Truera allows data scientists to keep a tab on the workings of the Machine Learning model.

Data scientists have long suffered from the impervious ‘black box’ that surrounds a machine learning model. That’s because data is crunched by the algorithm and decisions spit out. But what went on in between the input and the output remains opaque.

“Particularly for neural networks, where input data can undergo complex transformations in multiple layers of the algorithm, the model can become vastly complex and behave in unpredictable ways,” says The Lancet.

But help is at hand – Truera’s platform allows data scientists to analyze machine learning to improve model quality, address unfair bias, and build trust. (Intrado GlobeNewswire)

Truera’s Model Intelligence platform

The startup’s Model Intelligence platform is an industry-first with Fortune 100 customers in banking and insurance already using it.

Here’s Truera’s blog on why machine learning (ML) needs Model Intelligence. “Data scientists do not have sufficient visibility into models to effectively explain, analyze, and improve model quality during development and in production. Business users and risk and compliance teams do not understand ML models and their predictions deeply enough to assess model risk, guard against harms like unfair bias, and effectively use models to make decisions.”

COVID-19 pandemic

“Black box models raise societal concerns about fairness, bias, and transparency,” said Truera in a statement. “It’s harder to maintain black box ML model performance and trust over time when new data changes from the training data used to create the model. It’s a challenge called ‘concept drift’ that is top of mind during the current Coronavirus pandemic.”

Anupam Datta, Shayak Sen, and Will Uppington founded Truera in 2019. They wanted to create a model intelligence solution to solve the industry’s black box problem and make sure AI was implemented responsibly. Their efforts are termed as “AI Explainability” technology.

Datta and Sen are from Carnegie Mellon, while Uppington has previous experience in a startup. He is well aware of how a lack of model intelligence negatively impacts model development, selling, and customer success.

Truera has emerged from stealth and raised $5.1 million in a funding round led by Greylock. Additional investors including Wing VC, Conversion Capital, and Aaref Hilaly also participated.

How it works

The Truera platform uses AI.Q technology, claimed to be the best enterprise-class AI Explainability technology in the market.

It performs sophisticated sensitivity analysis that enables data scientists and non-data scientists to understand exactly why a model makes a prediction.

This understanding leads to higher quality, explainable models that achieve measurable business results, address unfair bias, and ensure governance and compliance.

Related Story:     An AI Bot Will Design the Next-Gen AI Chips