Elizabeth Gibney’s Nature article, Could machine learning fuel a reproducibility crisis in science?, is an intriguing exploration about reproducibility in disciplines that use Machine Learning with a particular focus on computational reproducibility. The challenges of training data from the same period or even including data in both training and evaluation data, or data leakage, are […]
Tag Archives: machine_learning
Causation, Correlation, and Method
Another short post but John Naughton’s latest column in the Observer is another one for book marking, Yes, DeepMind crunches the numbers – but is it really a magic bullet? . The computational achievements are there but the underlying question about how one understands it still need the scientific method. I might argue that, in […]
OpenAI codex
I was sent the link to the OpenAI Codex coding demo on YouTube, which was a lot of fun and interesting. It makes me think of the no code movement. At one level, I really enjoyed it and I like the fluency of the API that is being used. What worries me slightly is that […]