You may have come across from earlier posts that we, Clearbox AI, were the proud recipients of the first edition of the WomenTechEU awards of the European Commission. We have now come to the end of the 6-month programme and would like to reflect on our experience to be helpful to others who may be interested.
The award's goal is to select and support the 50 most innovative deep-tech startups in Europe with a woman leader in the startup team. Deep-tech startups, like Clearbox AI, are founded on engineering innovation and science advances and have longer R&D cycles. The WomenTechEU programme was an excellent and timely support opportunity in our efforts to transfer our R&D efforts into practice.
The main objectives of our successful project application to WomenTechEU are:
- Accelerate communication and dissemination of Clearbox AI’s value-based propositions of trust, bias mitigation, and inclusion
- Implement a bias and fairness assessment and mitigation module within our product offering including the release of open-source modules
We exceeded expectations in our communication and dissemination plan for the past 6 months. During the span of the WomenTechEU programme we created rich content through our blog posts, social media channels as well as our thought leadership activities in various events, and conferences. Here’s a selection of these activities:
- Speaker at the IEEE World Congress on Computational Intelligence 2022 (July 18-23, 2022)
- Panel speaker at the Italian association of banks on the use of synthetic data for trustworthy AI (July 14, 2022)
- Italian national television interview (RAI) on data ethics, bias and fairness (July 8, 2022)
- Interview on synthetic data for trustworthy AI and responsible innovation on Cybernews (June 14, 2022)
- Webinar on the use of synthetic data for bias mitigation and fairness, Data Phoenix (October 5, 2022)
- Speech on the role of synthetic data in the new-paradigm of data-centric AI (for privacy and fairness) with Italian data experts in banking and innovation, Annual Future Driven Banking (October 27-28, 2022)
Most recently, we also formalised a crucial partnership with BearingPoint, a prominent IT and management consultant company in Europe, to enhance their Machine Learning models with our synthetic data offering. The partnership announcement was published on Econopoly, the blog of Sole24ore, the leading Italian newspaper on November 4, 2022.
Our CEO Shalini Kurapati also participated in the various training and networking events organised by the European Innovation Council. It was an interesting experience to meet fellow entrepreneurs and researchers across Europe working on cutting-edge innovation. We are happy with our communication and networking activities. Still, we are even more excited about achieving our second and most impactful objective- the release of the open-source module on bias detection and mitigation.
We published our software on detecting bias, improving data quality, and implementing the fairness metrics with the following software modules:
- Data quality assessment and profiling module
- Bias detection, mitigation, and fairness metrics module
You can find the detailed documentation in the README files of each of the two software modules.
We have also documented the practical use of the two modules with a step-by-step guide and report that is publicly available on the zenodo public repository citing the EU funding and Grant number: https://zenodo.org/record/7266913
The work performed during the past six months was the first step towards data quality augmentation for trustworthy AI. Augmenting datasets is not trivial and usually corresponds to finding the ideal trade-off between local and global performance. However, synthetic data can be an excellent instrument for data scientists striving for better models, and we will continue to explore its potential even beyond the scope of the WomenTechEU project.