Advanced Computing in the Age of AI | Wednesday, July 24, 2024

Aporia Tackles AI Trust with $25M Series A Round 

Aporia, an observability platform for machine learning, has announced a $25 million Series A funding round, bringing its total funding to $30 million.

With the ever-expanding use of AI for business intelligence, data integrity and trustworthiness of predictions made by machine learning models have become important concerns.

There are notable examples of what happens when AI is used irresponsibly, including Zillow’s recent failure to accurately leverage its AI when forecasting home prices—a mistake that led to the company dissolving its Zillow Offers division, unloading 7,000 homes it had purchased, and laying off a quarter of its workforce.

Aporia’s machine learning model monitoring platform aims to empower businesses to trust their AI. Aporia offers a custom monitoring solution that integrates with existing infrastructure in minutes, which the company says can provide “full visibility of how models are performing in the real world,” according to today’s press release. Data teams can create customizable monitors that can detect biased predictions, unexpected changes in input data formatting, and the gradual degradation of a model’s performance. The platform then offers actionable insights to address these issues.

“Every organization using AI must have visibility on whether their ML models are accurate, equitable and fair,” said Liran Hason, CEO and co-founder of Aporia. “Data scientists require a tailor-made observability solution to ensure the safety of their highly specialized AI models, but it’s expensive and time-consuming to build one from scratch. Businesses that have embraced AI and are enjoying the benefits are eager to put safeguards in place to ensure they use their AI responsibly and don’t negatively impact their customers and business.”

Examples of existing data infrastructure into which Aporia says its platform easily integrates.

Demand is increasing for tools that address accuracy in AI and allow for greater visibility and explanation of ML models in production. Aporia reported a 600 percent growth in customers in the past six months, including New Relic, Lemonade and Armis.

Aporia’s platform now has new capabilities for explaining how models are making predictions. Blindly trusting black box AI models can be problematic, not only for the aforementioned example of what can happen with dubious models, but also for regulatory provisions for AI use that require an explanation of how a decision was made, such as with loan applications. Aporia’s platform can explain how models make predictions “by showing how much each input parameter contributed to the decision and how the prediction shifts with changes to the input data.”

Today’s Series A round saw participation from new investors Tiger Global and Samsung Next, as well as existing investors TLV Partners and Vertex. Aporia proposes to use the funds to triple its headcount this year and expand its U.S. presence while increasing possible uses cases for its platform.

“Aporia has demonstrated unbelievable growth since its launch and has amazing momentum, quickly becoming a leader in the space of ML observability,” said John Curtius, Partner at Tiger Global. “Executives at global enterprises understand the benefits of artificial intelligence and how it's impacting virtually every industry but the risks keep them awake at night. Aporia is positioned to be the solution every organization turns to for ensuring their responsible use of AI.”

A screenshot of the platform's dashboard shows customized models management alerts. Source: Aporia.