Skip to content
YIZE
02 / 学术ACADEMIC

数学、统计,以及那些不肯配合的部分Mathematics, statistics, and the parts that refuse to cooperateMathematics, statistics, and the parts that refuse to cooperate数学、统计,以及那些不肯配合的部分

我做的东西落在应用统计和测量相交的地方:能估的就去估,估不了的就明确说出来。

My work sits where applied statistics meets measurement: estimating what can be estimated, and being explicit about the rest.

领域Field
统计学 · 统计学习Statistics · Statistical learning
单位Institution
中国农业大学China Agricultural University
学位Degree
数学本科 → 统计硕士B.Sc. Mathematics → M.Sc. Statistics
01 / 研究兴趣RESEARCH INTERESTS

研究兴趣Research InterestsResearch Interests研究兴趣

作物表型Crop Phenotyping

把原本只能手工量的东西从田里取出来——根系、冠层、一个生长季里的变化——而不是让一个人坐下来一株株数。

Getting measurements out of a field that would otherwise be taken by hand — root systems, canopy, growth over a season — without asking a person to sit and count.

农业图像Agricultural Imaging

在田间真实存在的条件下做检测与分割:叶片互相遮挡、光照一直在变,照片也永远不会像基准数据集那样干净。

Detection and segmentation under the conditions a field actually has: overlapping leaves, changing light, and photographs that will never be as clean as a benchmark.

统计学习Statistical Learning

一个被报告出来的提升,在固定评估协议之后还能剩下多少;以及一个小而吵的样本能诚实支撑多少信心。

How much of a reported improvement survives a fixed evaluation protocol, and how much confidence a small, noisy sample can honestly support.

有限群论Group Theory

从消失元的结构反过来刻画有限群;以及追问一个融合系统在什么条件下会从它的来源群继承超可解性。

Characterising finite groups from the structure of their vanishing elements, and asking when a fusion system inherits supersolvability from the group it came from.

科研工具Research Tooling

让别人能真的把方法跑起来的界面与流程。一个除了自己没人跑得起来的方法,还不算产出了结果。

Interfaces and pipelines that let the next person run the method. A method nobody else can run has not produced a result yet.

反复出现的那个论证The recurring argument

我大部分项目都会撞上同一个障碍:主导误差的那一项,不是模型能控制的那一项。

Most of my projects run into the same obstacle: the term that dominates the error is not the one the model controls.

1
The decomposition that decides the evaluation protocol before the model is chosen.
02 / 学术经历TIMELINE
4 个条目4 entries

学术经历TimelineTimeline学术经历

  1. 2023

    开始读数学Started a degree in Mathematics

    中国农业大学理学院。第一次真正接触证明,也第一次意识到「理解了」和「会用」是两回事。

    China Agricultural University, College of Science. First real encounter with proof, and with the fact that understanding something is different from being able to use it.

    began
  2. 2024

    第一次认真做科研First serious research work

    从课程作业转向真实问题:图像数据、表型测量,以及在结果可信之前会坏掉的那一长串东西。

    Moved from coursework into applied problems: imaging data, phenotype measurement, and the long tail of things that break before a result is trustworthy.

    research
  3. 2025

    为科研写软件Building software for research

    明白了一个没人跑得起来的方法还不算方法。于是开始在做分析的同时写工具、写流程、写界面。

    Learned that a method nobody can run is not yet a method. Started writing tools, pipelines and interfaces alongside the analysis.

    software
  4. 2026

    统计,以及这个档案室Statistics, and this archive

    准备继续读统计。搭这个空间,是为了让这些东西有一个比笔记本电脑里某个文件夹更久的去处。

    Preparing to continue into statistics. Building this space so that the work has somewhere to live beyond a folder on a laptop.

    now
03 / 科研项目SELECTED RESEARCH

科研项目Selected ResearchSelected Research科研项目

  • 从图像序列中恢复根系结构Root structure recovery from image sequences

    2025 — presentongoing

    在不逐帧人工标注的前提下,低成本采集的图像里能还原出多少细根结构?

    How much fine-root structure can be recovered from low-cost captures without manual annotation of every frame?

  • 融合系统与超可解性Fusion systems and supersolvability

    2025writing

    在 F_S(G)-融合系统中,哪些结构性条件会强制超可解性?

    Which structural conditions force supersolvability in F_S(G)-fusion systems?

  • 植物病害分割中的基准稳定性Benchmark stability in plant disease segmentation

    2025complete

    固定评估协议之后,已发表的提升还能剩下多少?

    How much of a reported gain survives a fixed evaluation protocol?

04 / 论文与成果PUBLICATIONS

论文与成果PublicationsPublications论文与成果

Plant-level phenotyping of maize aerial root tiers from field video using occlusion-aware multi-frame fusion

2026

Li, Y., et al.

Computers and Electronics in Agriculture (Q1) — revise-and-resubmit recommended; in revision

Aerial roots are counted by hand, and counting them from video is unreliable: the roots overlap, occlude each other, and change appearance between frames. This work treats a sequence of captures as evidence rather than as one noisy observation. A cascade — EfficientNet-B3 for tier classification, YOLOv10 for detection, and an occlusion-aware multi-frame fusion step — lifted accuracy from under 50% to 88.9% over a structured set of more than ten thousand plants.

修改中revising第一作者first author

Application of ECV-UNet model in semantic segmentation of wheat stripe rust

2025

Li, Y., et al.

ICSIP 2025 · 10th International Conference on Signal and Image Processing · IEEE · Ei Compendex and Scopus

Stripe rust presents as thin, low-contrast lesions against a cluttered canopy — the case where a plain U-Net loses the boundary first. ECV-UNet keeps the U-Net backbone and changes two things: an attention module in the skip connections, and a Transformer block at the bottleneck.

已见刊published共同第一作者joint first author

New characterizations of G2(3^m) via its vanishing elements

2025

Li, Y., et al.

Czechoslovak Mathematical Journal (Springer) · SCIE-indexed · submitted March 2025

A vanishing element is one on which some irreducible character of the group takes the value zero, and the set of them constrains the group surprisingly tightly. This paper uses that set, together with the prime graph built from it, to give new characterisations of the Chevalley groups G2(3^m).

在审under review共同第一作者joint first author
05 / 教育经历EDUCATION

教育经历EducationEducation教育经历

  1. 2023 — 2027 (expected)

    数学与应用数学 · 本科B.Sc. in Mathematics

    中国农业大学 · 理学院China Agricultural University · College of Science

    主修分析、代数与概率,偏向应用统计。

    Core work in analysis, algebra and probability, with a concentration in applied statistics.

  2. 2027 (expected)

    统计学 · 硕士M.Sc. in Statistics

    中国农业大学China Agricultural University

    拟继续深造,方向是统计推断与面向科学数据的方法。

    Planned continuation, focused on statistical inference and methods for scientific data.

06 / 技能SKILLS
面试时我会这么写what I would claim in an interview

技能SkillsSkills技能

数学Mathematics
  • Real analysis
  • Probability
  • Linear algebra
  • Group theory
统计Statistics
  • Inference
  • Regression
  • Resampling
  • Model selection
编程Programming
  • Python
  • R
  • TypeScript
  • SQL
  • MATLAB
科研Research
  • Experimental design
  • Image analysis
  • Reproducible workflows
  • LaTeX
07 / 学术笔记ACADEMIC NOTES

学术笔记Academic NotesAcademic Notes学术笔记

看看做出来的东西See what came out of it