About Me

Personal Information

Nickname: Yang Sier
Email: hobart.yang@qq.com
Website: discover304.top
Motto 1: Everything is for the best.
Motto 2: If you are fragrant, the butterfly will come.
Motto 3: If you can’t find me, I’m either causing trouble or on my way to cause trouble.
Bio: Third-year student majoring in Artificial Intelligence; academically excellent; has a wide range of interests and is full of curiosity towards the unknown, coupled with an insatiable desire to learn; willing to try new things and has experience in organizing campus charitable activities and startups; positive and cheerful, aims to inspire those around them while in a team leadership or management role.

Skills

Proficient in Python, Java, Haskell programming.
Experienced in Linux system operations.
Knows Git as well as GitHub/Gitee/GitLab setup and usage.
Familiar with TensorFlow2 and PyTorch frameworks, understands PaddlePaddle framework.
Proficient in MMDetection3D framework.
Knowledgeable in Convolutional Neural Networks (CNN), Graph Neural Networks (GNN), Generative Adversarial Networks (GAN), and Flow Models; understands YOLO object detection algorithm and Bert model.
Familiar with MySQL database usage.
Understands static website development, HTML+CSS+JS stack, and Electron js web application packaging methods.
Basic knowledge in front-end, back-end, and networking; can implement simple front-back end interactions and database connections.

Education

2019/09 - Present: University of Edinburgh

Awards

2021: 🏅 Participated in Google Hash Code programming competition, team ranked 3755
2018: 🏆 Top 10 in Physics Bowl, Central China region
2018: 🥉 Bronze in British Physics Olympiad
2016: 🥈 VEX EDR Robot Asia Competition, team won silver and excellent design award
2016: 🥇 First place in Underwater Robot competition at the China Robot Convention

Experience

2022/03 - 2022/08

🧢 Research Project Eloss in the way: A Sensitive Input Quality Metrics for Intelligent Driving (arxiv.org)
Lead author; designed a sensor input quality metric based on the expected change in information. Achieved quality evaluation without affecting model accuracy.
Technologies Used: PyTorch, MMDetection3D, CNN/GNN/VoxelNet/PointPillars, Jupyter Lab, Docker cloud-based training, GPU cloud servers.

2021/08 - 2021/10

🧢 Research Project SinGlow: Singing Voice Synthesis with Glow - Help Virtual Singers More Human-like (ICAICE 2021)
Solo lead author; developed song feature extraction based on flow models and performed style transfer. Partially succeeded in transferring real human singing features to virtual singers.
Technologies Used: TensorFlow 2, librosa audio processing library, Glow flow models, Jupyter Notebook, Google Colab.

2021/07 - 2021/09

🧢 Research Project A New Breast Cancer Diagnosis Application Based on ResNet50 (IYSF-AICE 2021)
Co-first author; explored cancer detection applications based on ResNet50. Compared various deep learning models and implemented the application.
Technologies Used: TensorFlow 2, ResNet, Jupyter Notebook, PyQt for visualization.

2021/04

🎓 Coursework Project Maze Chess AI
A small game where the hero uses items in a maze to kill monsters and save the princess.
Used PDDL language to build a Logic Agent (AI) that plans the hero’s actions to achieve the goal of saving the princess.

2021/04 - 2021/05

🧢 Personal Project QQ Chatbot implemented in Java.
Used the open-source Mirai framework to build a QQ bot, integrating Turing AI for chat functionality.
Deployed on Microsoft Azure cloud server, later moved to Alibaba Cloud.

2021/01 - 2021/03

🧢 Personal Project Feasibility Analysis for AI Vocaloid Training, implemented in Python.
Investigated the feasibility of using machine learning techniques to train virtual singers by selecting high-correlation dimensions from a large dataset to reduce computation.

2020/04 - 2020/07

💼 Startup, Online Student Study Room (Universal Student Service Platform)
Team management and test operation lead, wrote the business plan, recruited members.

2020/04 - 2020/05

🎓 Charitable Activity, Organized and operated an online student study room based on Zoom + Facebook/Teams/WeChat/QQ.
Received support from the School of Informatics at the University of Edinburgh.
Over 100 students from the University of Edinburgh and the University of Toronto participated.

Hobbies

Writing my own blog and enriching its content.
Learning new knowledge and skills from the internet.
Enthusiastically helping classmates with academic and life issues.
Designing small inventions based on Arduino or Raspberry Pi and fabricating them through 3D printing.

Research Interest

Multi-Agent System, Federate Learning: focusing on robotics and math-intensive proofs for safe and private AI. I am interested in the interplay between these technologies and how they can be used to create more secure and efficient systems.
Multi-Modal Fusion, Deep Learning Interpretability, and Reliability: developing robust AI systems that combine information from different sources, including image, point cloud, text, and speech. I aim to investigate methods to enhance the interpretability and reliability of these systems, making them more practical and trustworthy for real-world applications.

