FYI
A Design Thinking Mindset for Data Science
https://towardsdatascience.com/a-design-thinking-mindset-for-data-science-f94f1e27f90
On Tue, Apr 9, 2019 at 6:07 AM Ali Ilhan <[log in to unmask]> wrote:
> >
> > Dear Sandra,
> >
>
> Thanks a lot for your email. What you describe sounds like a standard ML
> implementation process (except the parts that you integrate design, which,
> I think is very important.) I follow a similar process in my consulting
> work with large scale child-welfare data. Also there are many good industry
> examples such as:
>
> https://www.conversionlogic.com/
>
> One of our PhD students designed an "intelligent" retirement investment
> advising "thing" (I say a thing, because it is hard to fit their system
> into a category), which provides (as the name implies) recommendations for
> your investments. see https://www.akillibes.com.tr/en/
> But in this case, we have a lot of data such as long-term bond yields,
> individuals' risk taking behavior etc. And design plays an important role,
> especially in structuring how users interact with this "thing". However,
> this is a much more circumscribed problem, where ML algorithm can be fed
> and trained with a relatively structured database.
>
> ML algorithms are pretty dumb in a human sense, as you also know. Somebody
> needs to train the data, needs to monitor the bugs, biases, needs to decide
> when is an over-fit or an under-fit is really a problem, etc. At least in
> my experience, these type of algorithms are much more harder to implement
> in the fuzzy front end of design,especially when there is very little data
> or the data has a lot of noise. Hence the concerns I voiced in my previous
> email.
>
> That said, (and as you have also pointed out) design and designers, I
> think, can be very important especially in terms of bringing human-centered
> concerns back into this whole overly commercial, profit oriented AI frenzy.
> Most AI algorithms are very biased (especially against women, minorities,
> etc. // you can see the reference list at the end of this article
>
> https://medium.com/thoughts-and-reflections/racial-bias-and-gender-bias-examples-in-ai-systems-7211e4c166a1
> )
> and I see room for a lot of valuable future contributions from
> socially-responsible, humanities/social sciences conscious (sorry I could
> not find a better description) designers. ''
>
> Cheers,
>
> Ali
>
>
> -----------------------------------------------------------------
> PhD-Design mailing list <[log in to unmask]>
> Discussion of PhD studies and related research in Design
> Subscribe or Unsubscribe at https://www.jiscmail.ac.uk/phd-design
> -----------------------------------------------------------------
>
-----------------------------------------------------------------
PhD-Design mailing list <[log in to unmask]>
Discussion of PhD studies and related research in Design
Subscribe or Unsubscribe at https://www.jiscmail.ac.uk/phd-design
-----------------------------------------------------------------
|