Digital Culture 2020 - notes

Revisit block 3: Commons + F/LOSS

Individual Brain-Dumping (5 mins on hackmd - writing and reflecting)
complexity: free (cost/freedom) - control - open source/accessibility/power dynamics - licenses - commercial/capitalist economy - care/feminist perspective
MX017/Group - Follow this group feedback sequence (2->4->1->3->5->2), provide a peer feedback of block assignment 3 (comment on their hackmd)

Automated and Predictive culture + assignment

machine learning (pattern recognition)
predictive culture
data-driven culture
Block 4 assignment

ML workshop

Focus on data processing (input/output)

Teachable Machine: https://teachablemachine.withgoogle.com/v1/

source code: https://github.com/googlecreativelab/teachable-machine-v1

Activity: Cont'

Prepare three set of gestures that can be captured by the web camera.
Each gesture has to be repeatedly trained by long-pressing the colored "TRAIN" button
when there is a new image input (a so-called "test dataset"), the learning/teachable machine can classify those gestures with various confidence levels, and then predict the corresponding output results.

Train the machine using three different sets of gestures/facial expressions, then observe the predictive results shown as various outputs.
Test the boundaries of recognition or classification problems, such as having a different test dataset, or under different conditions such as variable lighting. What can, and cannot, be recognized?
What happens when you only use a few images? How does this amount of training input change the machine's predictions?

MX015/group

By trying out Teachable Machine version 1 (https://teachablemachine.withgoogle.com/v1/), discuss what is your understanding of machine learning, what interest you during the tinkering processes? Reflect on possible critical questions/angles that we can raise.

What's learning here

"In machine learning, it is commonly understood that the style is learnt from training datasets through techniques to process and analyze large amounts of (natural language) data. As such, machine learning techniques such as "style transfer" rely on a process of generalization in order to identify patterns." - Soon & Cox, forthcoming

Neutrality

'this "pattern recognition" is clearly not a neutral process as it involves the identification of input data, and the "discrimination" of information. It is clear that there is other kinds of discrimination in such processes, such as inherent stereotypes in voice assistants or in online translation tools, and other examples that might include the AI chatbot Tay that was regarded as racist, or how facial recognition in Amazon software, and other smart systems demonstrate gender and racial bias. Understood this way, pattern recognition is not only about smoothing tasks and making accurate predictions in terms of technical operations but also political ones as it creates "subjects and subjection, knowledge, authority" as well as classification and categorization.'
- Soon & Cox, forthcoming

What's learning in machine learning?

Where is the data
How do they learn
How do they identify pattern
How do they make prediction

Learning algorithms #1

Supervised learning- This model is based on a training dataset with input/output pairs as expected answers. (binary/multi classification)

Kittydar (Kitty Radar)
http://harthur.github.io/kittydar/

Learning algorithms #1 - cont'

ImageNet (2009) by Fei-Fei Li: http://www.image-net.org/

12 hours of ImageNet (2019) by Nicolas Malevé Nicolas Malevé

Scaling/labeling/labour issues:

a dataset of 14,197,122 photographs from ImageNet.
organised into 21,841 synsets (categories), taken from a lexical database called WordNet.
at a speed of 90 milliseconds per image
traversing the entire dataset in a period of two months

Learning algorithms #2

Unsupervised learning - One of the common tasks is "clustering" (algorithms such as K-means and Hierarchical Clustering). The goal of this technique is to find similarities, providing insights into underlying patterns, and relationships of different groups in a dataset using exploratory and cluster analysis.

Learning algorithms #2 - cont'

The clustering of images based on "cuteness" and "curliness" in the Anatomies of Intelligence workshop by Joana Chicau and Jonathan Reus.

Learning ???

What could we learn about learning from the dynamics of machine learning?
how can we think productively about the fact that a generation of humans and algorithms are learning together to look at data (such as images)?
How does machine learning change our ways of understanding/seeing things?
Will our thinking and creativity be automated? and that our ability to determine our potential futures will become compromised by predictive algorithms?

Text reading 1

Mackenzie, Adrian . (2017). “Introduction: Into the Data” in Machine Learners: Archaeology of a Data Practice, Cambridge, Mass.: MIT Press, pp. 1-19
https://i.imgur.com/QEwKW9D.png

Q0. What are the things interest you in this section? and why?
Q1. what's the problem statement in the section?

Text reading 2

Computer vision algorithms are not mere technical improvements but intervene in the common understanding of what an image is, what it can do and whether it can be trusted. These developments have been achieved by emulating algorithmically the ways humans see, interpret and produce images.

Q0. What is an image?
Q1. Why the understanding of image is changed with computer vision algorithms/technologies? What's Malevéan's point of view?

Revisit MX014 - glossary

https://hackmd.io/@siusoon/dc2020-discussion

Digital Culture #11