New York : A research team in the US has created a software that can quickly identify the information in cancer reports that would not only save time and work-hours but also reveal overlooked avenues in cancer research.
Much of the cancer-related data is drawn from electronic, text-based clinical reports that must be manually curated — a time-intensive process — before it can be used in research.
“The manual model is not scalable and we need to develop new tools that can automate the information-extraction process and truly modernise cancer surveillance in the US,” said Georgia Tourassi, director of the Health Data Sciences Institute at the US Department of Energy’s (DOE) Oak Ridge National Laboratory (ORNL).
After experimenting with conventional natural-language-processing software, the team found an answer via deep learning — a machine-learning technique that employs algorithms, big data and the computing power of GPUs (Graphics Processing Unit) to emulate human learning and intelligence.
“Our work shows deep learning’s potential for creating resources that can capture the effectiveness of cancer treatments and diagnostic procedures and give the cancer community a greater understanding of how they perform in real life,” Tourassi added.
GPUs, such as those in Titan, can accelerate this training process by quickly executing many deep-learning calculations simultaneously.