MACHINE LEARNING; TRANSITIONING PEOPLE
Transitioning people to a new attitude, skills, competences & the connect to real life solutions is challenge faced with every new technology. It is more conspicuous with disruptive technologies, where the issue is both of technology & its usage. From creating the right data sets, databases & using the conventional data for new usage can throw best of the professionals out of gear even in the world renowned institutions. Who knows this better than Prof. Daphne Koller, co-founder of Coursera; a leading practitioner of ML in areas of Biology & health. Issues raised in this area by her has universal appeal.
Technical vigour is only one part of the story, transitioning scientists & researchers to understand what it can achieve for real world problems is the core. Finally, it has provide succour to the patients. That ML cannot solve all problems & some hybrids can work more successfully in couple of areas has to ingrained. Scientists have a tendency to create ML for their needs & not necessarily match real world problems.
Bilingual Human Resources are most difficult to find. By bilingual resources we mean resources at the intersection of ML & Biology / Health in the present context. Being interdisciplinary is the clincher. Being few they have to be oriented to make the best out of this rare skill subset. Pathologists look at a problem from a different point of view based on their technical knowledge & limited data. Studying the tumour itself &studying the tumour micro ecosystem, can lead to immense difference in the findings.
Siloed existence of Pharma companies companies comes in way of their collaboration with ML practitioners, even if they have decided to do so. Mindsets & jargons differ. Giving conventional datasets for ML analysis is different from a data set created for the exact real world problem. What Prof. Daphne Koller regrets of the fact that she did not think of something big, before Coursera. The contribution gets limited.
TRANSITIONING PEOPLE IS A MAJOR CHALLENGE OF THE ML ECOSYSTEM.