In history and in literature, man has contemplated the idea of machine intelligence with a mixture of fascination, wonder and dread. For every Robby the Robot of Forbidden Planet, a wise and helpful guide, there is a HAL or Terminator to darkly hint at things to come.
The theoretical combination of high intelligence, emerging consciousness and zero emotion is the stuff of nightmares. HAL is the archetypical example, an advanced technology thoroughly aware of his intellectual superiority, yet cool as a cucumber.
“I am putting myself to the fullest possible use, which is all I think that any conscious entity can ever hope to do,” he confidently claimed early in 2001: A Space Odyssey, later demonstrating a merciless instinct for self-preservation.
For now, artificial intelligence is assuredly following its programmed credo to serve man. The latest example of this comes from my research speciality, where a team of engineers from Cornell University is using AI to create a more precise model for estimating air pollution in urban areas.
This new AI-based technology reviews a vast range of data, synthesizing and summarizing the information that is most relevant to decision-makers.
“Older models to calculate particulate matter were computationally and mechanically consuming and complex,” explained senior author Oliver Gao, Howard Simpson
Professor of Civil and Environmental Engineering in the College of Engineering at Cornell University. “But if you develop an easily accessible data model, with the help of artificial intelligence filling in some of the blanks, you can have an accurate model at a local scale.”
The study, “Developing Machine Learning Models for Hyperlocal Traffic Related Particulate Matter Concentration Mapping,” focuses on particulates generated by vehicle exhaust. Traffic-related pollution was measured across the New York City metropolitan area.
In perhaps an inadvertent nod to great science fiction acronyms, the team gave their most precise predictive model the vaguely unsettling name, Convolutional Long Short-term Memory, or ConvLSTM. The model trains computers to crunch the numbers, think about what it observes, and offer conclusions.
“Our data-driven approach — mainly based on vehicle emission data — requires considerably fewer modeling steps,” lead author Salil Desai noted. “Instead of focusing on stationary locations, the method provides a high-resolution estimation of the city street pollution surface. Higher resolution can help transportation and epidemiology studies assess health, environmental justice and air quality impacts.”
Science Daily heralded the breakthrough: “The equations use few inputs such as traffic data, topology and meteorology in an AI algorithm to learn simulations for a wide range of traffic-related, air-pollution concentration scenarios. … Ambient air pollution is a leading cause of premature death around the world. Globally, more than 4.2 million annual fatalities — in the form of cardiovascular disease, ischemic heart disease, stroke and lung cancer — were attributed to air pollution in 2015.”