Artificial Intelligence Vs. Climate ChangeJuly 26, 2019
Of all the many problems that face our globalized economies, climate change is the issue of the moment grabbing and holding the attention of business leaders around the world. Calls-to-action are coming from ever broader demographics as the scientific consensus has made the threat harder to ignore. Fostering economic growth without sacrificing the fragile balance of our ecology may one day be determined as the defining challenge of our age.
Alongside these developments, continuing advances in computational capacities and machine learning algorithms are having increasing impacts upon our daily lives. From manipulating the prices of services to adjust with predicted demand, to monitoring millions of online transactions and automatically detecting suspicious behavior, we are already entrusting machines with real-world tasks with actual effects. The question is: can we use these systems to rectify the pitfalls we have in our relationship with our environment, and what’s more, can we do it in time?
Fighting Fire With Data
A paper published last month written some of the world’s leading machine-learning experts has detailed many ways it may be of use. In the energy sector for example, there is an abundance of available data, but very little is being done in the way of making supply adaptive to real-time changes in demand. The benefits of this including reducing the expense of storing superfluous power, as well as reducing the overall amount of power produced. Google’s DeepMind was last year successful in trials of predicting wind turbine energy output.
Another notable example involved an AI that was trained to read scientific papers. In a study published this month in the journal Nature, a system had been fed over 3.3 million abstracts on the topic of material science. It then proceeded to create a neural network with its vocabulary of over 500,000 unique words. From this it produced predictions about materials which could be more effective at conducting electricity than those currently used. “It can read any paper on material science, so can make connections that no scientists could,” researcher Anubhav Jain said. “Sometimes it does what a researcher would do; other times it makes these cross-discipline associations.” Whilst it may take time to apply the lessons we can learn from the results of this application of machine-learning, it proves a convincing example of the potential it has to accelerate advancements in other areas of technology.
“The way that this Word2vec algorithm works is that you train a neural network model to remove each word and predict what the words next to it will be,” Jain said. “By training a neural network on a word, you get representations of words that can actually confer knowledge.” It is difficult to estimate the potential impact of AI on the efficiencies of current industrial systems. It seems we could be on the cusp of a transformation in how we relate to the data around us, which could hold untold benefits.
A Modeled Environment
But AI holds more than just the potential to mitigate some of the causes of climate change. As computational power increases, previously intractable problems might fall with the range of the algorithms ability to model virtual systems and predict results. Extreme weather events such as floods, droughts have complex causes. A great amount of work is being done to model these systems to predict emergencies in the future, as well as software to manage resource scarcities that are an inevitable consequence of natural disasters.
The steam powering the engine of machine learning and artificial intelligence is the rising sea of data that we our societies are generating at an alarming and accelerating rate. This new economy of data may determine the outcome of some of the biggest challenges faced in the world today, and it is tools like those discussed here which, through enabling a greater understanding of the world and our relation to it, will prove invaluable to our generation, and for the many to come.