Skip to content
YIZE
01 / 项目PROJECTS

做过的东西,以及哪里做错了Things I have built, and what went wrongThings I have built, and what went wrong做过的东西,以及哪里做错了

这一栏原来叫「作品」,比它需要的含糊。这里放的都是真的做出来的东西——一个参赛项目、一款软件、一个网站——每一条都按案例来写:问题是什么、做过哪些决定、哪一部分没成。截图是一个项目里最没意思的东西。

This section used to be called Work, which was vaguer than it needed to be. Everything here is a thing that got made — a competition entry, a piece of software, a site — and each one is written up as a case study: the problem, the decisions, and the parts that did not work. Screenshots are the least interesting thing a project has to offer.

条目Entries
11
检索标签Tags
竞赛 · 软件 · 网页 · 科研Competition · Software · Web · Research
排序Order
人工挑选,不按时间Curated, not chronological

11 / 11

P.00101

育种家|全流程一站式智慧育种综合平台Yuzhongjia · One-Stop Intelligent Breeding Platform

Yuzhongjia · One-Stop Intelligent Breeding Platform育种家|全流程一站式智慧育种综合平台

把一次真实试验从田间记录一路带到站得住的结论的育种工作台——采、选、洗、算、用五个环节,一条证据链,版本从 V1.0 迭代到了 V13.0。

A breeding workbench that carries a real experiment from field record to defensible conclusion — five stages, one evidence chain, and a version history that now runs from V1.0 to V13.0.

角色Role
项目负责人Project lead
类型Type
软件Software · 竞赛Competition
年份Year
2023 — 2026
  • Python
  • PyTorch
  • Workflow orchestration
  • Data governance
查看详情View case study/ View case study/ 查看详情
P.00202

理小助|一站式校园服务智能体Li-Xiaozhu · Campus Service Agent

Li-Xiaozhu · Campus Service Agent理小助|一站式校园服务智能体

2026 年 6 月下旬起在中国农业大学理学院投入使用的校园智能体:课表、作业、通知、校园事务收进同一个对话框,每个回答都能点回出处。

A campus agent in production at CAU's College of Sciences since June 2026: one conversation box for timetables, deadlines, notices and campus business, with every answer traceable to its source document.

角色Role
核心开发者 · 参与全流程Core developer, full pipeline
类型Type
网页Web
年份Year
2026 —
  • LLM agent
  • RAG
  • OCR pipeline
  • Web
查看详情View case study/ View case study/ 查看详情
P.00303

玉米气生根层级识别|遮挡感知的田间视频表型Maize Aerial Root Tiers · Occlusion-Aware Video Phenotyping

Maize Aerial Root Tiers · Occlusion-Aware Video Phenotyping玉米气生根层级识别|遮挡感知的田间视频表型

从田间视频里逐株数清气生根的层数。真实遮挡下,单帧检测的准确率只有 36.6–45.1%;本文的遮挡感知多帧融合把株级准确率提到 88.9%。

Counting aerial root tiers per plant from field video. Single-frame detection sits at 36.6–45.1% accuracy under real occlusion; the occlusion-aware multi-frame fusion described here takes plant-level accuracy to 88.9%.

角色Role
第一作者First author
类型Type
科研Research
年份Year
2025 — 2026
  • Python
  • PyTorch
  • YOLOv10
  • ByteTrack
  • EfficientNet-B3
查看详情View case study/ View case study/ 查看详情
P.00404

根系动态生长表型|测量方法与发明专利Root Growth Phenotyping · Measurement System and Invention Patent

Root Growth Phenotyping · Measurement System and Invention Patent根系动态生长表型|测量方法与发明专利

一项已授权(不是受理中)的发明专利:从时序图像高通量测量根系动态生长——时序差分、骨架提取、最短路径定长,最终长成生长曲线。

An invention patent — granted, not pending — for a high-throughput method that measures root length over time from image sequences: temporal differencing, skeleton extraction and shortest-path rooting, assembled into growth curves.

角色Role
发明人之一 · 测量软件独立完成Inventor · measurement software built independently
类型Type
科研Research
年份Year
2025 — 2026
  • Computer vision
  • Image calibration
  • Skeleton extraction
  • Python
查看详情View case study/ View case study/ 查看详情
P.00505

智能集成机器学习建模与可视化分析平台Integrated Machine-Learning Modelling and Visualisation Platform

Integrated Machine-Learning Modelling and Visualisation Platform智能集成机器学习建模与可视化分析平台

把常用的建模与可视化分析收进一个平台的登记软件(软著 2025SR1800232),是育种家平台三项软著组件之一。

A registered platform (软件著作权 2025SR1800232) that wraps standard modelling and visual analysis into one place — one of the three copyrighted components underneath the breeding platform.

