|日本語

Citadel Lens

Test your AI models and datasets against industry standards and improve AI performance.

AI Evaluation in Lens

LLM Evaluation in Citadel Lens - Citadel AI
Play Video about LLM Evaluation in Citadel Lens - Citadel AI
Computer Vision-Evaluation in Citadel Lens - Citadel AI
Play Video about Computer Vision-Evaluation in Citadel Lens - Citadel AI
Tabular Evaluation in Citadel Lens - Citadel AI
Play Video about Tabular Evaluation in Citadel Lens - Citadel AI

Challenges in AI evaluation

Citadel AI - AI Problem - Black Box Model
Black Box Models

AI predictions and failures are opaque and hard to debug. Lens provides an interactive interface to explain, inspect, and improve AI systems – opening the black box.

Citadel AI - AI Problem - Limited Test Suites
Limited Test Suites

It’s hard to comprehensively measure AI quality across metrics, data slices, robustness scenarios, and more. Lens’ built-in test suites systematically validate your AI against industry standards.

Citadel AI - AI Problem - AI Compliance
AI Compliance

Lens automates technical and process reports to support compliance against AI standards and regulations, such as the EU AI Act. Centralize your model registry and reporting across an organization.

Citadel AI - Lens - Diagram

Why you need Citadel Lens

Automatically test your AI models and datasets against industry standards and improve AI performance.


Lens helps engineering and compliance teams to measure AI quality based on industry best practices and international standards: robustness reports, explainability reports, fairness reports, data quality reports, and more.

Benefits

Citadel Lens - Actionable insights for better model performance

Actionable insights for better model performance

Easily identify areas of improvement with Lens model reports and dataset reports, and interactively debug a model in the Lens UI. For example, automatically evaluate a model’s robustness to lighting conditions in a specific geographic location.

Citadel Lens - Model-agnostic testing

Model-agnostic testing

Lens is designed to support models developed in any framework, allowing you to use the tools of your choice and evaluate all of your organization’s AI projects on a single platform.

Best-in-class XAI tools

Best-in-class XAI tools

Citadel Lens provides powerful explainability features to understand AI models. Lens can explain a model globally, regionally, and locally, such as counterfactual explanations (tabular models) and per-detection heatmap explanations (object detection models).

AI compliance reports

Citadel Lens supports AI developers and auditors with compliance against AI standards and regulations, such as the EU AI Act. Automatically generate, track, and approve technical or process reports against ISO standards and more.

The adoption of Citadel AI’s technology was a major turning point in the use of machine learning at Suntory. With Citadel AI, we were able to make agile improvements and achieve stable operation in our machine learning systems, successfully maximizing business efficiency.

Motoshige Takagi
Manager, Global System Department, 

Suntory System Technology Limited

Interested in learning more about Lens or Radar?

Let’s discuss how Citadel AI can improve your AI quality.

Benefits

Track all AI projects and reports across your organization from a single dashboard.

Automatically generate reports for your AI models and datasets based on international standards and industry best practices. (The reports shown here are examples.)

Fully customize Lens’ comprehensive test suites to your application – across data slices, metrics, robustness scenarios, and more.

Fully customize Lens’ comprehensive test suites to your application – across data slices, metrics, robustness scenarios, and more.

FAQs

How does Citadel Lens work?

Citadel Lens tests your AI models and datasets against industry standards to help improve model performance. After testing, Lens automatically creates reports: model quality reports, data quality reports, ISO technical reports, process reports, and more. These reports enable engineering teams to improve AI quality and compliance teams to decrease AI risk.

What testing does Citadel Lens actually provide?

Citadel Lens includes the world’s most sophisticated library of automated AI reports, enabling you to comprehensively test your AI models and datasets. Lens measures and visualizes all dimensions of AI reliability: explainability, robustness, fairness, bias, error analysis, backward compatibility, and more. These reports can also be fully customized for your specific use case.

What do I need to prepare to use Citadel Lens?

To use Citadel Lens, you’ll need a model that you want to test and a dataset to evaluate it against. Because Lens is model-agnostic, it can automatically connect to almost all tabular and image model formats, with LLM support coming soon. For a trial, we can also provide a sample model and dataset (or you can bring your own model).

We’re happy to discuss the best setup for your use case – contact us here anytime.

Is Citadel Lens secure?
We take security seriously at Citadel AI, and we’re ISO 27001 certified and GDPR compliant. Your data is encrypted in transit and at rest, and your Lens server can be hosted in a cloud or on-prem environment of your choice. For more information, visit our Trust Report or contact security@citadel-ai.com.
How much does it cost to use Citadel Lens?

We work with a wide range of users from startups to large enterprises, so our customers’ use cases and requirements differ significantly. Please contact us to determine the best pricing for your use case.

Request a Demo

Interested in a product demo or discussing how Citadel AI can improve your AI quality? Please reach out to us here or by email.