Announcing the integration of “Citadel Lens” and “MAGELLAN BLOCKS” to ensure Trustworthy AI
We are pleased to announce that Citadel AI Inc. and Groovenauts, Inc. have agreed to introduce a platform integration between Citadel Lens, an AI model testing product, and MAGELLAN BLOCKS, an AI model development platform.
Over 50% of companies in Japan utilize AI, and this number is increasing every year (*1). It’s important for these enterprise AI systems to be continuously evaluated and monitored throughout their lifecycle, including when models are retrained after the initial deployment.
Citadel Lens, our model testing product, automatically stress tests AI models to reveal reliability issues, including explainability (*2), model bias (*3), backward compatibility (*4), adversarial robustness (*5), and more.
On the other hand, Citadel Radar, our model and data monitoring product, automatically detects and protects AI systems against real-world data problems that cause performance deterioration in production, such as data drift, model drift, and outlier/invalid data points.
Groovenauts’ MAGELLAN BLOCKS AI platform enables users to easily build AI models on numerical, image, and text data. Many Groovenauts customers use this platform since it does not require programming, and smoothly handles the AI lifecycle across data collection and processing, model development, prediction, and accuracy validation.
This integration between MAGELLAN BLOCKS and Citadel Lens will enable AI developers to measure the quality of AI models in an intuitive and objective manner. This will help increase the trust, transparency, and explainability of AI applications. Both companies will continue to work together to provide solutions to promote and advance AI, including further collaboration on Citadel Radar.
Citadel Lens verifies AI quality
Citadel Lens automatically stress tests AI models, creating Model Reports that reveal performance and reliability issues. Users can easily analyze a single model, or compare a group of model versions, with actionable insights to improve performance. Citadel Lens ensures that AI developers can deploy new models with confidence.
(*1) From “2022 Japan Artificial Intelligence and Data Analytics: Enterprise User Survey” by IDC Japan K.K., an IT specialist research firm. https://www.idc.com/getdoc.jsp?containerId=prJPJ49020822
(*2) Explainability refers to being able to provide human-interpretable explanations of individual model predictions, and overall model behavior.
(*3) Model bias refers to inconsistent model performance across different sub-groups of a population. It can occur when training data is insufficient for a particular customer group, etc.
(*4) Backward compatibility refers to a new model version having predictions that remain consistent with those of a previous version, and minimizing the number of new errors.
(*5) Adversarial robustness refers to the ability of a model to withstand external attacks aimed at misprediction and mislearning, such as input data that is designed to cause mispredictions.
About Citadel AI Inc.
Citadel AI is an MLOps startup based in Tokyo, and raised seed funding from UTokyo IPC and ANRI in 2021. The co-founder & CTO was the former product manager for ML infrastructure at Google Brain, and is currently leading Citadel AI’s product and engineering teams. In April 2022, the company entered into a capital and business alliance with Suntory Holdings Ltd.
About Groovenauts, Inc.
Groovenauts is a technology company that develops and provides MAGELLAN BLOCKS, a cloud platform that enables the use of AI and quantum computing, based on advanced technical innovations and engineering capabilities. The company is also the first in the world to commercialize quantum computer applications. With a vision of “contributing to the realization of a prosperous and humane society,” the company promotes the utilization of advanced technologies such as quantum computers and AI to expand the potential and opportunities for the future of society.
Inquiries from the media should be addressed to
Citadel AI Inc.
Contact: Mr. Kobayashi
Contact: Ms. Kaneda (Public Relations)