A $200,000 competition challenges developers to design a program to automate management of oil and gas pipelines.
When you think about cognitive computing, you think of computers that can learn to solve the same complex problems that people solve – like IBM’s Watson that won on the TV show Jeopardy. While deep-learning software programs are available today, a commercial solution does not exist that allows a computer to learn how to scan technical documents, identify data patterns and populate databases that can be applied to analytical models for the energy industry.
To help fill this gap, Dynamic Risk is excited to announce the Cognitive Computing Challenge – a $200,000 incentive prize for developing real-time cognitive software that combines data mining and pattern recognition techniques to mimic the way the human brain works.
The goal of the challenge is to create a truly literate computer. We are asking engineers not only to find a solution to this challenge and invent a new field of technology, but also to join a global community that is dedicated to making radical innovation possible through incentive prize models. Rather than requiring a developer to code an application that anticipate every possible scenario, the desired solution should be interactive, contextual and adaptive so that models can sense, predict and infer what the user needs.
In the oil and natural gas industry, pipelines are a key component of the transportation network. Dynamic Risk helps many of the largest oil and gas operators in the world manage the safety and reliability of these and ther assets. Pipelines traverse our fields and forests, run down our streets and, in many cases, run through our backyards. They operate safely the overwhelming majority of the time, but when they do fail, the results can be catastrophic.
We believe that by building a cognitive computing system that can receive, understand, interpret and infer information about the condition of a pipeline, we can dramatically change the industry and bring pipelines to an even higher level of safety and performance. We think the public deserves and expects that of us.
Big data for pipelines can include data from millions of inspection locations collected on just a ten-mile section of pipeline. There are tens of dozens of data sets that get considered for every foot of pipe. If we can bring that unstructured data together for real-time analysis, we can offer solutions that have the potential to change the industry.
In one scenario, a system might combine data on heavy rains in an area with inspections reporting thinning on a section of pipe and a more corrosive mix of liquids flowing through it. The system could decide to shut down a section of pipe until repairs are made. This scenario, once thought to be a far off reality, is possible now with cognitive computing.
Dynamic Risk is launching its Cognitive Computing Challenge on the HeroX Web site because the first step in our vision is creating a solution that can learn and adapt to receiving data in multiple forms and prepare it for advanced analytics. By reaching out to the broad innovator community, we dramatically increase the likelihood of solving this difficult challenge.
Breakthroughs often come from people outside of an industry because they are without constraints or biases and look at problems differently. We are excited to see what ingenuity comes to light. The challenge invites submissions from teams until to December 31, 2015. Submissions will be judged in January 2016.
--Trevor MacFarlane is President of Dynamic Risk, a pipeline integrity management provider.