Winnie Soon, AI and Machine Learning
Winnie’s talk was both as informative as inspiring.
It gave a well needed and practical understanding of
the structure and implementation of ML as well as engaging
all of us in thought provoking critical discussions while
framing it within her own research, interests and projects.
Some parts I took note of during the talk and discussions were to name
a few: Winnie’s discussions of (in)visibility/the black box,
performativity and the language embedded in code (the notions of
read/write/execute) and generativity(chaos/order/structure) in software
studies. This along with the concepts of authorship and responsibility in
ML, AI and beyond.
Winnie also posed an open question that stayed with me and probably is
going to be something I would like to research further in the future. The
question was about prediction and imagination, what they are, how differ,
how they interact and influence each other. It made me think back to Le
Guin, about the importance of imagination for the ability to act and
create. How imagining is creating possibilities. Is it more likely that
a predicted event will occur? If so, what are the limitations of
predictive systems? Is it a way of imagining? Is it a way of shaping the
world? the future? Is a prediction only possible to be made if the outcome
has been understood, defined and imagined? Can computing help us bypass
mental and cultural blockages or are the outcomes of the algorithms
dependant on the imaginative capabilities of their authors/creators? Can
we create prediction systems that is not just reinforcing our beliefs but
extending our interaction with the world. Are the predictions only
reflections of our imagination? As always, more questions then answers.
But I do think it is an interesting and important topic.
The presentation of Winnies workings with Weibo, censorship, visibility
and erasure highlighted several important and interesting areas. One of
them (among many) were the ways of bypassing restriction, inventing new
ways of communication and trying to find loopholes and the never ending
creativity of how to spread thoughts, ideas and messages when they are
being suppressed. It was also interesting in the context of using ML and
similar tools to counter and reveal the destructive use of very similar
technologies. It opens up for (or maybe revitalize is a better word)
important discussions about tools and technologies and how they are used
//Jakob Jennerholm Hammar | 10/30/19