Food Info Network System

Name: Food Info Network System
Author: University of Notre Dame

University of Notre Dame

DOCUMENTED

Published in academic literature

For:Researchers & AcademicsGeneral Public & Enthusiasts

Free

Download on the App Store Get it on Google Play

App Summary

The Food Info Network System (FINS) provides users with an optimized list of grocery items and recipes tailored to their personal dietary needs, budget, and preferences. The app's recommendation engine is grounded in a novel hierarchical graph attention network that models the complex relationships between user history, recipes, and ingredients. The associated research concludes that this network-based approach improves the quality and relevance of food recommendations compared to standard methods.

App Screenshots

Food Info Network System screenshot 1 of 8

Food Info Network System screenshot 2 of 8

Food Info Network System screenshot 3 of 8

Food Info Network System screenshot 4 of 8

Food Info Network System screenshot 5 of 8

Food Info Network System screenshot 6 of 8

Food Info Network System screenshot 7 of 8

Food Info Network System screenshot 8 of 8

Detailed Description

Functionality & Mechanism Developed at the University of Notre Dame, the Food Info Network System optimizes grocery item selection based on user-defined constraints. The interface captures dietary requirements, budget limitations, and a general shopping list. The system then leverages a graph-based recommendation engine, grounded in network science, to analyze complex relationships between food items and user profiles. This process, which integrates textual and structural data, generates a refined and specific purchasing list to guide consumer decisions.

Evidence & Research Context

The system's recommendation engine is based on published research detailing novel hierarchical graph attention networks (HGAT) and heterogeneous network embedding models (rn2vec).
Associated research demonstrates that these graph-based models can effectively learn complex representations by integrating textual, structural, and nutritional data for food-related tasks.
In computational experiments, the underlying algorithms demonstrated superior performance compared to standard baseline methods on recipe recommendation and classification benchmarks.
The development is supported by RecipeNet, a large-scale, structured corpus of recipe data created to facilitate network-based food studies.

Intended Use & Scope This system is intended for consumers seeking to optimize food purchasing and for researchers in computational food science. Its primary utility is as a decision-support tool for generating grocery lists aligned with dietary and budgetary goals. The tool does not provide medical nutrition therapy and is not a substitute for professional clinical guidance from a registered dietitian for managing health conditions.

Studies & Publications

2 publications

Peer-reviewed research associated with this app.

Non-Evaluative Reference

Recipe recommendation with hierarchical graph attention network

Tian et al. (2021) · Frontiers in Big Data

Referenced in academic literature; no direct evaluation of the app

Recipe recommendation systems play an important role in helping people find recipes that are of their interest and fit their eating habits. Unlike what has been developed for recommending recipes using content-based or collaborative filtering approaches, the relational information among users, recipes, and food items is less explored. In this paper, we leverage the relational information into recipe recommendation and propose a graph learning approach to solve it. In particular, we proposeHGAT, a novel hierarchical graph attention network for recipe recommendation. The proposed model can capture user history behavior, recipe content, and relational information through several neural network modules, including type-specific transformation, node-level attention, and relation-level attention. We further introduce a ranking-based objective function to optimize the model. Thorough experiments demonstrate thatHGAToutperforms numerous baseline methods.

Non-Evaluative Reference

Recipe representation learning with networks

Tian et al. (2021) · Proceedings of the 30th ACM International Conference on Information & Knowledge Management

Referenced in academic literature; no direct evaluation of the app

Learning effective representations for recipes is essential in food studies for recommendation, classification, and other applications. Unlike what has been developed for learning textual or cross-modal embeddings for recipes, the structural relationship among recipes and food items are less explored. In this paper, we formalize the problem recipe representation learning with networks to involve both the textual feature and the structural relational feature into recipe representations. Specifically, we first present RecipeNet, a new and large-scale corpus of recipe data to facilitate network based food studies and recipe representation learning research. We then propose a novel heterogeneous recipe network embedding model, rn2vec, to learn recipe representations. The proposed model is able to capture textual, structural, and nutritional information through several neural network modules, including textual CNN, inner-ingredients transformer, and a graph neural network with hierarchical attention. We further design a combined objective function of node classification and link prediction to jointly optimize the model. The extensive experiments show that our model outperforms state-of-the-art baselines on two classic food study tasks. Dataset and codes are available at https://github.com/meettyj/rn2vec.

In the Media

Using data to feed the world

University of Notre Dame researchers developed the Food Info Network System to address global hunger using data science tools and techniques, supported by nearly $5 million in funding from the Bill & Melinda Gates Foundation. Associate Professor Jaron Porciello emphasized that "we already grow enough food to feed the world" but noted that "we need a broad set of interventions, and we need the data to make those decisions possible." The system supports the Juno Evidence Alliance, which uses Porciello's taxonomy for evidence-based agriculture to foster collaboration between nonprofits, data analysis firms, and governments.

NdRead article

Michelle Sawwan Presents FINs Research Poster

Michelle Sawwan from the Center for Civic Innovation presented research on the Food Info Network System at the National Science Foundation meeting, developing a mobile app to optimize healthy food procurement for lower-income households through NIFA funding. Early project data demonstrated that lower-income heads of households already use technology to inform food procurement decisions by finding sales, searching online for healthy recipe ingredients, or finding foods with short preparation times. The app aims to augment shoppers' existing technology strategies while accounting for their constraints to improve purchasing decisions and health outcomes.

NdRead article

Oasis in the Desert

Professors Ron Metoyer and Ann-Marie Conrado developed the Food Info Network System to address food access challenges in food deserts, using ethnographic research to understand residents' shopping and eating behaviors. "As a design professor with a background in ethnography and qualitative research, my job is to really go out and connect with people and understand not only their needs, their challenges, their frustrations, but also their values, their aspirations, their goals," Conrado said. The study focuses on two food deserts in Detroit and South Bend, where the average poverty rate reaches nearly 36 percent.

NdRead article