Please enable JavaScript.
Coggle requires JavaScript to display documents.
The Just City: Machine Learning´s Social and Political Foundations,…
The Just City: Machine Learning´s Social and Political Foundations
Striking pattern of thefts
Thief acting systemically
Predictable
Patterns in large databases can get overlooked
Patterns in crimes within the cities
Algorithms
The Five "W"
When
Where
What
Why
Who
Computational Systems
Learns pattern-general similarities (proximity in space and time)
Identify pattern-specific similarities
Machine Learning
Predictive analytics
Detect more subtle and complex patterns
Training Data
Gamil
Spam Filters
Spam
Junk mail
Not-spam
Gmail considers the trade-off
False positives
False negatives
Improving police operations
Policing
Sending police only to observe crime/punish offenders
Doesn´t curb crime efficiently
They don´t prevent crime
Warrior mindset
They deal with
Homelessness
Mental health
drug crisis
Computers can cause significant improvements to whatever they´re added to.
Predictive policing
stop crime before it starts
Oversimplified model that doesn´t take into account the patterns it uses
They make racially biased predictions about policing models
Police activity and priorities
Police targeted urban minority communities
Discriminatory treatment
Policing algorithms
Machine learning makes sense of complex patterns and predicts hitherto inscrutable.
Data-driven algorithms
Priorities
Design chioces
Beliefs
Predicting the past
We can risk unleashing algorithms that make inaccurate or unfair decisions
Mental Health
People with mental illness have a higher probability of going to jail
Makes policing look like it is the problem
Most of the communities can't afford to have mental health services
People can´t adress their vulnerabilities
Highest-risk individuals
"These are not issues we can arrest or incarcerate our way out of"
Almendra Teycalco García García - A01746426