This research explores the justification and implications of incorporating consumption variety into mobile-based food recommendation systems. Our study makes use of data from a popular mobile fitness app, in which we can observe large volumes of daily food logs of thousands of users. We first confirm that consumption variety is associated with lower overall calories consumed, higher vegetable consumption, and lower snack consumption. Motivated by these empirical patterns about variety and eating, we then seek out to design a novel multicriteria food recommendation system (FOODVAR) that can accommodate for variety in recommended foods. We then assess the impact of including this additional variety criterion in recommendation system performance, where we show that the incorporation of variety improves the algorithm’s evaluation metrics.

Typical recommender systems try to mimic the past behaviors of users to make future recommendations. For example, in food recommendations, they tend to recommend the foods the user prefers. While the recommended foods may be easily accepted by the user, it cannot improve the user’s dietary habits for a specific goal such as weight control. In this paper, we build a food recommendation system that can be used on the web or in a mobile app to help users meet their goals on body weight, while also taking into account their health information (BMI) and the nutrition information of foods (calories). Instead of applying dietary guidelines as constraints, we build recommendation models from the successful behaviors of comparable users: the weight loss model is trained using the historical food consumption data of similar users who successfully lost weight. By combining such a goal-oriented recommendation model with a general model, the recommendations can be smoothly tuned toward the goal without disruptive food changes. We tested the approach on real data collected from a popular weight management app. It is shown that our recommendation approach can better predict the foods for test periods where the user truly meets the goal, than the typical existing approaches.

Sentence ordering aims to arrange the sentences of a given text in the correct order. Recent work frames it as a ranking problem and applies deep neural networks to it. In this work, we propose a new method, named BERT4SO, by fine-tuning BERT for sentence ordering. We concatenate all sentences and compute their representations by using multiple special tokens and carefully designed segment (interval) embeddings. The tokens across multiple sentences can attend to each other which greatly enhances their interactions. We also propose a margin-based listwise ranking loss based on ListMLE to facilitate the optimization process. Experimental results on five benchmark datasets demonstrate the effectiveness of our proposed method.