Guiding Eyes for the Blind keeps extensive records on how dogs are bred and raised and needed to shine a light on the dark data for a clearer view of why dogs succeed or fail.
Using natural language processing (NLP) on structured and unstructured data, the system is trained to find correlations to success among myriad genetic, health, temperament and environmental factors.
10% point expected increasein graduation rates, helping the organization meet growing demand for guide dogs
Reduced costs expectedthrough greater efficiency in the guide dog program, from breeding to placement with a human in need
New insights into the personalitiesof dogs and matching to the volunteer puppy raisers
Business challenge story
Unstructured data keeping insights hidden
Guiding Eyes spends USD 50,000 on each dog that goes through its 20-month training program, producing exceptional service animals that have been bred, selected, socialized and trained for high loyalty and intelligence. But the program has a modest success rate, with less than half of 500 puppies graduating each year. With demand for service animals on the rise, Guiding Eyes needs to better understand what makes dogs succeed and boost graduation rates. Though the organization has spent years collecting detailed information from its program, including genetic maps, medical records and questionnaires filled out by trainers and foster families, it never had the ability to properly analyze the data. The nonprofit needed a way to tap into large volumes of unstructured data and unlock meaningful insights that could help improve its program.
Dramatic advances with cognitive computing
Guiding Eyes is working with researchers at San Jose State University (SJSU) to make unprecedented advances in the training of service animals, using cognitive computing to uncover meaningful insights about the variables that contribute to a guide dog’s success. The organization began by moving half a million medical and genetic records and more than 65,000 temperament records to the cloud, making years of structured and unstructured data available for sophisticated analysis. Now researchers are using advanced cognitive technology to evaluate key characteristics of the dogs so that it can more effectively pair animals with people. They are also training the system to recognize correlations between certain markers or clues in the text and a dog’s level of success. The organization can correlate these insights with genetic and medical data to gain a deeper understanding of how multiple variables impact the ultimate success of the service animal. The solution learns from an ever-expanding corpus of knowledge, constantly revising and updating its understanding of the interrelationships among genetics, health, environment, personality and training approach.
Program efficiency improved
The Watson AI solution is expected to help Guiding Eyes improve graduation rates by 10 percentage points—from 49 percent to 59 percent—and meet the growing demand for guide dogs. The organization also expects to reduce costs by improving the efficiency of the program, from breeding and selection to training and placement. This represents a dramatic transformation in the way dogs are bred and trained, with more accurate insights into their personalities and temperaments.
Game-changing outcome—Using natural language processing (NLP) to examine personality traits, Watson AI is transforming the breeding and training of guide dogs into a more exact science, finding the unseen correlations among canine genetics, health and temperament—and the humans who help train them. These capabilities help Guiding Eyes improve its breeding and training processes, ultimately increasing graduation rates for the dogs it raises.
Before-after impact—Much of the organization’s valuable information is captured in questionnaires filled out by trainers and foster families—unstructured data that, in the past, one person had to collect and synthesize manually, which meant that valuable insights went unnoticed. Now, with linguistic analytics and NLP, the system discovers hidden connections in what used to be dark data, helping the organization optimize its program for higher success rates.
Systems of cognitive discovery, engagement or decision making—The solution creates a system of cognitive discovery and decision making by finding meaning in large volumes of unstructured data, helping Guiding Eyes make small but consequential improvements in its program and find unexpected insights about how dogs perform under different circumstances.
Guiding Eyes for the Blind
As a 501(c)(3) nonprofit entity, Guiding Eyes for the Blind provides guide dogs to people with vision loss as well as service dogs for children with autism. The organization is passionate about connecting exceptional dogs with individuals and families for greater independence. Throughout its 70-year history, the organization has trained and graduated more than 7,300 guide dogs.