New MINT Demo 2 tool tests AI models for training data inclusion

Researchers have developed a Membership Inference Test (MINT) Demo 2, a framework to enhance transparency in machine learning training. This tool can experimentally determine if specific data was used in a model's training process. Demonstrations on a face recognition model and four large language models (LLMs) showed up to 90% accuracy in detecting training data. The framework, which includes variations like aMINT and gMINT, is now available on a web platform to audit image and text models, aiming to support AI transparency and regulatory compliance. AI

IMPACT This tool could help ensure AI models are trained ethically and in compliance with data privacy regulations.

RANK_REASON The cluster describes a research paper detailing a new framework for membership inference testing on AI models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• DagsHub
• Hugging Face
• LLMs
• MINT Demo 2
• Mint Minimal Information Neuro Symbolic Tree

• paper
• safety
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →