The process of selecting and training a seeing-eye dog is a costly, timely endeavor for Guiding Eyes, a nonprofit that trains canines for blind and visually impaired people. Training for one dog takes nearly two years and cost up to $40,000. Fewer than 37 percent of puppies complete the program successfully. The puppies need to have the right health profile, temperament and ability to interact with humans, but many of the necessary traits only reveal themselves with time and training.
SJSU Computer Science Professor Chris Tseng and his students have been working with Guiding Eyes on data analysis to help the nonprofit find a pattern that predicts which puppies are most likely to complete their training successfully. Through its Canine Development Center, the nonprofit has collected health records of more than half a million dogs and 65,000 temperament records that they have migrated to IBM Cloud.
Tseng’s students are using IBM Watson Personality and Natural Language Processing on IBM Bluemix, to analyze the vast amounts of data. By May, the group is hoping to establish a process for identifying data patterns and correlating traits, characteristics, environmental conditions, and personalities – of both dogs and trainers – to help improve Guiding Eye’s dog graduation rates, and to better match young dogs with trainers and ultimately owners.
“Guiding Eyes, is a great example of how IBM Cloud can help organizations innovate new business models and processes that were heretofore unthinkable,” said William Karpovich, General Manager, IBM Cloud Platform. “Through the IBM Cloud, Guiding Eyes is now able to advance even further its critical work in breeding, raising and training service dogs for those in need.”