on company websites and in webinars. Companies describe using machine learning as one of their strengths and yet sometimes it has a dark shadow and is even cast as a villain. In the documentary “the Social dilemma” they explain how advanced algorithms are used to lure us in and create an addiction to social media. Machine learning is also associated with intelligent robots that think for themselves and we fear that they will, in the future, overrule us. So, what is machine learning and is it good or bad, to our advantage or disadvantage? And what are its implications from a pharma point of view?
Machine learning is a branch of Artificial Intelligence and is a computer program that learns, through algorithms and data, to see patterns, calculate and recognize best outcomes. By punishment and reward, that to a certain extent resembles the way humans learn, you can teach a computer to drive, recognize words and images, diagnose illness, and so much more.
It is pretty impressive that a program such as Alexa or Siri can, through machine learning, understand human language and execute your commands. Companies can also profit from machine learning by for instance using collected data to predict what a consumer is most likely to buy. However, there is more to machine learning than coolness and profit. By using machine learning a computer can process an enormous amount of data at a very rapid speed which we humans cannot. With self-driving cars this means that we can avoid many accidents as most crashes are due to driver error. In the pharma industry machine learning can in many cases, based on historic data, diagnose a disease much faster which means we can start treatment in an earlier stage. Furthermore, machine learning can play a role in overcoming a pandemic such as the one we are in now. It can help in detecting, understanding, and predicting the spread of the disease which can provide early warning signs and give suggestions on effective ways to handle a situation. However, a large amount of data needs to be collected to do this. Companies around the world are proposing phone-based apps that can track people’s contact with those diagnosed with or recovering from the virus. While this data is beneficial, some critics suggest however that these types of solutions pose many ethical questions on privacy.
There is a growing ethical aspect to machine learning because the more accurate we want it to be the more data we need to feed the computer. As this field is growing and it is becoming easier to collect data, the more crucial it is that there are rules and regulations on how this data is used and that ethical aspects are considered. Though we sometimes paint AI and machine learning as self- thinking robots (which they are to a certain extent), but they aren’t self-made. We created these algorithms, and we control them. It is imperative to always evaluate what we ask a computer to do, and what parameters we let the computer act within. If we do this right, machine learning can in many cases literally be lifesaving.
From a business point of view, machine learning can help better understand the data available and allow for more informed decisions, which is crucial to stay competitive and adapt to trends as we live in a rapidly changing world. This can be difficult though because not every company has this kind of expertise in-house. In that case it can be a great idea to partner up with research centers that have a vast experience in this field and can help with guidance and knowledge.
In fact, we at Captario are working together with scientists from one of these centers, Fraunhofer Chalmers Centre (FCC), to identify value-adding areas for application of machine learning and AI. Captario SUM is a platform that Captario provides for pharma companies to help them build models that allow them to test different scenarios in order to make better informed decisions. This insight that SUM provides is especially useful because there are millions of different factors that affect the drug development process. Therefore, suitable methods and algorithms will be developed and implemented as features in Captario SUM, enabling swifter and more reliable decision support for users of SUM.
Captario has also started a podcast to shed light on these exciting Data-and-Pharma topics and explore their interconnectivity. Experts from these fields share their knowledge and experience while doing it over Fika – a Swedish social coffee culture. In the first episode we talk about machine learning with Mats Jirstrand, an Associate Professor in Automatic Control and the head of Systems and Data Analysis department at FCC. Mats breaks down machine learning and explains the different ways various fields can benefit from using AI. You should definitely tune in:
In conclusion, machine learning can be a great advantage for everyone involved if we continuously evaluate what we ask a computer to do and consider the ethical aspects of the job.
If you are in pharma, we would love to hear your take on how machine learning can be beneficial in this field! And if you want to know more about Captario, don’t hesitate to reach out! Contact me at stephanie.mardini@captario.com