Modulos presents its Data-Centric AI platform at #MWC22

ZURICH, SWITZERLAND, February 24, 2022 — Modulos AG is presenting its unique Data-Centric AI platform designed to speed the development of trustworthy AI applications and help companies comply with forthcoming AI regulatory frameworks.

The Modulos platform is designed to serve users that need reliable machine learning (ML) models that can be easily integrated into their business activities. The platform ensures fairness, accuracy, and robustness during development and deployment. As machine-learning models are built, humans are kept in the loop, ethical considerations are always paramount, and there is a focus on explainable decision-making throughout the journey.

Iterating data is the next frontier in improving the performance of AI applications while the search for the most optimal solution is delegated to automated machine learning (AutoML). The Modulos Data-Centric AI system provides step-by-step recommendations that guide users as they address flaws in their datasets that could undermine the accuracy or fairness of their final model.

As our CEO Kevin Schawinski puts it: “The tools for building machine learning models have gotten so good that what limits their performance is the data we use to train them. Bad data imprints its flaws on any model trained on them, leading to a loss of accuracy and fairness.”

The powerful implications of the unique Modulos methodology are already clear. Early test results have shown that by simply addressing the most important 10% records of a dataset, an 80% overall accuracy score was immediately achieved. Cleaning data randomly is not effective and can be an expensive exercise that ultimately harms performance. Modulos Data Quality Management and Modulos AutoML can provide maximum model performance while minimizing the costs of data preparation.

“If the previous decade of ML/AI system research was driven by scalability, efficiency, and automation, we believe among the defining challenges of ML/AI systems for the coming decade is the need to manage, facilitate, and enforce trustworthiness,” explains Modulos co-founder and ETH Zürich professor Ce Zhang.

Modulos can be found in the Swiss Pavilion (located at 5C41)—which will feature the most prominent innovative Swiss startups—at the Mobile World Congress 2022 in Barcelona. Visit us to experience a live demo of the Modulos platform and engage in serious discussion regarding AI-related matters with our on-site experts.

More formal meetings with Modulos can be arranged at our booth (5C41). Alternatively, our email address is [email protected]

About Modulos

Modulos AG operates in the AI software industry. The platform is designed to make the creation and deployment of trustworthy AI applications which are safe, reliable, transparent and fair, accessible to non-ML experts.

Modulos takes a revolutionary approach by centering the AI journey on data, rather than models, our so-called Data-Centric AI (DCAI). The search for good models can now be automated with our commercial AutoML, while the factor that limits model performance is now the data itself.

We solve this by providing actionable insights on how to address shortcomings in datasets—such as dirty labels, outliers, missing or incorrect labels, and missing and dirty feature values—to instead build ML models that result in desired performance outcomes. This new data-centric approach is based on cutting-edge research done by leading AI researchers, including our co-founder at ETH Zurich, one of the world’s top AI research.