Analytics Predictions 2019: Innovations for Ethical AI
Blog: Enterprise Decision Management Blog
By any measure, 2018 was the Year of Artificial Intelligence—“31 flavors of AI” and then some. For me, personally, 2018 was a period of intense AI-related creativity; my role as Chief Analytics Officer is tailor-made for my desires to rigorously create exciting new technologies in the context of solving the real-world problems of FICO customers. 2018 was the perfect year to do that, as AI and machine learning matured from magical curiosities into powerful business tools.
As 2018 draws to a close, I’m winding down the process filing my seventh and eighth US patent applications for the year. This brings my career total to 95: 45 patents granted and 50 pending. As one of my colleagues jokingly observed, I’m in the final sprint of “the race to 100 patents.”
It’s no coincidence that much of my recent patent work has laid a foundation for what I predict will be the tech industry’s biggest development in 2019: ethical AI. As a data scientist and member of the global analytics community, creating ethical analytic technology is very important to me, particularly in my role of serving FICO customers.
Racing to Contribute Toward Ethical AI
Here’s how my recent patent work unfolds to support the industry’s embrace of ethical AI.
Blockchain: Even though Bitcoin, the most famous instantiation of blockchain, had a lousy year, this underlying technology is on fire in novel business applications such as car rentals. In 2018 I turned my thoughts on blockchain inward, producing a patent application (16/128,359 USA) around using blockchain to ensure that all of the decisions made about an AI model are being recorded and are auditable. These include the model’s variables, model design, data utilized and selection of features, as well as the ability to view its latent features, and all scientists who built portions of the variable sets and model weights. The sum and total record of these decisions provides the visibility required to effectively govern models and satisfy regulators.
Explainable latent features: Another patent (15/985,130 USA) addresses the immaturity of the AI industry overall—which is painfully evident when the conversation turns to machine learning algorithms’ explainability. Specifically, for all of data scientists’ talk about deep learning being game-changing technology, questions about the details of learned patterns in a shallow or deep neural network are usually answered with quizzical silence, even at the largest companies. This is completely unacceptable for anyone who has to talk to a customer about the model or represent it to a regulator. One only need look at GDPR, which must be addressed in many of our customer’s locales, to understand that “I dunno” does not fly.
My patent for explainable latent features “explodes” a neural network model in a sparely connected multi-layered model, such that each hidden node can be explained succinctly. I recently talked about explainable latent features at an innovation workshop at the U.S. Federal Reserve, and the audience was very enthusiastic about building transparency into models in this way. Further, this capability has now been integrated into FICO® Analytics Workbench™, allowing FICO customer organizations to build these explainable models.
Bias removal: This ethical AI topic is little broader. It looks at restricting the type of data that would go into a model build, to prevent the introduction of bias—a conceptual cornerstone of ethical AI. I’ve filed two patents (15/981,755 USA and 15/819,338 USA) to facilitate decision-making on whether particular data and derived variables are suitable in a model or not. For example, a model that factors in a person’s height would be useful in calculating the production cost of a pair of blue jeans (which typically have the same price, irrespective of the inseam length), but not a loan applicant’s earning potential or credit worthiness.
All of these patents fit neatly under an umbrella of regulatory embrace of Ethical AI, which will accelerate dramatically in 2019 as governing bodies enforce the requirement to understand and explain the AI models set to power today’s financial decisions—including the billions of decisions powered by all of FICO’s products, from the industry-leading anti-fraud FICO® Falcon Platform to our latest AML and cybersecurity solutions.
The post Analytics Predictions 2019: Innovations for Ethical AI appeared first on FICO.