machine learning

Scala version of Collective Intelligence Euclidean distance algorithm

While reading the excellent book, Programming Collective Intelligence recently, I decided to code up the first algorithm in the book using Scala instead of Python (which the book uses). This is a Euclidean distance algorithm, and it provides one way to compare two sets of data to each other, and attempts to score the similarity between the data sets.

Without any further introduction (and assuming you have the Collective Intelligence book), here's the Scala source code for the Euclidean distance algorithm as described in the book:

Google “CLoud TPUs” available in beta

Google is making “Cloud TPUs” available in beta. From their announcement: “Starting today, Cloud TPUs are available in beta on Google Cloud Platform (GCP) to help machine learning (ML) experts train and run their ML models more quickly. Cloud TPUs are a family of Google-designed hardware accelerators that are optimized to speed up and scale up specific ML workloads programmed with TensorFlow. Built with four custom ASICs, each Cloud TPU packs up to 180 teraflops of floating-point performance and 64 GB of high-bandwidth memory onto a single board.”

Industrial robot costs to drop 65%

According to ARK Invest, the cost of industrial robots will drop 65% by 2025. As they write, “Combined with advances in machine learning and computer vision, this drop in costs should cause an inflection point in the demand for robots as they infiltrate new industries with more provocative use cases.” (Image from the ARK Invest website.)

AlphaZero generalized to learn more games by itself

From this Cornell University page, Google’s AlphaZero algorithm has been generalized to learn new games given only the game rules: “In this paper, we generalise this approach into a single AlphaZero algorithm that can achieve, tabula rasa, superhuman performance in many challenging domains.