HOME > 상세정보

상세정보

Deep learning approaches for representing food entities and recommending food pairings

Deep learning approaches for representing food entities and recommending food pairings

자료유형
학위논문
개인저자
박동현, 朴東炫
서명 / 저자사항
Deep learning approaches for representing food entities and recommending food pairings / Donghyeon Park
발행사항
Seoul :   Graduate School, Korea University,   2020  
형태사항
vii, 62장 : 천연색삽화, 도표 ; 26 cm
기타형태 저록
Deep Learning Approaches for Representing Food Entities and Recommending Food Pairings   (DCOLL211009)000000232138  
학위논문주기
학위논문(박사)-- 고려대학교 대학원: 컴퓨터·전파통신공학과, 2020. 8
학과코드
0510   6YD36   385  
일반주기
지도교수: 강재우  
서지주기
참고문헌: 장 58-62
이용가능한 다른형태자료
PDF 파일로도 이용가능;   Requires PDF file reader(application/pdf)  
비통제주제어
인공지능 , 음식 , 식재료 , 푸드페어링 , 추천시스템 , 딥러닝,,
000 00000nam c2200205 c 4500
001 000046048373
005 20200921092612
007 ta
008 200629s2020 ulkad bmAC 000c eng
040 ▼a 211009 ▼c 211009 ▼d 211009
085 0 ▼a 0510 ▼2 KDCP
090 ▼a 0510 ▼b 6YD36 ▼c 385
100 1 ▼a 박동현, ▼g 朴東炫
245 1 0 ▼a Deep learning approaches for representing food entities and recommending food pairings / ▼d Donghyeon Park
260 ▼a Seoul : ▼b Graduate School, Korea University, ▼c 2020
300 ▼a vii, 62장 : ▼b 천연색삽화, 도표 ; ▼c 26 cm
500 ▼a 지도교수: 강재우
502 1 ▼a 학위논문(박사)-- ▼b 고려대학교 대학원: ▼c 컴퓨터·전파통신공학과, ▼d 2020. 8
504 ▼a 참고문헌: 장 58-62
530 ▼a PDF 파일로도 이용가능; ▼c Requires PDF file reader(application/pdf)
653 ▼a 인공지능 ▼a 음식 ▼a 식재료 ▼a 푸드페어링 ▼a 추천시스템 ▼a 딥러닝
776 0 ▼t Deep Learning Approaches for Representing Food Entities and Recommending Food Pairings ▼w (DCOLL211009)000000232138
900 1 0 ▼a Park, Dong-hyeon, ▼e
900 1 0 ▼a 강재우, ▼g 姜在雨, ▼e 지도교수
945 ▼a KLPA

전자정보

No. 원문명 서비스
1
Deep learning approaches for representing food entities and recommending food pairings (43회 열람)
PDF 초록 목차

소장정보

No. 소장처 청구기호 등록번호 도서상태 반납예정일 예약 서비스
No. 1 소장처 과학도서관/학위논문서고/ 청구기호 0510 6YD36 385 등록번호 123064856 도서상태 대출가능 반납예정일 예약 서비스 B M
No. 2 소장처 과학도서관/학위논문서고/ 청구기호 0510 6YD36 385 등록번호 123064857 도서상태 대출가능 반납예정일 예약 서비스 B M

컨텐츠정보

초록

Over the last decade, as the era of deep learning has arrived, we are witnessing the explosion of various types of data. Among various fields of data, interest in food-related data is also increasing day by day. Food is what we face and eat every day, but the research interest in data on food has not been so great. Collecting data on food is more difficult than expected, and even if there is data, food has been a subject that has been difficult to computationally represent and analyze. Nevertheless, the field of food is still an area that needs to be explored and is important. To support this, we believe that the recent increase in quality food data and deep learning technology can solve this.

This dissertation introduces deep learning approaches for representing food entities and recommending food pairing. First, we define the task and importance of computationally representing food entities. Then, we introduce our previous studies on how we utilized food databases and food recipes to computationally represent food entities. Inspired by NLP tasks, various types of embedding techniques were applied. We demonstrated that our representation of food entities can be applied to many food-related tasks, such as food pairing, food-compound prediction, and so on. We believe our contributions of this dissertation can further improve the food-related researches in near future.

목차

Contents
Abstract
Contents i
List of Figures iv
List of Tables vi
1 Introduction 1
1.1 Background and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Task Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Our Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Embedding Food Ingredients Based on Chemical Combination in Dense
Vector Space 8
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Dataset & Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.2 Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.3 Food ingredient embedding using Doc2vec . . . . . . . . . . . . . . 12
2.4 Results and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
i
2.4.1 Visualization of food ingredient embedding based on chemical combination
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.2 Evaluation on the similarity of food ingredients embedding . . . . 14
2.4.3 Experiment results and analysis . . . . . . . . . . . . . . . . . . . . 14
2.5 Conclusion and Future Research . . . . . . . . . . . . . . . . . . . . . . . 16
3 KitcheNette: Predicting and Ranking Food Ingredient Pairings using
Siamese Neural Networks 17
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.1 Food Related Research . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.2 Siamese Neural Networks . . . . . . . . . . . . . . . . . . . . . . . 21
3.3 Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.3.1 Dataset Description and Preprocessing . . . . . . . . . . . . . . . . 21
3.3.2 Food Ingredient Pairing Score Generation . . . . . . . . . . . . . . 22
3.4 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.4.1 Learning Ingredient Representations . . . . . . . . . . . . . . . . . 24
3.4.2 Predicting Food Ingredient Pairing Scores . . . . . . . . . . . . . . 25
3.4.3 Model Training Details . . . . . . . . . . . . . . . . . . . . . . . . 26
3.5 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5.1 Baseline Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5.2 Main Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.5.3 Ablation Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.6 Qualitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.6.1 Finding Unknown Pairings . . . . . . . . . . . . . . . . . . . . . . 30
3.6.2 Comparison of Food Pairing Ranking Results . . . . . . . . . . . . 32
3.6.3 Discovering New Drink-Food Pairings . . . . . . . . . . . . . . . . 33
3.7 Conclusion & Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . 33
ii
4 FlavorGraph: A large-scale food-chemical graph for generating food
representations and recommending food pairings 35
4.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.3.1 Representing food ingredients and chemical compounds in a vector
space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.3.2 Node clustering by food category . . . . . . . . . . . . . . . . . . . 41
4.3.3 Case Study 1: Using the flavor representations of FlavorGraph for
food pairing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.4 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.4.1 Building FlavorGraph . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.4.2 Graph Node Embedding in FlavorGraph . . . . . . . . . . . . . . . 51
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5 Conclusion 56
Bibliography 58