Fiddler Announces Giga-Scale Model Performance Management with Deeper Understanding of Unstructured Models and Fine Discoverability to Launch New AI Initiatives

New NLP and computer vision monitoring and class imbalance capabilities will drive increased adoption of responsible AI

SAN FRANCISCO–(BUSINESS WIRE)–Fiddler, the pioneer in Model Performance Management (MPM), today announced major improvements to its MPM platform, including model ingestion at giga-scale, natural language processing (NLP) and computer vision (CV) monitoring, class imbalance, and an intuitive and streamlined user experience. With these new features, the Fiddler MPM platform is delivering a deeper understanding of unstructured model behavior and performance, and enhanced scalability, discoverability of rare and nuanced model drifts, and ease of use.

According to IBM’s Global AI Adoption Index 2022, a majority of organizations have not taken key steps to ensure their AI is trustworthy and responsible. Such steps include reducing bias (74%), tracking performance variations and model drift (68%), and making sure they can explain AI-powered decisions (61%). By operationalizing machine learning in a safe and trustworthy way and going beyond metrics to explain results, the Fiddler MPM platform makes more transparent AI models a reality.

The combined power of the new Fiddler capabilities will help enterprises across finance, banking, insurance, healthcare, defense, criminal justice, retail, and travel advance their AI efforts and deliver responsible models. Specifically, Fiddler has added capabilities to empower global organizations and Fortune 500 companies to:

• Monitor NLP and CV – Fiddler will now provide vector monitoring capabilities to help organizations gain a deeper understanding of more complex ML models in production that involve unstructured data such as text, images, embeddings, and intermediate model representations. For example:
  • Medical practitioners can gain greater accuracy in recognizing patterns as variants of illnesses change
  • Manufacturers can be alerted when the type of defects change in a defect detection CV model
  • Automakers can improve passenger safety by enabling cars to detect a change in weather patterns or a change in the road environment (e.g., construction zones)
• Address Class Imbalance – While the probability of a low-frequency event occurring is rare, it is also relatively hard to detect. Fiddler helps organizations discover nuanced model drifts in minority segments and surface fraud-like use cases in finance, retail, gaming, manufacturing, and education industries, among others. For example:
  • Gaming companies are notified when fraudulent transactions including subtle variations occur that could cost millions of dollars in potential revenue
  • Organizations can protect their advertising efforts by detecting higher-than-usual ad click rates to identify malicious behavior that impacts their bottom line
  • Ecommerce platforms are immediately alerted when a slight change in purchase patterns occurs within an attribute subset of a product category

Fiddler has also created a single pane of glass UI, providing data science and MLOps teams with command center-like visibility into every model’s behavior and performance across training and production. Teams can view, prioritize, and manage updates, alerts, versions, traffic, and drifts from a centralized homepage. Additionally, customers can now choose where to deploy their models to fit their company’s needs – whether that be on-premises, cloud, or both.

“AI model development is pivotal to our company’s ability to hire, vet, and match the world’s top software development and engineering talent with quality job opportunities,” said Jonathan Siddharth, CEO and Founder, Turing, a Fiddler customer. “We have invested in Fiddler to monitor models and ensure that our AI is consistently fair and explainable. Their platform and new enhancements demonstrate the path forward to making responsible AI a reality for every organization.”

“In order for organizations to be successful with AI, it is critical they ensure that underlying machine learning models are robust to shifts in the data, are not relying on spurious features, and are not unduly discriminating against minority groups,” said Krishna Gade, Founder and CEO, Fiddler. “Our new capabilities provide enterprises with even greater visibility throughout the entire model lifecycle. The ability to understand and explain unstructured data and discover rare but costly model drifts is game changing, and opens up tremendous AI opportunities across a plethora of use cases and a diverse set of industries. With Fiddler, organizations can deliver enterprise-scale AI models that power greater business results and responsible and fair outcomes for consumers.”

The new features will be demoed in a webinar hosted by Sree Kamireddy, Director of Product Management at Fiddler, taking place Tuesday, August 16 at 10 a.m. PT/1 p.m. ET. Register here to attend the webinar and read the blog for more information.

To learn more about Fiddler, visit: www.fiddler.ai

About Fiddler

Fiddler is a pioneer in Model Performance Management for responsible AI. The Fiddler platform’s unified environment provides a common language, centralized controls, and actionable insights to operationalize ML/AI with trust. Model monitoring, explainable AI, analytics, and fairness capabilities address the unique challenges of building in-house stable and secure MLOps systems at scale.

Unlike observability solutions, Fiddler integrates deep XAI and analytics to help you grow into advanced capabilities over time and build a framework for responsible AI practices. Fortune 500 organizations use Fiddler across training and production models to accelerate AI time-to-value and scale, build trusted AI solutions, and increase revenue.

For more information, visit www.fiddler.ai or follow us on Twitter @fiddlerlabs.

Contacts

Rachel Levy

V2 Communications for Fiddler

Fiddler@v2comms.com