SWIPE N’ BITE: Food Discovery and Social Matching Platform and AI-Driven Recommendations

Abstract

Food discovery platforms today often overwhelm users with extensive listings, reviews, and manual filtering mechanisms, leading to decision fatigue and reduced engagement. Also, most food recommendation systems don\'t use social context or real-time user intent. This study introduces Swipe N\' Bite, a smart food discovery and recommendation platform that features Food Cupid, a new food-interest-based social matching mechanism, along with swipe-based preference learning and machine learning-driven personalization. The system learns about users\' unspoken preferences through swipe interactions, makes individualized food suggestions using a content-based similarity model, and allows for real-time social matching when several users show interest in the same food item within a short time frame. To make user interaction even better, an AI food assistant is added to help people find and ask about food in a conversational way. The platform is a web app built with modern frontend technologies and an architecture that can grow with it. Compared to traditional list-based food platforms, experimental testing and qualitative analysis show that more users are interested, it takes less time to find things, and the recommendations are more relevant.

Introduction

The text discusses how traditional food discovery apps—using ratings, reviews, and list-based searches—often overwhelm users with too much information, leading to decision fatigue and disengagement. Modern human-centered design research highlights the effectiveness of swipe-based, implicit feedback mechanisms, which reduce cognitive load while capturing strong signals about user preferences. However, most current food platforms do not fully utilize these interactions, nor do they integrate real-time social features that connect users based on shared food interests.

To address these gaps, the research proposes Swipe N’ Bite, a food discovery platform combining:

1. Swipe-based preference learning: Users swipe left or right on food items to indicate disinterest or interest, respectively. This implicit feedback is logged and used to learn preferences without requiring manual ratings.

2. Food Cupid social matching: A novel, temporal social matching system connects users who express interest in the same food item within a short time window. Matches are interest-first, not profile-based, and enable opt-in chat interactions.

3. Machine learning-based recommendations: A content-based engine suggests similar dishes based on users’ saved preferences, continuously updating as new interactions occur.

4. Conversational AI (Gemini API): A chatbot allows users to explore food preferences and discover new options via natural language queries, complementing the swipe-based implicit learning.

The platform integrates these components into a unified workflow: swipe interactions capture preferences, which feed personalized recommendations, social matching, and AI-assisted exploration. This hybrid approach minimizes cognitive load, enhances personalization, encourages social engagement, and creates a more intuitive, user-centered food discovery experience.

Conclusion

This research introduced Swipe N’ Bite, an intelligent food discovery and social matching platform that incorporates swipe-based preference learning, machine learning-driven recommendations, and an innovative interest-based social matching system known as Food Cupid. The system was made to fix some of the main problems with traditional food discovery platforms, such as decision fatigue, the need for clear feedback, and the lack of socially contextual discovery mechanisms. The platform effectively captures changing food preferences with little effort from users by using implicit user interactions like swipe gestures. The content-based recommendation engine lets users get personalized food suggestions based on their saved preferences. The Food Cupid mechanism adds a social aspect by connecting users who are both interested in the same food items at the same time. The addition of a conversational AI assistant also makes exploratory discovery better by allowing natural language interaction and modeling behavioral preferences. The current prototype makes some real-world factors easier to understand, like a wide range of users and restaurants that are open at the time, but it does prove the main idea: using implicit interaction modeling, machine learning-based personalization, and interest-driven social matching can greatly enhance food discovery experiences. The long-term goal of this project is to turn Swipe N\' Bite into a smart, scalable platform that helps people find food in a way that is socially enriching and focused on people. Swipe N\' Bite\'s long-term goal is to become a smart, scalable platform that helps people find food in a way that is good for society and the environment. The platform is in line with the broader Sustainable Development Goals (SDGs) like SDG 11 (Sustainable Cities and Communities) by helping people make smarter food choices in cities and communities, and SDG 12 (Responsible Consumption and Production) by letting people make food choices based on their preferences and needs within their own ecosystems.

