Department of Food Science (KU FOOD)
University of Copenhagen, Denmark
Hello! My name is Mansi Goel. I am a Postdoctoral Researcher in the Section for Chemometrics and Analytical Technology at the Department of Food Science, University of Copenhagen, Denmark, where I work with Prof. Rasmus Bro and Prof. Morten Arendt Rasmussen. My research brings together Machine Learning, Natural Language Processing (NLP), Large Language Models (LLMs) and chemometrics to make food, flavour and nutrition computable—an emerging domain known as Computational Gastronomy.
My current work spans multimodal models for dietary assessment from food images, knowledge graphs and ontologies for semantic interoperability in food data, NLP-driven sensory analysis, and Bayesian and multi-block predictive modelling of analytical data.
Before joining the University of Copenhagen, I completed my PhD at IIIT-Delhi, India. My doctoral research, supervised by Prof. Ganesh Bagler, explored how AI can make food computable, leading to tools for novel recipe generation, predictive models for molecular toxicity, sweetness and umami taste, and comprehensive databases like FlavorDB2. I am passionate about developing AI-driven methods that bridge the gap between traditional food science and modern computational techniques.
During my PhD, I was a Visiting Researcher at the National Institute of Informatics, Japan. My earlier academic journey includes a Dual Degree (B.Tech & M.Tech) in Computer Science and Engineering from JIIT Noida, where I laid my foundation in algorithms and computational modeling.
Thank you for visiting my website! I am always interested in meaningful collaborations in computational gastronomy, predictive modelling, and applied AI in life sciences. Feel free to reach out if you would like to discuss research ideas or potential collaborations.
Research Metrics
CV & Skillset
- Machine Learning
- Deep Learning
- NLP
- Data Science
- Web Development
- Generative Algorithms
Education
-
Indraprastha Institute of Information Technology Delhi (IIIT-Delhi)PhD, Computational Gastronomy (advisor: Prof. Ganesh Bagler)2020 - 2025
-
Jaypee Institute of Information Technology (JIIT), NoidaDual Degree B.Tech-M.Tech in Computer Science & Engineering2015 - 2020
Experience
-
University of Copenhagen, DenmarkPostdoctoral Researcher, Chemometrics and Analytical Technology (with Prof. Rasmus Bro & Prof. Morten Arendt Rasmussen)2025 - present
-
National Institute of Informatics, JapanVisiting ResearcherOct 2024 - March 2025
-
IIIT-DelhiTeaching Assistant2020 - 2024
-
NTPC NoidaWeb DesignerMay 2018 - July 2018
Selected Publications (view all )
Food Image Nutrition Estimation with Open-Weight Large Multimodal Models: Fine-Tuning, Multi-View Ensembles, and Retrieval Augmentation
Zhi Ye, Mansi Goel, Morten Arendt Rasmussen
British Machine Vision Conference (BMVC) 2026 Accepted
Zero-shot benchmarking, LoRA fine-tuning, multi-view ensembling, and FNDDS retrieval-augmented generation with open-weight multimodal models (Gemma 3, Qwen3-VL) for estimating nutrient content from food images on Nutrition5k and ACETADA.
Food Image Nutrition Estimation with Open-Weight Large Multimodal Models: Fine-Tuning, Multi-View Ensembles, and Retrieval Augmentation
Zhi Ye, Mansi Goel, Morten Arendt Rasmussen
British Machine Vision Conference (BMVC) 2026 Accepted
Zero-shot benchmarking, LoRA fine-tuning, multi-view ensembling, and FNDDS retrieval-augmented generation with open-weight multimodal models (Gemma 3, Qwen3-VL) for estimating nutrient content from food images on Nutrition5k and ACETADA.
Universal statistical laws governing culinary design
Ganesh Bagler, Gopal Krishna Tewari, Aditya Raj Yadav, Akshat Singh, Pranay Bansal, Ujjval Dargar, Mansi Goel, Madhvi Kumari Sinha
arXiv 2026
Statistical regularities that recur across world cuisines, pointing to universal constraints on how recipes are composed.
Universal statistical laws governing culinary design
Ganesh Bagler, Gopal Krishna Tewari, Aditya Raj Yadav, Akshat Singh, Pranay Bansal, Ujjval Dargar, Mansi Goel, Madhvi Kumari Sinha
arXiv 2026
Statistical regularities that recur across world cuisines, pointing to universal constraints on how recipes are composed.
AllerStack: Predicting Allergenic Proteins with a Stacked Ensemble Approach
Harshika Sharma, Mansi Goel, Devkul Sahu, Mazhar Ayaz Sayed, Pratham Mangla, Pratham Shekhawat, Shubham Yadav, Ganesh Bagler
bioRxiv 2025
AllerStack is a stacked ensemble machine learning framework that integrates sequence-based descriptors with antigenicity predictions to identify allergenic proteins
AllerStack: Predicting Allergenic Proteins with a Stacked Ensemble Approach
Harshika Sharma, Mansi Goel, Devkul Sahu, Mazhar Ayaz Sayed, Pratham Mangla, Pratham Shekhawat, Shubham Yadav, Ganesh Bagler
bioRxiv 2025
AllerStack is a stacked ensemble machine learning framework that integrates sequence-based descriptors with antigenicity predictions to identify allergenic proteins
IEML-A New Frontier for Semantic Computing and Collective Intelligence in the Cooking Domain
Mansi Goel, Frederic Andres, Pierre Levy
IEEE 41st International Conference on Data Engineering Workshops (ICDEW) 2025
Explores the transformative potential of the Information Economy Meta Language (IEML) in revolutionizing cooking recipes' management, personalization, and innovation
IEML-A New Frontier for Semantic Computing and Collective Intelligence in the Cooking Domain
Mansi Goel, Frederic Andres, Pierre Levy
IEEE 41st International Conference on Data Engineering Workshops (ICDEW) 2025
Explores the transformative potential of the Information Economy Meta Language (IEML) in revolutionizing cooking recipes' management, personalization, and innovation
A Multilingual Ontology and Knowledge Graph for Recipes
Mansi Goel, Frederic Andres
IEEE 41st International Conference on Data Engineering Workshops (ICDEW) 2025
Multilingual recipe ontology capturing critical relationships between ingredients, nutrition, cooking actions, and recipe planning
A Multilingual Ontology and Knowledge Graph for Recipes
Mansi Goel, Frederic Andres
IEEE 41st International Conference on Data Engineering Workshops (ICDEW) 2025
Multilingual recipe ontology capturing critical relationships between ingredients, nutrition, cooking actions, and recipe planning
All publications
Projects
Turing Test for Chefs
Evaluated recipe generation models with human experts.
Flavor Profile Prediction
Web server for predicting the flavor profile of molecules and ingredient categories.
Foodle
Food-related game to learn about food.
Recipe Generation
Generative models for cultural recipes.
SustainableFoodDB
Sustainable food database and carbon footprint calculator for recipes
Computational Gastronomy APIs
APIs for recipes and flavor molecules for computational gastronomy research.
Highlights