角色Role
开发成员Developer
类型Type
软件Software
年份Year
2025
  • Python
  • Machine learning
  • Visualisation
查看详情View case study/ View case study/ 查看详情
P.00606

自适应可视多平台主成分分析软件Adaptive Multi-Platform PCA for Very Wide Matrices

Adaptive Multi-Platform PCA for Very Wide Matrices自适应可视多平台主成分分析软件

能在 GWAS 量级的基因型矩阵上活着跑完的主成分分析:分块读取、CPU/GPU 双模式、三种主成分数确定方式,计算过程全程可视。

Principal component analysis that survives GWAS-scale genotype matrices: block-wise reading, CPU or GPU execution, three ways to choose the number of components, and a visual trace of the computation while it runs.

角色Role
开发成员Developer
类型Type
软件Software
年份Year
2024
  • Python
  • GPU computing
  • Blocked algorithms
  • PCA
查看详情View case study/ View case study/ 查看详情
P.00707

面向多需求的生物数据处理软件Bio-Data Processing · From HMP to Analysis-Ready Matrices

Bio-Data Processing · From HMP to Analysis-Ready Matrices面向多需求的生物数据处理软件

GWAS 里不起眼的那一半工作的登记软件(软著 2024SR1711525):hmp 转纯数阵、数字矩阵插值、正定分块协方差矩阵、定比随机分组。

A registered utility (软件著作权 2024SR1711525) for the unglamorous half of GWAS: HMP-to-matrix conversion, imputation, positive-definite block covariance matrices, and proportional random grouping.

角色Role
开发成员Developer
类型Type
软件Software
年份Year
2024
  • Python
  • Bioinformatics
  • GWAS preprocessing
查看详情View case study/ View case study/ 查看详情
P.00808

基于天球理论的可视化日照时长预测软件Solar Duration from the Celestial Sphere

Solar Duration from the Celestial Sphere基于天球理论的可视化日照时长预测软件

预测某栋建筑、某个点日照时长的桌面软件(软著 2025SR1116569):用天球几何加上真实遮挡物的形状算阳光,单日与全年两种可视化模式。

A desktop tool (软件著作权 2025SR1116569) that predicts sunshine duration for a building or a point, from celestial-sphere geometry and the actual shapes that block the sky — with single-day and whole-year visual modes.

角色Role
开发成员Developer
类型Type
软件Software
年份Year
2025
  • Python
  • Computational geometry
  • α-shape
  • Qt GUI
查看详情View case study/ View case study/ 查看详情
P.00909

同质与异质集成方法在预测与特征选择上的比较Homogeneous and Heterogeneous Ensembles for Prediction and Feature Selection

Homogeneous and Heterogeneous Ensembles for Prediction and Feature Selection同质与异质集成方法在预测与特征选择上的比较

对同质与异质集成方法在预测与特征选择上的系统比较,投至 ACML 2026、在审中——rebuttal 已交,审稿人的问题已逐条回应。

A systematic comparison of ensemble methods for prediction and feature selection, under review at ACML 2026 — with the rebuttal written and the reviewers' questions answered.

角色Role
第七作者Seventh author
类型Type
科研Research
年份Year
2026
查看详情View case study/ View case study/ 查看详情
P.01010

基于特征标零点素图的零散单群刻画Characterising Sporadic Groups from the Vanishing-Element Prime Graph

Characterising Sporadic Groups from the Vanishing-Element Prime Graph基于特征标零点素图的零散单群刻画

从消失元构成的素图出发,刻画部分零散单群与 Chevalley 群。这项工作带来了一篇并列第一作者、目前在审的论文。

Starting from the graph built on vanishing elements, this project characterises some sporadic simple groups and Chevalley groups — work that produced a joint first-author paper now under review.

角色Role
核心成员Core member
类型Type
科研Research
年份Year
2025 — 2026
查看详情View case study/ View case study/ 查看详情
P.01111

个人数字空间Personal Digital Space

Personal Digital Space个人数字空间

就是本站。它被当作一个要长期存在的档案室来搭,而不是一份作品集;文档也写好了,让下一版比这一版更好写。

This website. Built as a long-lived archive rather than a portfolio, and documented so that the next version is easier to write than this one was.

角色Role
设计与开发Design and development
类型Type
软件Software · 网页Web
年份Year
2026
  • Next.js
  • TypeScript
  • Tailwind CSS
  • MDX
  • KaTeX
查看详情View case study/ View case study/ 查看详情