Sponsored by:
Produced by:
Wolfgang Halter
Head of Data Science for the Life Science business of Merck
Innovation in data science can mean the methods, or the actual applications. Current methods are very mature but what is keeping us from unlocking potential is the application of tools which reduce barriers and hurdles.
To what extent do you think innovation in data science holds the key to an organisational financial performance? Please discuss.
Innovation in data science can mean the methods, or the actual applications. Current methods are very mature but what is keeping us from unlocking potential is the application of tools which reduce barriers and hurdles. We need to integrate these novel approaches into well-established legacy processes and systems and rarely have the luxury of beginning from a greenfield.
The biggest factor to leverage more useful data and gain those deeper insights is to think more about the way data insights are implemented, ensuring they are used in the business day to day. That requires better reliability and transparency to drive higher adoption.
It is a significant barrier trying to get researchers and users to trust the system and tools and if you can’t provide that, then change management is far more difficult. In my experience, gaining the trust of researchers requires you to show the potential and power of tools but to also show the limitations.
Building in-house means you gain ownership of the technology and long-lasting ownership structures for that project.
What in your mind should organisations be doing to maximise ROI from data science?
Firstly, organisations can help to drive ROI by prioritising quality over speed and shifting from small proof of concept projects to look at the big problems. If you only look for the quick wins, you will never unlock the full potential that comes with bigger projects of several years with proper resources, development teams and planning.
Additionally, there has been a shift in the industry to externalise data science. I think the potential downside of this change is that you start to lose control of analysis and data. If you want to build something that will deliver long-term ROI then it shouldn’t be outsourced, I think it needs to be built in-house. I have seen too many outsourced projects fail, because they are built by data consultants not internal teams. Building in-house means you gain ownership of the technology and long-lasting ownership structures for that project.
There’s the idea that the currently available technology is a general AI that will solve every problem that you give it. This, in my opinion, will not materialize soon. If I was to predict the future, I would predict that many AI tools will fail due to excessively high expectations. This doesn’t need to be a bad thing: the few tools that will stick, will have a huge potential.
What potential do you see for AI in data science and life science?
AI platforms and services will be developed by the Life Science industry, and I already see many people asking for this technology and producing very powerful use-cases.
However, there’s the idea that the currently available technology is a general AI that will solve every problem that you give it. This, in my opinion, will not materialize soon.
AI shows immense potential in transforming Data Science and Life Science with its ability to generate insights from complex and big data. Nonetheless, it is essential to acknowledge the limitations of AI, for example in verifying the content generated by Language Models. Generative AI can speed up processes and open new possibilities, but it's not a one-size-fits-all solution. Classical Data Science and advanced statistics remain critical in providing reliable and accurate insights.
If I was to predict the future, I would predict that many AI tools will fail due to excessively high expectations. This doesn’t need to be a bad thing: the few tools that will stick, will have a huge potential.
Life Science business of Merck is a science and technology company that drives discoveries and creates technology. In healthcare, the company is discovering unique ways to treat the most challenging diseases, while life science experts empower scientists by developing tools and solutions that help deliver breakthroughs more quickly.
Sponsored by
We are a global player in life-science R&D informatics, providing customers with cutting-edge multi-omics data management and integration software. We develop innovative off-the-shelf and custom solutions for clients in pharma, biotech, agriscience, consumer goods, and research centers. Omics Data Manager (ODM), our flagship product, helps organisations create a FAIR catalogue of multi-omics investigations (studies, samples, omics data), with powerful tools for curating rich and standardised metadata in bulk, as well as optimised RESTful APIs for scalable cross-study, cross-omics integrative search.
Up next: Amir Emadzadeh
