Highlights:

  • Many businesses struggle to get their AI models to provide meaningful or usable results because of their limited, sparse, or one-dimensional datasets.
  • Dark Matter fits into data pipelines between feature engineering and model training and inference operations, according to Ensemble co-founder and CEO Alex Reneau.

Ensemble AI Inc. secured USD 3.3 million in the recent funding round to resolve issues around data quality and help businesses in developing robust AI models.

Salesforce Ventures led the round, with participation from Amplo, M13, and Motivate. They are supporting Ensemble because the company has developed a ground-breaking method of representing data to improve the efficiency of AI models without overloading them with information or building more intricate model architectures.

By assisting them in finding hidden correlations between their datasets, the business is enhancing AI models using machine learning techniques. The company claims that AI need access to more and higher-quality data to be able to tackle real-world problems. Many businesses struggle to get their AI models to provide meaningful or usable results because of their limited, sparse, or one-dimensional datasets.

Data scientists spend hours trying to repair their data to combat this, considerable progress has been achieved with more advanced AI model architectures, however, such initiatives demand large resources and technical knowledge that not every organization has.

To address these problems, Ensemble has developed a unique embedding model known as Dark Matter, which creates richer data representations for prediction tasks by utilizing an “objective function.”

According to the business, Dark Matter can perform a lightweight data transformation that allows it to understand the intricate, nonlinear relationships found inside datasets. It reduces the intricacy of these connections to a straightforward “data representation,” enabling programmers to create AI models of higher caliber that are capable of handling far more challenging issues.

Dark Matter fits into data pipelines between model training and feature engineering and inference operations, according to Ensemble Co-founder and CEO Alex Reneau.

“We’re able to enable customers to maximize their own data that they’re working with, even when it’s limited, sparse or highly complex, allowing them to train effective models with less comprehensive information. This foundational technology frees up data scientists to focus on experimentation and makes ML viable for problems previously unable to be modeled, unlocking new capabilities for our customers,” he added.

The startup thinks Dark Matter is a better option than synthetic data, which is frequently utilized by AI programmers to make up for sparse or low-quality datasets. It clarifies that although Dark Matter introduces additional variables, the mechanics remain essentially unchanged.

As synthetic data merely reproduces pre-existing distributions from Gaussian noise, no new information is formed in reality. The company clarified that there is no appreciable effect on predicted accuracy because the synthetic data only replicates the statistical characteristics of the existing data.

However, Dark Matter gains the ability to generate new embeddings that have essentially distinct statistical distributions and features, which leads to quantifiably better prediction accuracy.

According to Caroline Fiegel of Salesforce Ventures, Ensemble presents a viable solution that might hasten the adoption of AI. She clarified that because of problems with poor data quality and the possible usage of personally identifiable information, many firms are finding it difficult to implement AI models in production.

“When you peel that back and really start to understand why, it’s because the data is disparate. It’s kind of low-quality. It’s riddled with PII,” she added.

According to Ensemble, a number of early adopters have already used Dark Matter with encouraging outcomes in fields like biotechnology, healthcare, customization, and advertising technology. For example, according to the biotech customer, its technology was utilized to develop a model that more accurately forecasts virus-host interactions inside the gut microbiome.

In the future, Ensemble plans to accelerate its go-to-market strategy, grow its workforce, and develop new products with the capital raised in the funding round.