March 24, 2016 | Written by: mlcullen
There are many limits to traditional analytics—problems need to be predefined, there’s no capacity for handling ambiguity, semantics for structured and unstructured data must be known, and interaction with the end user is through formal digital means. Cognitive computing however opens up possibilities where machines can learn new problem domain, reason through hypotheses, resolve ambiguity, evolve towards more accuracy, and interact in natural means. This creates vast opportunities for complex problems solving across all industries and sectors.
Take for example the increasingly complicated problem of energy. For the most part, the electricity grid is simple enough: a power station creates enough energy to run our homes and businesses, taking into consideration our fairly regular consumption patterns (i.e., more during the day than at night).
Yet, solar panels are increasingly being used as energy sources—sources that vary due to weather. On sunny days, when homes and buildings generate more electricity than they need, they can sell the excess back to the grid; on cloudy days, they need to supplement their own energy with energy purchased from the grid. To avoid energy outages and waste, as energy cannot be effectively stored, it is necessary to correctly predict the energy consumption considering the effect the weather has on it.
As more homes and buildings turn to the use of solar energy, it will become increasingly difficult to predict energy consumption. This type of problem has so many unpredictable variables, it requires an algorithm that improves itself as it goes.
It is exactly this type of complex problem that the IBM Research – Ireland lab is exploring by using cognitive computing techniques and proprietary high-resolution weather forecasting technology.
The goal is to help utility companies make better operational decisions, such as scheduling conventional generation or demand response. This will ultimately help integrate more renewable energy into the power grid.
And that’s just one example of the many innovations underway at the lab, where over 100 researchers are dedicated to helping cities not only collect and analyze data, but actively learn from and adapt to it, in order to make life for citizens easier and more efficient. “It’s our job to tackle the city problems nobody else can solve,” said Jean-Charles Gautier, who works in the lab on Business Development Strategy.
Rather than race to find short-term solutions, the lab is strategically selecting city challenges that other, smaller companies would struggle to solve. One team at the lab is working diligently to create cognitive buildings—structures that are aware of their inhabitants and learn to regulate light, temperature and airflow to suit their needs.
Another team is designing an Internet of Things platform that connects parked cars using various devices, including built-in computers, GPS and cameras, to detect gas leaks, locate missing objects, find parking spaces and help spot suspicious activities. Yet another team is using cognitive algorithms to create a holistic approach to care, by connecting medical, social and public records, which means that medical caregivers can quickly understand patients’ medical and psychological history, plus their network of friends and family, and their means for accessing care.
Each of these examples demonstrate the immense power for transformational change behind cognitive computing. You can learn more about the lab here, and cognitive computing here.