Latest News & Research
Stay updated with Hexagone AI's latest achievements, research publications, and company news.
Hexagone AI presentation at HR Congress 'The Future of HR Data: Ethical AI & Anonymization'
We explore ethical risks and opportunities of GenAI in HR, followed by a discussion on how Hexagone AI’s anonymization technology can help organizations securely capture the value of sensitive employee data.
Mariia Zameshina Selected as One of 16 Female Entrepreneurs to Watch in 2025
Hexagone AI's CEO Mariia Zameshina has been recognized by STATION F as one of the 16 high-potential female entrepreneurs to watch in 2025 as part of their Female Founders Fellowship. The program celebrates women leaders shaping the future of tech, highlighting Mariia's contribution to AI innovation and ethical data practices.
Hexagone AI presentation at PPAI workshop @ AAAI conference in Philadelphia
'PrivacyGAN: robust generative image privacy' paper by Hexagone AI researchers is presented at PPAI workshop @ AAAI conference in Philadelphia
Hexagone AI is Selected for Microsoft GenAI Studio Program
Hexagone AI has been selected as one of 15 startups for Microsoft's first GenAI Studio program at Station F. The program includes up to $150,000 in Azure credits and support from industry leaders like MistralAI, Nvidia, and GitHub. Hexagone AI was recognized for its innovative multimodal anonymization solution.
Evolutionary Retrofitting
New research exploring evolutionary algorithms for model adaptation, introducing innovative techniques for efficient model updating and enhancement.
M Videau, M Zameshina, A Leite, L Najman, M Schoenauer, O Teytaud · arXiv preprint arXiv:2410.11330
Log-normal Mutations and their Use in Detecting Surreptitious Fake Images
Novel approach to detecting AI-generated images using log-normal mutations, providing robust detection methods for synthetic content.
I Labiad, T Bäck, P Fernandez, L Najman, T Sander, F Ye, M Zameshina · arXiv preprint arXiv:2409.15119
Agnostic latent diversity enhancement in generative modeling
Innovative research on improving diversity in generative models through agnostic latent space optimization techniques.
M Zameshina, M Videau, A Leite, M Schoenauer, L Najman, O Teytaud · HAL Science
PrivacyGAN: robust generative image privacy
Groundbreaking work on protecting visual privacy through generative AI, introducing robust methods for image anonymization.
M Zameshina, M Careil, O Teytaud, L Najman · arXiv preprint arXiv:2310.12590