A winning recipe for machine learning

A winning recipe for machine learning


Machine learning is about automating the
ineffable. It is about unlocking a second mode of communication with machines. I
think that’s incredibly powerful. I’m Cassie Kozyrkov I serve as chief decision
scientist I’m with Google cloud what I’m really passionate about is turning
information into better action I work on decision intelligence it’s the applied
side of machine learning in a nutshell machine learning is about explaining
your wishes to computers in a different way from the traditional programming so
think about it like this traditional programming to get a computer to do
something for you you have to give explicit instructions to the computer
machine learning is asking for what you want with examples instead of
instructions the biggest challenge that I see marketers face with machine
learning is how to get started there’s the research side machine
learning and then there’s the applied side and the problem area for a lot of
businesses is that you don’t know which one they want to be the research side is
all about building machine learning algorithms for other people to use
that’s a lot like building kitchen appliances where is applied machine
learning is all about cooking and cooking at scale and that is a
completely different art just because you know how to build the microwave it
doesn’t make you a great chef and vice versa and so what they end up doing this
is the big mistake businesses and marketers will fail is they think they
need to build their own microwave in order to cook they don’t realize that to
apply machine learning to a specific business problem you don’t need to
reinvent the wheel and build that algorithm yourself there are warehouses
upland warehouses of these algorithms created by research teams and once it’s
created everyone else can take it off the shelf and start applying it
innovating in their models the key to success here is rigorous testing making
sure that what you asked for has been delivered keep testing the model on new
data to see if it’s working and bear in mind there’s only so much that the
hotest algorithm can help you with if your data isn’t very good put your
effort into making sure that your data your ingredients are relevant to the
task and if you have the most beautiful ingredients you can get away with a very
simple kitchen you

3 thoughts to “A winning recipe for machine learning”

Leave a Reply

Your email address will not be published. Required fields are marked *