Prioritizing neglected food species in nutritional studies using expert-knowledge and explainable AI

2026-03-02 · Source: Machine learning : nature.com subject feeds · Field: Agriculture & Food Systems — Artificial Intelligence & Machine Learning, Agricultural Sustainability & Climate, Nutrition, Fitness & Lifestyle Medicine · Depth: Intermediate, long

Summary

A study conducted in Brazil inventoried 369 neglected food species, including algae, aquatic fauna, wild terrestrial vertebrates, insects, mushrooms, and plants, to prioritize them for nutritional research. Utilizing a mixed-methods approach that combined expert knowledge with explainable AI (LightGBM and SHAP value analysis), researchers identified key factors influencing prioritization for food composition and consumption studies. The inventory revealed that plants (29.5%) and wild vertebrates (24.4%) dominate the species count, with significant data gaps in nutritional information, especially for algae, insects, and wild vertebrates. Over 36,000 recipes involving these neglected species were identified. The analysis showed that the number of recipes and species occurrence across different states were the most influential features, with R² values of 0.677 for food composition and 0.782 for consumption studies, highlighting the importance of cultural uses and local accessibility in shaping research priorities.

Key takeaway

For AI Scientists and nutrition researchers aiming to address food biodiversity gaps, this study demonstrates a robust methodology. You should consider integrating explainable AI, specifically LightGBM and SHAP analysis, with expert knowledge to prioritize neglected species. Focus on features like recipe prevalence and geographic distribution, as these strongly predict research prioritization and can guide resource allocation for data collection and intervention strategies in tropical regions.

Key insights

Explainable AI and expert knowledge can effectively prioritize neglected food species for nutritional research.

Principles

• Food biodiversity is crucial for human health.
• Cultural uses and local accessibility drive research priorities.

Method

A mixed-methods approach combining expert knowledge and explainable AI (LightGBM and SHAP value analysis) was used to identify key prioritization factors for neglected food species.

In practice

• Inventory local neglected food species.
• Analyze recipe data for cultural relevance.
• Map species occurrence for accessibility.

Topics

• Explainable AI
• Neglected Food Species
• Nutritional Research Prioritization
• LightGBM
• Food Biodiversity

Code references

• eliasjacob/paper_bionut
• microsoft/LightGBM
• shap/shap

Best for: AI Scientist, AI Researcher, Data Scientist, Research Scientist

Related on AIssential

• Data Solutions for Tailoring Agronomic Support to Meet Regional Needs - with Tami Craig Schilling of Bayer Crop Science — Generative AI provides localized, precise agronomic advice by integrating diverse environmental and genetic data.
• AI's Role in Saving the Endangered White-Headed Langur — AI-powered monitoring platforms can effectively drive conservation efforts for endangered species in challenging terrains.
• AI for inclusive and resilient agri-food systems: Potential ways forward — AI can significantly enhance food security, resilience, and sustainability in agri-food systems by addressing key challenges and opportunities.
• Batch-Invariant Spectral Intelligence for Robust and Explainable Insect Authentication — BISN uses learnable preprocessing and adversarial training to achieve robust, explainable insect authentication despite batch variations.
• Exploring NLP Applications in Food Research: my ATRIUM TNA visit to the GATE group — NLP tools significantly enhance social media data analysis for consumer behavior and food research.
• Zuckerberg's $500M AI biology swing — Large-scale data generation and advanced AI models are converging to revolutionize disease understanding and food science.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.