Logistic Regression, L1, L2 regularization, Gradient/Coordinate descent
MLE v.s. MAP ~ L1, L2 Math Derivation
XGBoost math Derivation
Introduction to Convex Optimization Basic Concepts
Convex Optimization: Primal Problem to Dual problem clearly explained
SVM/ Dual SVM math derivation, non-linear SVM, kernel function
EM (Expectation–Maximization) Algorithm
Gaussian Mixture Model (GMM) & K-means derivation
Intro to Deep Learning & Backpropagation
Knowledge Distillation clearly explained
Intro to NLP & Spell Correction with Noisy Channel
Distributed representation, Hyperbolic Space, Gaussian/ Graph Embedding clearly explained
Skip-gram clearly explained
Hidden Markov Model (HMM) derivation
Probabilistic Graphical Model (PGM) clearly explained
Log-Linear Model & Conditional Random Field (CRF) derivation
RNN, LSTM clearly explained
Transformer step-by-step explaination (slides)
BERT and RoBERTa clearly explained
What Does BERT Look At? An Analysis of BERT’s Attention
Common Multi-lingual Language Models (M-Bert, LASER, MultiFiT, XLM) clearly explained
Cross-lingual learning for text processing: A survey
Don’t Stop Pretraining: Adapt Language Models to Domains and Tasks (slides)
A Review and Outlook for Relation Extraction
Document-level Neural Relation Extraction with Edge-oriented Graphs
Introduction to Graph Neural Network
Graph Convolutional Neural Network - Spatial Convolution
Graph Convolutional Neural Network - Spectral Convolution
The More You Know: Using Knowledge Graphs for Image Classification
XRayGAN: Consistency-preserving Generation of X-ray Images from Radiology Reports (slides)
An image is worth 16x16 words: Transformers for image recognation at scale (slides)
Decoupling Representation and Classifier for Long-Tailed Recognition
Kaggle: Google Quest Q&A Labeling, my first Silver Medal summary
Kaggle: Tweet Sentiment Extraction, Part 1/2: common methods summary
Kaggle: Tweet Sentiment Extraction, Part 2/2: top solutions summary
Kaggle: Jigsaw Multilingual Toxic Comment Classification, top solutions summary
