Hi, this is a simple academic summary of

HAORAN LUO

My life has fallen into a local minimum, and I'm trying to adjust the learning rate.

Currently pursuing a Ph.D. in the field of AI at Waseda University.

Serving as a part-time Special Researcher at the Japan Science and Technology Agency.

Thank you for visiting my personal homepage, I hope you can enjoy your time here.

About Me

I am a doctoral student at Waseda University’s Graduate School of Creative Science and Engineering. My current research focuses on the development and industrialization of business digital transformation technologies for Japan’s Society 5.0, funded by the Japan Science and Technology Agency (JST). During my master’s and doctoral studies, I have concentrated on integrating computer science-based research with industry to drive innovation and economic development. This is also the core topic of my doctoral research. Specifically, I am advancing the development of a smart retail system integrated with artificial intelligence technology, aimed at addressing labor shortages and the challenges posed by an aging population.

I have interned at three leading Chinese internet companies including ByteDance and Tencent. My roles have ranged from algorithm engineer to game operations and strategy analysis. My personal research philosophy and life philosophy is: Innovation+Motivation=Application.

Here are a few fields I've been closely following:
  • LLMs
  • EEG Analysis
  • CV (Style Transfer)
  • Object Detection
  • NLP (Sentiment Analysis)
  • Consumer Behavior Modeling

Experience

Multimedia algorithm engineer(intern) - Tencent
January 2022 - March 2022
Tencent Oteam-OCR project
  1. Tencent Oteam-OCR Open Source Collaborative Project: responsible for the collaborative development of some algorithms for image distortion correction, detection, and recognition , and at the same time for code debugging, integration, testing under the remote server linux system.
  2. Tencent NLG (table text generation ) start -up project : responsible for table analysis mode, text preprocessing, structure design and algorithmization of table preprocessing, for efficient processing and analysis of Tencent’s various game data.
Game data operation(intern) - Bytedance-MOONTON
April 2022 - October 2022
MLBB M5
  1. Organize the No. 1 MOBA game in South Asia “MLBB” local events in Pakistan, Nepal, Bangladesh and more.
  2. Responsible for localized community operations and management of creator projects in South Asia which contains over 100 people.
  3. Use data analytics to analyze and predict game user data and calculate return on investment.
Strategy Analyst(intern) - Lilith Games
July 2023 - October 2023
The developer of Rise of Kingdoms
  1. Participate in the project establishment of the company ’s internal AI native pet management game , including meeting minutes , brainstorming, project BP, etc.
  2. Industry financial reports and annual report analysis , game competitive product monitoring , AI-driven game market forecast , etc. Game types include but are not limited to: FPS, match-3 and SLG.
  3. Conduct project investment and cooperation talks with leading game companies and new power studios in Europe , the United States, South Korea and other regions.
  1. Aspect-Level Sentiment Analysis Framework.
  2. Consumer Intelligent Guidance System for Smart Retail.
  3. Face Shape Adaptive Makeup Transfer (FSAMT).
  4. Regulator-Feedback Based Attention (RFA) for EEG Sentiment Classification.
  5. …

Projects

FSAMT Face Shape Adaptive Makeup Transfer
Computer Vision Style Transfer Makeup Transfer Face Recognition
FSAMT Face Shape Adaptive Makeup Transfer
A high-level makeup artist at zero cost for everyone! Makeup transfer is the process of applying the makeup style from one picture to another, allowing for the modification of characters' makeup styles. To meet the diverse makeup needs of individuals or samples, the makeup transfer framework should accurately handle various makeup degrees, ranging from subtle to bold, and exhibit intelligence in adapting to the source makeup. This paper introduces a "3-level" adaptive makeup transfer framework, addressing facial makeup through two sub-tasks. 1. Makeup adaptation, utilizing feature descriptors and eyelid curve algorithms to classify 135 organ-level face shapes; 2. Makeup transfer, achieved by learning the reference picture from three branches (color, highlight, pattern) and applying it to the source picture. Just imagine replicating Audrey Hepburn's makeup in 1 second!
Code Review
Aspect-level cross-linguistic multi-layer sentiment analysis framework
LLM BERT-XLNET Sentiment Analysis Consumer Behavior Analysis NLP
Aspect-level cross-linguistic multi-layer sentiment analysis framework
As an important epidemic prevention product, the sales of protective masks are increasing day by day. What are people's preferences for different types of masks? For example, do they care more about the protective function or comfort of the mask, or the brand? For example, does this tendency have anything to do with gender, local economy, and geographical location? Starting from this topic, we built a highly granular commodity sentiment computing framework that includes sentiment polarity, intensity, and advanced deep learning and economic modeling methods to try to accurately restore consumers' true attitudes.
Code Review
AGO-I, A real-time consumer intelligent guidance system based on IoT and multi-tasking user portraits
Smart City User Portrait Computer Vision IoT
AGO-I, A real-time consumer intelligent guidance system based on IoT and multi-tasking user portraits
In the architecture of Society 5.0 actively promoted in Japan, the innovative retail industry is an essential component of the smart city, driving the high integration of network and physical space in the consumer industry. However, the AI facilities in various small and medium-sized stores in Japan are still at the level of some basic interactive applications. In order to improve the user's consumption experience and effectively promote the intelligence, customization and scientific management of the shopping system, we have designed a consumer intelligent guidance system based on deep learning and IoT technology. The system has two branches. 1. A high-granularity user profiling analysis framework; 2. A low-cost product management kit. Through this system, we can efficiently and accurately predict the age, gender, occupation and current physical state of consumers in real-time without infringing on personal privacy.
Code Review

Achievements

Kaggle TensorFlow (Help Protect the Great Barrier Reef)
Got Bronze Medal/TOP 6% in this game. Detect crown-of-thorns starfish in underwater image data.
Kaggle OTTO (Multi-Objective Recommender System)
Got Silver Medal/TOP 4% in this game. Build a recommender system based on real-world e-commerce sessions.
Kaggle UW (Madison GI Tract Image Segmentation)
Got Bronze Medal/TOP 6% in this game. Track healthy organs in medical scans to improve cancer treatment.
UCSanDiego Micro MBA Certificate
A Certificate of Completion from Institute for Business and Finance, Waseda University and Rady School of Management. The first step I took towards to the management.
Japan Science and Technology Agency Full Scholarship
I was awarded by the full scholarship for outstanding doctoral students initiated by JST.

Get in Touch

My inbox is always open. Whether you have a question or just want to say hi, I’ll try my best to get back to you!
Mail me