In other words, simple problems — Recently, I posted on when not to use neural networks. This time, I thought it would be nice to go the other way around and talk about why these fancy networks are so damn effective at vision and language problems. …

A guide on model selection and the inexistence of right choices — A while ago, a friend asked me for some guidance on picking AI models. At the time, most of his AI education was centered around neural networks and their many variants. …

Reading suggestions to keep you up-to-date with the latest and classic breakthroughs in AI and Data Science. — Today I return to my series of AI paper recommendations bringing you ten new AI paper suggestions (and many more links). My older readers should recall the first, second, and third lists. Now, nearly two years after, comes the fourth installment. This is a pretty opinionated list for those new…

State-of-the-art pose estimation with a built-in tracker at your fingertips — Recently, I was hired to develop an MVP that required real-time pose estimation. In previous years, the dominating technique was CMU’s OpenPose, which, despite being a bit cumbersome, works quite well once you manage to set it up correctly. Alternatives included CenterNet, AlphaPose, and PoseNet. Collectively, all these techniques lack…

On how error builds up on complex AI pipelines — In the previous article, I briefly mentioned how using AI sequentially is a nightmare. Whenever an AI is used as input for another, the individual errors of each model quickly add up to unacceptable levels or, simply put, catastrophic failure. Moreover, as you add more nodes to the chain, the…

Professional advice for data scientists out there — After nearly six years working with machine learning, AI, and, development, I list seven insights I acquired over the years with both successful and failed projects. The central theme here is how, despite learning about countless models and techniques, being an effective AI professional is all about avoiding complexity as…

But you definitively should to be a better data professional — Data science is a massive topic in its own right, which takes a lot of time to grasp, let alone master. So, not surprisingly, many universities and educational institutions now offer specialized data science education, detached from traditional courses. …

Learn how AI tools can solve non-AI problems — At a high level, differentiable programming is the paradigm of writing software that can be automatically differentiated. In plain English, we can quickly inspect how changes in the program inputs affect the program outputs. For instance, say we code an automatically differentiable car simulator. Using this system, we can simulate…

11 tips and tricks to make the most out of Keras and TensorFlow — Keras is a beautiful project. While TensorFlow and PyTorch used to compete for being the state-of-the-art, Keras aims at us, the professionals that need to get stuff done. We are okay with using last year’s models instead of betting on the next Transformer-killer or ResNet wannabes. Being Keras part of…