Segna Newsletter — 23 December 2021
What we’ve been reading
AI reveals that the Sahara actually has 1.8 billion trees and shrubs
AI analysis of satellite images has revealed that the Western Sahara contains an estimated 1.8 billion trees and shrubs. This is notable as “trees outside of forested areas are usually not included in climate models, and we know very little about their carbon stocks. They are basically a white spot on maps and an unknown component in the global carbon cycle.” Working with NASA satellite data, researchers trained an AI system to identify trees — “We counted hundreds of millions of trees in the desert alone. Doing so wouldn’t have been possible without this technology.”
DeepMind makes bet on AI systems that can play poker, chess, Go, and more
DeepMind’s Player of Games is a system, which, unlike other game-playing systems, it performs well as both perfect information games (e.g., Go and chess) as well as imperfect information games (e.g., poker) — this is important because, as in games, tasks like route planning around congestion, contract negotiations, and interacting with customers involve compromise and consideration of how people’s preferences coincide and conflict. Systems like Player of Games, which can reason about others’ goals and motivations, may pave the way for AI that can work more successfully with others.
Read the arXiv paper here.
We don’t need data scientists we need data engineers
KDnugget’s most popular article of 2021 provocatively points out that there are 70% more open roles in data engineering than data science. However, the high demand for data engineers reflects an evolution for the broader field. Frameworks like Tensorflow and PyTorch have done well to democratize the ability to get started with deep learning and machine learning. The bottleneck is now in helping companies get machine learning and modeling insights into production centers.
As we train the next generation of data and machine learning practitioners, let’s place more emphasis on engineering skills.
Engineers teach AI to navigate ocean with minimal energy
Researchers have developed a reinforcement learning system that enables autonomous drones to use ocean currents to aid their navigation — analogous to the way eagles and hawks ride thermals in the air to maneuver to a desired location with minimum energy expended. Surprisingly, the researchers discovered that their algorithm could learn navigation strategies that are even more effective than those thought to be used by real fish in the ocean — “we were initially just hoping the AI could compete with navigation strategies already found in real swimming animals, so we were surprised to see it learn even more effective methods.”
This puzzle challenge brings joy to the world of code
MIT Tech Review
Every December since 2015, Eric Wastl has published Advent of Code — an advent calendar for coders where everyday participants are faced with “playfully mathy problems and then write computer mini-programs that do the solving.” You can solve the puzzle using your programming language of choice (Python is most popular), although Wastl comments that a surprising number solve the puzzles in Minecraft.
It’s a great annual opportunity to tune up your programming skills, try a new language, or learn to code. Join in the fun here
If you’d like to receive this newsletter in your inbox, you can subscribe here.