⏱ Last two days to subscribe to TheSequence with a unique 60% discount. Share with your colleagues! And thank you for your constant support ⏱

SUBSCRIBE FOR $20/year

💡 ML Concept of the Day: A/B Testing for ML Models 

Continuing our MLOps series, today we would like to discuss the topic of A/B testing in ML models. A/B testing is a well-established practice in modern software applications, but it’s not trivial when it comes to ML pipelines. After all, testing an ML model does not only involve testing the model itself but also the corresponding datasets and hyperparameters. From that perspective, A/B testing in ML models is relatively different and more complex than in traditional software applications →read more with the subscription

🔎 ML Research You Should Know: How Meta AI Uses ML A/B Testing for Improving its News Feed Ranking 

Meta News Feed is a massively content selection engine that powers platforms such as Facebook and Instagram. Such a large system requires a sizable number of online A/B tests to determine relevant content for a given user. However, at that scale, traditional ML A/B testing techniques such as Bayesian optimization can be challenging due to resource limitations →read along to know what alternative they implemented to improve the news feed ranking

🤖 ML Technology to Follow: Weights and Biases is one of the Top ML Experimentations Platforms on the Market 

Why Should I Know About This: As a data scientist, be prepared to spend a lot of hours optimizing hyperparameters configurations. The toolsets for this task are still in very early stages, and W&B is one of the leaders in this developing area of the ML markets →more