This week I participated at my first poster (A0 formatđą) presentation! Kudos to me đ yay!
My graduate school organised an interdisciplinary day with academics from every lab presenting their research plus PhD students showing posters of their work. Truthfully, the stakes were not very high; neither was the quality of my job. Incidentally, the day-long multidisciplinary overview was as interesting as headache-causing. đ¤Ż
Anyway, here goes
some key points of the poster,
+ my comments.
Fishing for attention đŁ
1st problem: research leads you at the borders of knowledge, therefore abandoning the beaten tracks and â hopefully â find a pristine isle of overspecialisation⌠Where you stand alone.
To connect back with people, the plan was to undertake the trip the other way around, find a starting point they could relate and embark them on the journey all the way into my little isle. đş
Hence, I sought for a clickbait title:
A Healthcare Revolution?
I wished for ârevolutionâ to strike at the heart of French peopleâŚÂŻ\_ (ă)_/ÂŻ
Revolution: drastic change in the structure and nature of an organization. Its technical and/or its social elements affect and are altered by such changes as they are interrelated.
More generally, I like to adopt a socio-technical point of view of our human world where âelementsâ (objects, our behaviour, etc.) establish a complex network of interactions. This view stands as the middle ground between technological determinism (technology alone causes change and shapes society) and social constructivism (purely social constructs define our world). Both happen and influence each other.
For instance, knowledge on nuclear power (a technical element) can be geared towards mass destruction weapons or novel medical techniques, depending on society. Nonetheless, the knowledge exists, and the affordances it offers will modify society.
Disclaimer: Itâs just a brief and lacking presentation of the sociotechnical lenses. There are whole books of theoretical developments out there.
In this regard, a revolution is when many of those connections break/change/appear and result in a radically novel form of society.Â
Now, how to properly check if thereâs a revolution at stake in the healthcare industry?
How to spot a revolution? đĽ
Relying on socio-technical system studies, I propose that:
Implementation of novel tools and alternative perspectives of care may allow observation of the changes in the healthcare production organization.
In the healthcare industry, I spot at least two of such ânovelâ elements:
Machine Learning (ML) tools are novel, mainly as:
⢠New resources and advanced skills are needed,
⢠The outputs are non-deterministic,
⢠Many unanswered issues are to be addressed.
Patient care path personalization
Personalization is at the heart of health care, yet many approaches are still being debated. The three key paradigms are: Personalized medicine, Patient-centered care, Evidence-based medicine. Each approach implies organisational consequences and conveys various sets of values.
Inspecting how those elements take root in the system, or not, could be a way to unveil how the whole structure is changing â if it is.
Study project đ
To carry this analysis, I would love to follow the whole process of ML devices creation, form the problem definition to the effective use â or failure of adoption â of such tools.
This is just a linear representation but those steps are iterative and interactive!
However, the world is not ideal⌠While designing this poster, I somehow realised it. Securing several use cases at the âproblem definitionâ step and expect to witness their development until the end was a pipe dream in the PhD timeline.
Part of the job is to cope with reality. A necessary skill in the researcher toolbox is to be able to collect relevant data as it exists, not as you expect it should be. Although data collection is key, I have overlooked this fundamental problem until now.
A depreciated approach is to discover several of them at various stages; follow them as long as I can; assemble the insights in a patchwork of sorts; fill the blanks somehow⌠Letâs see how it turns out!
Takeaways đ
Well, the post is complicated and long enough as it is. I hope you enjoyed it, nonetheless. Thanks for reading! Feel free to grab some takeaways while you go, itâs their raison dâĂŞtre:
A procedural view of an ML project. Although generally most efforts (or talks) are focused on the modelling, each step is key and has issues that can jeopardise the rest of the project!
An ethical conundrum: I try to unveil some mechanisms behind a revolution. If successful, this would become a technical element (a piece of knowledge) that offers new affordances. Would I be responsible for its usages? Or would it be up to you to employ it according to your own values, either to instigate or to nip in the bud a potential revolution?
An esoteric thought: are socio-technical lenses a philosophical by-product of the (difficult to grasp) concept of ĹĹŤnyatÄ? ĹĹŤnyatÄ can loosely be translated as the void of essence of every âthingâ. Each âthingâ exists, not per se, but in an interrelated state with the rest of the âthingsâ. More (?) on wikipedia.