Routing Enforced Generative Model for Recipe Generation

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 3797–3806, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 3797 Routing Enforced Generative Model for Recipe Generation Zhiwei Yu∗,1,2, Hongyu Zang∗,1 and Xiaojun Wan1 1Wangxuan Institute of Computer Technology, Peking University 1The MOE Key Laboratory of Computational Linguistics, Peking University 2Microsoft Research Asia yuzw, zanghy, wanxiaojun@pku.edu.cn Abstract One of the most challenging part of recipe generation is to deal with the complex restric- tions among the input ingredients. Previous re- searches simplify the problem by treating the inputs independently and generating recipes containing as much information as possible. In this work, we propose a routing method to dive into the content selection under the inter- nal restrictions. The routing enforced gener- ative model (RGM) can generate appropriate recipes according to the given ingredients and user preferences. Our model yields new state- of-the-art results on the recipe generation task with significant improvements on BLEU, F1 and human evaluation. 1 Introduction Food is a critical contributor to physical well being, a major source of pleasure, worry and stress, a major occupant of waking time, and across the world, the single greatest category of expenditures for human beings (Rozin et al., 1999). Recipes are a specific genre of instructional language to teach people how to prepare delicious food. They have been gaining interests in recent researches as recipes contain immensely rich information about the real world (Yagcioglu et al., 2018). Among previous efforts towards computational recipe studies, there are two lines in obtaining cook- ing recipes for users: recipe retrieval and recipe generation. Recipe retrieval (Chen and Ngo, 2016; Min et al., 2017) matches the entities from the given dish pictures or text inputs to find the corre- sponding recipes. Provided with ingredients (Yang et al., 2017), recipe titles (Kiddon et al., 2016), or dish photos (Salvador et al., 2019), recipe gener- ation models introduce additional mechanisms to assure the generated recipes containing as much ∗ The two authors contributed equally to this paper. Con- tribution was done at Peking University. Figure 1: Recipe Examples given ingredients as possible. In previous studies, the target recipes are exactly composed of the given ingredients. However, in practice, people usually have a number of ingredients at hand and do not know what to cook. They have difficulty in choos- ing appropriate set of ingredients. And it can be hard for them to input an accurate recipe title for the models (Kiddon et al., 2016; Majumder et al., 2019). What’s more, users may have preferences on some ingredients (e.g. “olive oil”) or categories (e.g. “Low Sugar”). As Figure 1 shows, given the same ingredient list, there can be different sets of in- gredients contributing to recipes with different user demands. Previous researches have not discussed this common scene of life. There is a clear need in finding suitable cooking recipes that match user demands. As increasing the variety and creativity of daily dishes can promote our happiness, in this work, we manage to obtain …

Content truncated to 500 words