Tag
Introduces OmniFood-Bench, a benchmark for evaluating vision-language models on nutrient reasoning and personalized health advice. Experiments show VLMs struggle with mass estimation and safety-critical recommendations.
Nutrepedia is a nutrition information tool available in 29 locales, built using Clojure and Htmx.
Researchers at UC Davis developed an AI model that suggests small ingredient swaps to improve nutrition and reduce meal costs by up to 34%, as reported in PLOS Digital Health.
This paper proposes Mixed Integer Goal Programming (MIGP) for personalized meal optimization, addressing the limitations of fractional servings and infeasibility from hard constraints. It introduces integer variables for practical serving counts and goal programming deviations with inverse-target normalization, achieving 100% feasibility and better solutions than rounding-based approaches, with an open-source Python implementation.
The article lists nine foods claimed to boost testosterone levels in men, starting with eggs.
A social post shares 12 prompts that turn Claude into a free registered dietitian for weekly meal-prep, macro tracking and grocery savings.
Healthify, a health and fitness AI platform, partnered with OpenAI to enhance its AI-powered nutritionist Ria and food recognition feature Snap, overcoming limitations in accuracy, scalability, and multilingual support. The collaboration represents a significant upgrade to Healthify's decade-long AI-driven health coaching platform.