( AI Journey, Secured )
AI promises revolution in every industrial sector. Yet, foundational models bring compliance issues: they are often built on unauthorised personal data, copyrighted materials, and untruthful sources. Parameter Lab addresses these challenges, safeguarding individuals and businesses in their AI journey.
Problems We Solve
We tackle an array of challenges to elevate the use of AI. We address privacy and copyright issues associated with training data for Large Language Models (LLMs). Simultaneously, we are tackling faithfulness, hallucination, and calibration issues prevalent in LLMs. In a bid to foster trust in AI systems, we are also enhancing the explainability and interpretability of LLMs. Additionally, we dedicate ourselves to provide tools that facilitate research towards safer and more reliable AI models, thereby enabling responsible AI evolution.
We partner with
Naver is perhaps the most widely known search engine in South Korea. Naver’s ML research groups are one of the best in the world : Naver Clova and Naver Labs Europe. Naver has internalised 100B-scale language models.
We are a member of
CyberValley is Europe's largest AI research consortium. It bridges academic and private sector partners to turn basic research into real-world applications.