References

[1] S. Jannach and M. Jugovac, “Recency-based Evaluation of Recommender Systems,” ACM Trans. Inf. Syst., vol. 38, no. 2, pp. 1–29, Apr. 2020. [2] A. Deldjoo, M. T. Ribeiro, F. Ricci, and V. Ntsifmon, “Towards Explainable and Interactive Content-Based Recommendation,” in Proc. ACM RecSys, 2021, pp. 314–319. [3] Q. Bai, Y. Zhang, and Y. Gao, “Personalized food recommendation by multi-modality deep learning,” IEEE Trans. Multimedia, vol. 22, no. 12, pp. 3181–3196, Dec. 2020. [4] J. L. Herlocker, J. A. Konstan, L. G. Terveen, and J. T. Riedl, “Evaluating Recommendation Systems,” ACM Trans. Inf. Syst., vol. 22, no. 1, pp. 5–53, 2004. [5] F. Ricci, L. Rokach, and B. Shapira, “Introduction to Recommender Systems Handbook (2nd ed.),” Springer, 2022. [6] Y. Zhang, X. Chen, and Z. Huang, “Deep Learning Based Recommender System: A Survey and New Perspectives,” ACM Comput. Surv., vol. 52, no. 1, pp. 1–38, Feb. 2020. [7] P. Hu, J. Liu, and Q. Yang, “A Survey on Recommender Systems Integrating Social Network Analysis,” IEEE Trans. Knowl. Data Eng., vol. 33, no. 10, pp. 3997–4018, Oct. 2021. [8] J. Tang, H. Wang, and H. Liu, “Recommendation with Social Context,” in Proc. ACM SIGKDD, London, UK, Aug. 2019, pp. 3096–3099. [9] S. Rendle, “Factorization Machines with libFM,” ACM Trans. Intell. Syst. Technol., vol. 11, no. 3, pp. 1–23, May 2020. [10] A. Kumar and R. Sharma, “Hybrid Deep Learning Recommender Systems: A Review,” IEEE Access, vol. 9, pp. 132705–132728, 2021. [11] C. Zhao, D. Wang, F. Zhu, and M. Zhang, “Neural Social Recommendation with User Interest Alignment,” IEEE Trans. Knowl. Data Eng., vol. 33, no. 7, pp. 3096–3108, Jul. 2021. [12] R. Palumbo and F. Ricci, “Conversational Recommender Systems: Framework and Research Agenda,” ACM Trans. Interact. Intell. Syst., vol. 12, no. 3, pp. 1–36, Sep. 2022 [13] D. Zhang, X. Zhu, and S. Wang, “Chatbot-Based Interactive Recommendations: A Survey,” IEEE Access, vol. 9, pp. 30581–30602, 2021. [14] I. Guy, Y. Zilberstein, and Y. Maarek, “Social Matching in Recommender Systems: Overview,” ACM Comput. Surv., vol. 51, no. 2, pp. 1–39, Jun. 2019. [15] E. Pizzutilo, A. Seno, and F. Paternò, “Context-Aware Food Recommender Systems: Trends and Challenges,” IEEE Access, vol. 7, pp. 164768–164792, 2019. [16] H. Wang, F. Zhang, M. S. Yang, and A. Sun, “Neural Graph Collaborative Filtering,” in Proc. WWW, San Francisco, CA, USA, 2019, pp. 165–175. [17] J. Wang, Q. Liu, and Y. Ren, “Deep Learning for Recommendation: A Survey,” IEEE Trans. Big Data, vol. 8, no. 6, pp. 1222–1247, Dec. 2022. [18] B. Rashid, R. Qu, and C. Salter, “User Satisfaction in Recommender Systems: Challenges and Perspectives,” IEEE Trans. Syst., Man, Cybern. Syst., vol. 51, no. 6, pp. 3599–3612, Jun. 2021. [19] T. Luo, H. Liu, and Z. Zhou, “Swipe-Based User Preference Modelling for Real-Time Interaction,” in *Proc. ACM UIST*, 2022, pp. 451–463. [20] C. Lin, Y. Chen, and K. Yang, “Interest-Based Social Matching Algorithms for Dynamic Recommendations,” IEEE Trans. Knowl. Data Eng., vol. 35, no. 4, pp. 3455–3470, Apr. 2023.

Copyright

Copyright © 2025 Saurabh Kumar Singh, Irena Nidhi, G. Suseela. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.