Over the course of Fiddler’s two and a half years, we’ve taken the time and effort to build strong partnerships with stalwarts in the industry who are heavily invested in AI explainability, model monitoring, and fairness. We’ve been fortunate to partner with the likes of Lockheed Martin & Amazon Alexa, and are honored to continue these strategic and innovative partnerships. Today, we’re proud to add two more important names to this list. Over the last few months, we’ve been fortunate to build strong relationships with the team at In-Q-Tel & join a FinRegLab research project. Below are a few notes on these important initiatives. We’re excited and look forward to working with both of these esteemed organizations.
Recommended ITech News: Pavilion Partners with Cyber Bytes Foundation to Provide High Speed Data Transfer and Storage Capabilities for Research and Innovation
In-Q-Tel & Fiddler
We’re thrilled to announce a strategic investment from In-Q-Tel, the not-for-profit strategic investor that accelerates the development and delivery of cutting-edge technologies to U.S. government agencies.
The need for Explainable AI and Model Monitoring is strong in all sectors of the economy including in government institutions. Based on the critical nature of their work, the need to explain and debug model predictions and analyze model behavior over entire datasets or for subsets of a dataset is essential to the deployment of models and their ultimate success. There is also a need to continuously monitor model performance once they are deployed and identify outliers in data and model predictions to continue to maintain high performance.
“The US Intelligence and Defense communities are increasingly relying on AI/ML in support of national security objectives, but ML models are only effective if they can be understood and trusted over time,” said A.J. Bertone, Partner at In-Q-Tel. “Fiddler’s platform efficiently combines both explainable AI and model monitoring in order to ensure that AI/ML systems function reliably and transparently. We are thrilled to partner with Fiddler in helping our government partners better manage and monitor their AI/ML capabilities.”
“Fiddler is the first startup to bring true explainability and continuous monitoring to the enterprise, enabling companies to effectively and responsibly manage their AI solutions at scale. We cannot allow algorithms to operate with a lack of transparency. We need accountability to build trust between humans and AI. Fiddler’s mission is to build trust with AI using a centralized Model Performance Monitoring solution that continuously monitors models and unlocks the AI black box with explainability,” says Krishna Gade, Founder & CEO, Fiddler.
Recommended ITech News: Jeeva Introduces World’s Lowest Power Wireless Chip
FinRegLab & Fiddler
We’re honored to participate in FinRegLab’s study with the Stanford University Graduate School of Business to evaluate the transparency and fairness of machine learning underwriting models. Here’s a short excerpt and you can read more about FinRegLab’s project on their website here.
This research comes as profound societal challenges have lenders and policymakers alike looking to new underwriting approaches with unprecedented urgency: a viral pandemic; a severe, uneven economic shock; and a mass movement for racial justice. Adopting machine learning to assess the creditworthiness of loan applicants is among the more promising options for improving current underwriting in consumer credit. In machine learning, stakeholders see significant opportunity to help improve the efficiency, fairness, and inclusiveness of lending. But the potential complexity and opacity of machine learning models pose risk, particularly with respect to their reliability in the face of changing economic conditions and questions about their ability to help overcome, rather than replicate, historical inequalities.
Despite the potential for machine learning models to improve credit risk assessment, the potential for these algorithms to reinforce biases or amplify other risks is a source of great concern. Many lenders are choosing not to use machine learning in credit underwriting – or only to use it in an indirect way that sacrifices much of its potential benefits – because of uncertainty about being able to manage explainability and fairness concerns with sufficient skill to satisfy themselves, their regulators, and other potential critics.
Recommended ITech News: Redwood Software Raises $383 Million Strategic Investment from Turn/River Capital to Expand Cloud Automation Solutions