个人信息

昵称: 杨丝儿
邮箱: hobart.yang@qq.com
网站: discover304.top
座右铭1：一切都是最好的安排。
座右铭2：你若芬芳，蝴蝶自来。
座右铭3：你若找不到我，那我不是在搞事情，就是在去搞事情的路上。
简介: 人工智能专业大三，在校成绩优异；兴趣广泛，对未知事物充满好奇，并拥有不停学习的动力；勇于尝试，有组织校内公益活动以及创业经历；阳光开朗，在团队中作为调节或者管理的职位，努力以自己的心态感染身边的人。

技能

掌握 Python、Java、Haskell 编程。
掌握 Linux 系统操作。
掌握 Git 以及 GitHub/Gitee/GitLab 的配置与使用。
熟悉 TensorFlow2 和 PyTorch 框架，了解PaddlePaddle 框架。
熟悉 MMDetection3D 框架。
熟悉深度学习中的卷积神经网络CNN、图神经网络GNN、对抗生成网络GAN、流模型Flow、了解 Yolo 目标检测算法、Bert 模型。
熟悉 MySQL 数据库的使用。
了解静态网站开发，了解 HTML+CSS+JS 模式，了解 Electron js 网页软件打包方法。
了解前后端以及网络通信基础知识，能实现简单的前后端交互以及与数据库之间的连接。

学历

2019/09 - 现在: University of Edinburgh (爱丁堡大学)

获奖

2021: 🏅 参加 Google Hash Code 编程比赛，小组排名3755
2018: 🏆 Physics Bowl (物理碗) 中国中部地区前十
2018: 🥉 British Physics Olympiad (英国奥林匹克物理竞赛) 铜奖
2016: 🥈 VEX EDR 机器人亚洲赛小组银奖和优秀设计奖
2016: 🥇 中国机器人大会水中机器人大赛小组第一

经历

2022/03 - 2022/08

🧢 论文项目 Eloss in the way: A Sensitive Input Quality Metrics for Intelligent Driving (arxiv.org)
第一作者，一种参考通信模型设计的，使用信息量期望变化作为标准的，传感器输入信息质量指标。在不影响模型精度的同时，实现了输入信息质量的判断。
开发技术：PyTorch 框架 + MMDetection3D 目标检测框架 + CNN/GNN/VoxelNet/PointPillars 深度学习模型 + Jupyter Lab 开发环境 + Docker 云端训练部署 + GPU 计算云服务器

2021/08 - 2021/10

🧢 论文项目 SinGlow: Singing Voice Synthesis with Glow - Help Virtual Singers More Human-like (ICAICE 2021)
独立第一作者，基于流模型的歌曲特征提取，并进行风格迁移。一定程度上实现了将真实人声歌曲的特征迁移到虚拟歌手的歌曲上。
开发技术：TensorFlow 2框架 + librosa 音频处理库 + Glow 深度学习流模型 + Jupyter Notebook 开发环境 + Google Colab 训练平台

2021/07 - 2021/09

🧢 论文项目 A New Breast Cancer Diagnosis Application Based on ResNet50 (IYSF-AICE 2021)
并列一作，基于 ResNet50的癌症检测应用探索。对比分析了多种深度学习模型，并通过方法上的创新实现了模型的应用落地。
开发技术：TensorFlow 2框架 + ResNet 等深度学习模型 + Jupyter Notebook 开发环境 + PyQt 可视化界面

2021/04

🎓 作业项目 迷宫棋游戏AI
英雄利用迷宫中的道具杀怪物救公主的小游戏
使用 PDDL 语言搭建 Logic Agent (逻辑AI) 计划英雄行为流程，达到拯救公主的目标

2021/04 - 2021/05

🧢 个人项目 基于QQ的聊天机器人，Java语言实现。
采用开源 Mirai 框架，搭建QQ机器人。并调用图灵AI接口实现对话聊天。
项目部署在微软 Azure 云服务器，后转到阿里云服务器。

2021/01 - 2021/03

🧢 个人项目 虚拟歌手AI调教可行性分析，Python语言实现。
虚拟歌手AI调教计划使用机器学习技术实现，可行性分析是在探索输入数据集，从大量的维度中选择存在高相关性，减少计算量。

2020/04 - 2020/07

💼 创业，线上学生自习室 (通用学生服务平台)
小组管理以及测试运营负责人，策划书撰写，招募成员

2020/04 - 2020/05

🎓 公益活动，组织并运营基于 Zoom + Facebook/Teams/微信/QQ 的线上学生自习室。
基于 Zoom + Facebook/Teams /微信/QQ 的线上学生自习室。运营期间获得了来自爱丁堡大学信息学院的支持和帮助。
超过100名来自爱丁堡和多伦多大学的学生参加学习室活动。

爱好

撰写我自己的博客，丰富其内容
从网络上学习新的知识，获得新的技能
热心帮助同学解决学习以及生活上的问题
设计基于 Arduino 或 Raspberry Pi 的小发明，并使用3D打印的方式制作

個人情報

ニックネーム: 楊丝儿（ヤン・シー・アール）
メール: hobart.yang@qq.com
ウェブサイト: discover304.top
座右の銘1: すべては最善の配置です。
座右の銘2: 花が香れば、蝶が来る。
座右の銘3: 私を見つけられない場合、私は何かをしているか、それをする途中です。
自己紹介: 人工知能の専門、3年生、優秀な学業成績；広範な興味、未知に対する好奇心あり、学習への意欲も高い；積極的に新しいことに挑戦、学内の公益活動や起業も経験。