IBM AI EXPERIMENTS AND RESOURCES
IBM Watson Cloud
AI and Law Enforcement
AI and Sentencing Guidelines- Who Should Get Parole?Even the best risk assessments, such as COMPAS, yield probabilities, not certainties. That means they label as “high risk” some people who won't commit another crime and label as “low risk” some people who will. This simulation lets you sort offenders into risk categories based on the results of an assessment. Think we should rarely lock up anyone who wouldn’t reoffend? Set the “low risk” threshold high and the “high risk” threshold even higher. Have little tolerance for recidivism? Try the opposite. In the real world, policymakers have to strike a balance
Other Resources
Generative Adversarial Network Playground
https://gan-paint-demo.mybluemix.net/
A new kind of neural network, called a generative adversarial network (GAN), can be used create complex outputs, like photorealistic images. How do GANs do this, and can we influence what they produce? Try painting with a GAN below to collaborate and create your own custom image.
https://gan-paint-demo.mybluemix.net/
A new kind of neural network, called a generative adversarial network (GAN), can be used create complex outputs, like photorealistic images. How do GANs do this, and can we influence what they produce? Try painting with a GAN below to collaborate and create your own custom image.
AI Experiments from IBM Research