IRS-PM-2023-10-29-IS24FT-GRP24-Mango-AcademicKnowledgePlatform

---

SECTION 1 : PROJECT TITLE

Mango - An Academic Knowledge Platform

---

SECTION 2 : EXECUTIVE SUMMARY

Mango is an academic knowledge platform for college students and researchers.

It has two main components -

• A Search Engine for essays with their summaries.
• An Academic Papers Recommender for the latest and valuable academic papers.

The platform aims to assist young college students and researchers to quickly grab a general understanding of a particular field.

• It can also help them to keep up to date on the forefront development and progress of that field.

---

SECTION 3 : CREDITS / PROJECT CONTRIBUTION

Official Full Name	Student ID	Work Items	Email
Li Jiacheng	A0285823W	Data Preprocessing, Whole System designing and implementation Search and Recommendation system designing	[email protected]
Goh Min Hua	A0285810A	Data Preprocessing, Recommendation algorithm and system designing and development	[email protected]
Mao Zhihong	A0285799X	Data Preprocessing, Search algorithm and engine designing and development	[email protected]

---

SECTION 4 : VIDEO OF SYSTEM MODELLING & USE CASE DEMO

• Promotion: https://youtu.be/VvDoVLDkDYM
• System: https://youtu.be/ODpb9tNP02g

---

SECTION 5 : USER GUIDE

Please refer to the document "Mango - User Guide" within the Project Report folder.

---

SECTION 6 : PROJECT REPORT

Refer to Github Folder: ProjectReport

1. EXECUTIVE SUMMARY 3
2. INTRODUCTION 4 2.1. Business Problem Background 4 2.2. Current Solutions 4 2.3. Market Research 5 2.4. Project Objectives 8 2.5. Success Measurements 8
3. PROJECT SOLUTION (SYSTEM DESIGN AND MODEL) 9 3.1 Architecture Overview 9 Detailed Workflow 9 Technical Stack and Architecture 9 3.2 Data Management 10 Data Sources 10 Data Preparation 10 Pre-processed Training Dataset 11 Vector Database for ANN indexes 11
4. PROJECT IMPLEMENTATION 12 4.1 Information Retrieval System – Dense Passage Retrieval 12 4.2. Recommendation System (Trending Page) 13 Dual Tower Model 13 Training Process 14 4.3. Recommendation System (Relevant Papers) 15
5. PROJECT PERFORMANCE AND VALIDATION 15
6. CONCLUSION 19 6.1 Challenges and Solutions 19 6.2 Future Considerations 19 APPENDIX A – PROJECT PROPOSAL 20 APPENDIX B – MAPPED SYSTEMS FUNCTIONALITIES AGAINST KNOWLEDGE, TECHNIQUES AND SKILLS OF MODULE COURSES 30 APPENDIX C – INSTALLATION AND USER GUIDE 31 APPENDIX D – INDIVIDUAL PROJECT REPORT 35 Li Jiacheng (A0285823W) 35 Mao Zhihong (A0285799X) 36 Goh Min Hua (A0285810A) 37 REFERENCES 38

---

SECTION 7 : MISCELLANEOUS

Refer to Github Folder: Miscellaneous

Mango - Survey

The survey for the mango project.