Machine Learning in Elixir

Thinking in Tensors

Data is arguably the most important component in a machine learning system. Even with the most complex, cutting-edge model, you still need significant amounts of high-quality data to train a quality model. Bad data inevitably leads to bad models.

Researchers are generally divided into two schools of thought on the importance of data versus the importance of models in machine learning. Some researchers believe in a data-centric approach. The argument for a data-centric approach is that large amounts of high-quality data are the key to engineering quality machine learning systems. While other researchers believe in a model-centric approach. The model-centric approach focuses on spending more time researching new models to ...