News

Synthesized launches AI-driven platform to detect and remove biases in data

Wednesday 11 November 2020 15:00 CET | News

Synthesized has released the first publicly available solution to detect and remove biases in data, the Community Edition of its data platform for Bias Mitigation

Released as a freemium version, the offering incorporates AI research and techniques to enable financial organisations to identify potential biases within their data and immediately start to remediate these flaws.

The platform was designed by the UK-based firm to understand a wide array of regulatory and legal definitions regarding contextual bias. It can automatically identify bias across data attributes like gender, age, race, religion, sexual orientation, and more. 

Synthesized is making the capability available immediately, requiring no coding or deep technical expertise to get started. Financial users simply upload a structured data file, like a spreadsheet, to kick off the analysis process. The data platform could be used in finance to create fairer credit ratings.

Beyond this analysis and bias detection, the platform also offers another feature: to automatically remove the biases present in an entire dataset in a process called rebalancing. 

While there are a number of existing, limited techniques to rebalance biased data, Synthesized has developed a proprietary algorithm within its platform that is quicker and more accurate. The AI-driven platform has the ability to make randomised changes, at scale, to an original, biased dataset to construct a new, entirely synthetic dataset. With the generation of synthetic data, Synthesized’s platform gives its users the ability to equally distribute all attributes within a dataset to remove bias and rebalance the dataset completely. Users can also manually change singular data attributes within a dataset, such as gender, providing granular control of the rebalancing process. 

Community Edition for Bias Mitigation - How It Works
Free sign up: Your financial organisation can sign up here.
Easy to get started: Upload a structured data file, like an Excel spreadsheet, to kick off the analysis process. Users can also connect to relational database services including AWS, Azure, Google Cloud, Oracle, and others, to build custom datasets for analysis. The platform learns the structure of the data in real-time, and the analysis process can crunch over four million rows of data in roughly ten minutes. 
Bias summary and score: Once the analysis is complete, users are provided with a Synthesized Total Fairness Score that shows what percentage of the dataset contained biased data. The platform also highlights areas of the data in which bias was detected.

Rebalancing: As mentioned, the final feature available in this process is the ability to automatically rebalance biased data.

The Community Edition is one part of Synthesized’s data platform. The complete platform uses AI to automate all stages of data provisioning; the process of making data available in an orderly and secure way. This level of automation enables organisations to generate synthesized datasets, allowing them to better test data for new products and tools, validate mathematical models, or train machine learning models.  

Synthesized was founded in 2017. Just two weeks before the UK went into lockdown in March 2020, the company closed a seed funding round of GBP 2.2 million. More recently it collaborated with the Financial Conduct Authority (FCA) to launch a collection of synthetic fraud datasets for secure third-party collaboration in the Digital Sandbox Pilot, jointly launched by the FCA and City of London Corporation.


Free Headlines in your E-mail

Every day we send out a free e-mail with the most important headlines of the last 24 hours.

Subscribe now

Keywords: Synthesized, AI, biases, data, credit rating, bias identification, data ethics, rebalancing
Categories: Banking & Fintech | Online & Mobile Banking
Countries: United Kingdom
This article is part of category

Banking & Fintech