| Paper | Why |
|---|---|
| Sampling techniques | Stratifield sampling is popular. |
| A Comparative Study of Efficient Initialization Methods for the k-means Clustering Algorithm and kmeans initialization, Coursera | Fundamentals about kemans, favourite topics in interviews i.e: LinkedIn |
| A Comprehensive Survey of Clustering Algorithms | Clustering fundamentals |
| A Tutorial on Spectral Clustering | Sepctral clustering is intuitive and quite popular. |
| Partial residual plot | Useful for model diagnosis. |
| Compare GINI index and Information Gain | Intuition behind Decision Tree, RandomForest |
| Explain tf-idf | Fundamentals about tf-idf. |
| Understanding L-BFGS | Advance about optimization, rarely asked in interview |
| Optimizer Quasi newton method | Advance about optimization. |
| Paper | Why |
|---|---|
| Understanding the Difficulty of Training Deep Feedforward Neural Networks | Classic paper (2010) about initialization, sigmoid etc |
| Delving Deep into Rectifiers - Surpassing Human-Level Performance on ImageNet Classification | Classic paper (2015) about ReLU, PReLU |
| Batch Normalization - Accelerating Deep Network Training by Reducing Internal Covariate Shift | Classic paper about BatchNorm |
| Dropout - A Simple Way to Prevent Neural Networks from Overfitting | Classic paper about Dropout |
| Deep Residual Learning for Image Recognition | Classic ResNet |
| On Large-Batch Training for Deep Learning - Generalization Gap and Sharp Minima | Practical technique for large batch training |
| Paper | Why |
|---|---|
| Calibration in modern neural network | Important topics in ML system design i.e: facebook |
| Attention model | Fundamentals in Attention, powerful architecture in NLP |
| Ilya's thesis | Network in network |
| Paper | Why |
|---|---|
| Adaptive Importance Sampling to Accelerate Training of a Neural Probabilistic Language Model | Classic paper (2008) in NLP |
| Natural Language Processing (Almost) from Scratch | Classic paper (2011) in NLP |
| Word2vec | The classic paper in NLP, still popular in industry: Uber, DoorDash, Twitter etc |
| GloVe - Global Vectors for Word Representation | Classic paper (2014) in NLP |
| Bag of Tricks for Efficient Text Classification | Cool tricks in NLP tasks |
| BERT - Pre-training of Deep Bidirectional Transformers for Language Understanding | The famous BERT |
| Smart Reply - Automated Response Suggestion for Email | NLP application, useful for ML system design |
| Enriching Word Vectors with Subword Information | Simple and fast method to train NLP task in Facebook |
| Neural Approaches to Conversational AI | Comprehensive survey (2018) about chatbots |
- If you're interested to learn more about paid ML system design course on educative.io with more examples, click here.
- If you find this helpful, you can Sponsor this project. It's cool if you don't.