Artificial Intelligence (AI) is the theory and development of computer systems able to perform tasks that would normally require human intelligence. This term was introduced by American computer scientist, John McCarthy in 1956 during a summer research workshop at Dartmouth College. The research project proposal stated “The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.”

It took early researchers years to realize that hard-coded algorithms alone were not sufficient for machines to solve these problems previously reserved for humans. They realized they needed to understand how humans “learn” to perform complex tasks. Humans do this by learning the rules and exceptions, looking at a variety of examples, and by practice and repetition they work their way up to more complicated and complex tasks.

The idea behind Machine Learning (ML) is similar in a sense that some codes and algorithms enable computers to learn and improve performance of certain tasks through iterations without being explicitly programmed to do so.

Deep Learning (DL) is an advanced area in ML. It can recognize the relevant features required to solve a problem through automatic feature extraction. The computer uses multiple layers to progressively extract high-level features from the raw input.

How Are Water Resources Engineers Using DL?

Water resources engineers and scientists have started applying DL in their projects. According to a recently published paper titled A Transdisciplinary Review of Deep Learning Research and Its Relevance for Water Resources Scientists deep learning could help hydrologists in many areas, including:

• Discovering and understanding the interactions between hydrology and systems that were previously considered exogenous to hydrology
• Extracting abstract but useful information from the extraordinary volume of diverse data sets that were not available previously and measure relevance
• Exploring and discovering complex relations between variables
• Generating realistic scenarios by capturing complicated data distributions
• Providing an approach that enables engineers and scientists to model difficult processes

I’m fascinated by AI. I’m not even a mediocre programmer, but I have the luxury of working with people on the resilience solutions team who have amazing skills in this field. By expanding our capabilities in this area, we’re hoping to provide more AI and ML services that complement and improve traditional hydrologic and hydraulic analysis and design.