Data Digest: AI Bias, Machine Learning Lies, and Poisoning AI
Recent research has tried to create less biased algorithms, falsely represented ML accuracy, and shown how AI can be corrupted with bad data.
- By Upside Staff
- December 5, 2019
Researchers are testing a framework that could create algorithms smart enough to avoid bias.
Read more at Los Angeles Times
Some machine learning claims aren't just unbelievable, they're actually misleading.
Read more at Predictive Analytics Times
Experiments show that AI algorithms could potentially be undermined or controlled by specific changes to training data sets.
Read more at Wired