Blog | super.AI

We explore the final two areas where the concealed complexities of machine learning can create costly surprises: configuration and R&D debt.

Messy real-world input data can quickly wreck havoc on a machine learning (ML) system that isn’t adequately prepared. Here, I show you how to stay ahead.

In this article we talk about how the algorithms ML models learn are determined by input data, rather than by a programmer’s handwritten rules—has profound system-level performance effects in ML and introduce the topics of covariance and prediction shift detection.

What is sentiment analysis and how does it work? We show you can apply this ML technique in your business.

Santa reached out to Canotic to help him build some machine learning models to make his work a little easier. Here's how we helped.

95% of the code in an ML model is actually just “plumbing”: code that handles configuration, feature extraction, monitoring, analysis, resource management, serving production models, etc.. In this post we talk about glue code, processing spaghetti code and how to tackle dead experimental code paths.

AI is transforming agriculture. We explore use cases showing how machine learning is applied in the industry to increase yields and resourcefulness.

In ML systems, data dependencies carry a lot of depth but are more difficult to predict. In this article, we explore the importance of quality input data and introduce a new danger: feedback loops.

The second costly surprise: machine learning often evolves into a complex web of interlinking systems. Changing anything in the system can have wide-reaching and potentially negative effects far from the source of the change. This post explores how to avoid the problems such as a system poses.

Real-world applications of ML require extremely high accuracy to provide a consumer benefit. Increasing accuracy is a balancing act between bias and variance. To minimize both, and thereby increase accuracy, ML models requires plentiful high quality training data.