The Hungry Cat

The new Titans architecture from Google Research is turning heads in the AI community, and for good reason. By elegantly integrating a learnable neural memory module into the core model architecture, Titans takes a significant step towards AI systems that can accumulate and synthesize knowledge over extended interactions [1]. It’s an impressive technical achievement, but it’s important to recognize that it’s not a complete solution to the challenge of long-term memory and learning in AI.

Purpose

A practitioner’s guide to implementing a XGBoost model with hyperparameter optimization using the Dask API for distributed computing. This article will limit its scope to a working script running on a yarn cluster. A official guide on the Dask API for XGBoost is here. We will limit this to simply building the model and finding the best hyperparameters for a XGBoost classification model. We can extend this to a XGBoost regression model as well.

I’ve now managed to keep a thriving planted tank for more than two years. It is quite rewarding to have a thriving ecosystem in a waterbox. I’m going to make an attempt to share some of my learnings as i have fumbled to maintain balance in this man made contraption. Should you find yourself in a similar journey, I hope you find this useful .

One odd specification - An outdoor aquarium

The planted tank needed to be outdoors. The limited indoors space mean’t I could only salvage what space remained in the patio. I assure you, several measurments were taken to ensure a habitable ecosystem and it is safe for the water beings. I’ll write another article to highlight parameters that you need to measure before you begin a planted tank.