Research & ApplicationsMultilingual AIEN
Data Scaling for LoRA Domain Adaptation
A research note on how to study the effect of domain data size in LoRA-based machine translation adaptation.
Foundations + Research
Xiaojing Yangfrom foundations to applied systems
A bilingual knowledge base on mathematics, statistics, machine learning, NLP, LLMs, model evaluation, and applied AI research. The goal is simple: make the foundations reusable, then connect them to real experiments.
Foundations
The concepts I want to keep sharp
This track is where existing high-quality resources become my own explanations, notes, examples, and interview-ready mental models.
Mathematics
Linear algebra, calculus, optimization, graphs, and representation spaces explained through ML and NLP examples.
Statistics
Probability, distributions, confidence intervals, bootstrap testing, hypothesis tests, and uncertainty in evaluation.
Machine Learning
Splits, generalization, model selection, cross-validation, losses, regularization, and practical training workflows.
NLP & LLMs
Tokenization, embeddings, Transformers, fine-tuning, prompting, sequence models, and language-specific evaluation.
Research Engineering
Experiment repositories, config-driven pipelines, logging, annotation sheets, demos, and reproducible analysis.
Research & Applications
Broad enough for future work, concrete enough for real projects
This track avoids boxing the blog into one thesis topic. It groups research by capability: building, evaluating, explaining, retrieving, and applying AI systems.
Multilingual AI
Machine translation, multilingual NLP, low-resource adaptation, terminology, domain adaptation, and language variation.
Model Evaluation
Metrics, benchmark design, statistical testing, ablations, error analysis, human evaluation, and robustness.
Explainability & Responsible AI
Interpretability, bias analysis, attribution, model auditing, SHAP/LIME, Shapley values, and controlled evaluation.
Retrieval & Knowledge Systems
Information retrieval, RAG, long documents, graph retrieval, evidence grounding, and document intelligence.
Applied ML Systems
Data pipelines, classification, forecasting, information extraction, deployment, product demos, and applied analytics.
Latest notes
View all posts »Posts may be in English, Chinese, or bilingual. Foundations posts often start from strong public resources; research posts connect those foundations to my own experiments.
Research & ApplicationsMultilingual AIEN
My LoRA NMT Project: Research Question, Method, Findings, and Limitations
A research-story version of my LoRA NMT project: problem framing, data, method, evaluation, findings, limitations, and future work.
Research & ApplicationsModel EvaluationEN
Terminology as an Evaluation Problem in Domain MT
Why terminology should be evaluated explicitly in domain machine translation, especially when automatic metrics can hide critical technical errors.
Research & ApplicationsMultilingual AIEN
Why LoRA Fits Low-Resource Domain Adaptation
A research-facing explanation of why LoRA is a good fit for low-resource domain machine translation: controlled adaptation, lower experimental cost, and reduced overfitting risk.
Research & ApplicationsMultilingual AI中文
LoRA 领域适配中的数据规模实验
如何研究领域数据规模对 LoRA 机器翻译适配的影响：数据多少才够，什么时候收益递减，什么时候是数据质量问题。
Research & ApplicationsMultilingual AI中文
我的 LoRA NMT 项目：研究问题、方法、发现与局限
用研究叙事方式解释我的 LoRA NMT 项目：问题定义、数据、方法、评估、发现、局限与未来工作。
This blog complements my portfolio.
Portfolio pages show what I built. These notes explain the foundations, decisions, experiments, and tradeoffs behind the work.