Description:

This site uses the openai interface (the latest gpt4o) to cater for users uploading excel chart analyses.
Distributed Rate Limiting - Limit single-user access frequency to prevent malicious use of system resources.
Asynchronous - Implement asynchronous AIGC using a custom I/O-intensive thread pool and task queue.
Distributed Message Queue - Use RabbitMQ to receive and persist task messages, forwarding them through a Direct exchange to a decoupled AI generation module for task processing.
Additionally, a dead-letter queue is created to handle tasks with AI Generation Error, enhancing system reliability.

Description:

This website features two main interfaces: a blog display page and an admin management page (access requires approval).
The primary function is to share problems encountered in my studies and daily life. The display page includes a complete login system and integrates Redis to update view counts in real-time.
Each blog post supports comments. Users also have their own user center where they can modify their attributes.
On the management page, users can publish posts, save and modify them multiple times, and control user permissions.
There are many valuable blogs on the site!

Description:

A WeChat mini program that combines on-campus forum, second-hand platform, and more. This project was completed in 2022.
We decided to develop a WeChat Mini Program featured in on-campus online forum and second-hand trading platform, and named it AllWhale (originally oxcean).
The reason for using WeChat Mini Program was that we wanted to popularize WeChat in Oxford.
The WeChat Mini Program Account: AllWhaleLYZ.
The server IP address: 34.41.117.214

Description:

Our primary objective is to develop a comprehensive tool that provides insights into how social media activities around top technology companies influence their stock market performance.
By examining likes, comments, and retweets, we aim to uncover trends and patterns that could suggest a relationship between social media sentiment and stock market movements.

Description:

This project aims to develop an AI system focused on interpreting images through the application of deep learning technologies, primarily Convolutional Neural Networks (CNNs) and Transformers.
The aim is to create a tool capable of processing and describing visual information, which could have practical implications in several areas such as aiding the visually impaired,
improving recommendation systems, and facilitating the integration of Computer Vision (CV) and Natural Language Processing (NLP). Initial results have shown promising performance,
with the model vividly and accurately describing the contents of images in test datasets, which highlights the feasibility of deploying this method in practical applications.
Provide a pdf of the paper, and code that can be reproduced

Description:

I built an interactive Spotify-themed web application using React and Node.js backed by a MySQL RDS database
The site was connected to my AWS mysql database, which stored information about each of Taylor’s songs and playlists.
I separated the back-end of the project and deployed it on different servers. The back-end used mp2 for continuous deployment of the node project
The project realizes the functions of paging query and data analysis.

Description:

A dataset of healthcare workers’ clothing is collected and processed, and the corresponding model is trained using yolo.
The model is used to assess the standard of dress and is commercially available in hospitals.Where due to the small sample, data enhancement was used to save on filming costs.
Due to the actual deployment environment limiting the yolo version, using a higher version can have better results.
Cannot provide the code…

Description:

Developed a simulation environment for indoor positioning in Python, preliminarily verified relevant algorithms, and collected pertinent training data sets.
Proposed intelligent indoor positioning algorithms to minimize positioning errors, detect NLOS, and save energy by disabling several sensors based on machine learning algorithms, i.e., LSTM neural networks, reinforcement learning, Double DQN, etc.
know more to see the PDF