> ML_LITERATURE_ATLAS_v1.0
Research Literature Atlas
253 qualified literature records from foundational statistical learning to frontier reasoning LLMs: verified DOIs, arXiv IDs, and original bilingual syntheses.
Nouvelles méthodes pour la détermination des orbites des comètes (Method of Least Squares)
Foundational treatise introducing the method of least squares to reconcile contradictory astronomical observations by minimizing sum of squared errors.
Theoria Motus Corporum Coelestium in Sectionibus Conicis Solem Ambientium
Seminal astronomical work rigorously formalizing the normal distribution and proving that least squares is the maximum likelihood estimator under Gaussian errors.
An Essay towards solving a Problem in the Doctrine of Chances
Posthumous essay introducing the concept of inverse probability, establishing what is now known globally as Bayes Theorem.
The Use of Multiple Measurements in Taxonomic Problems
Seminal paper introducing linear discriminant analysis to find linear combinations of features that maximize between-class variance relative to within-class variance.
The Regression Analysis of Binary Sequences
Classic paper establishing logistic regression by introducing the logit link function for binary response variables.
The Nature of Statistical Learning Theory
Monumental work establishing statistical learning theory, moving machine learning from empirical heuristics to rigorous mathematical generalization guarantees.
Support-Vector Networks
Seminal paper introducing the soft-margin Support Vector Machine, combining the kernel trick with quadratic optimization for robust classification.
Nearest Neighbor Pattern Classification
Classic paper providing the rigorous mathematical error bounds of nearest neighbor classification relative to the Bayes decision rule.
Random Forests
Foundational paper introducing Random Forests, proving that bagging randomized decision trees asymptotically prevents overfitting as forest size grows.
Classification and Regression Trees (CART)
The definitive monograph introducing the CART methodology for non-parametric classification and regression through binary tree structuring.
Greedy Function Approximation: A Gradient Boosting Machine
Foundational paper generalizing boosting to arbitrary loss functions by fitting trees greedily to negative gradient pseudo-residuals.
XGBoost: A Scalable Tree Boosting System
Highly influential systems paper presenting XGBoost, an engineered gradient boosting library delivering scalable second-order tree boosting.
LightGBM: A Highly Efficient Gradient Boosting Decision Tree
Pioneering work introducing LightGBM, speeding up gradient boosted tree training by orders of magnitude via histogram binning and GOSS.
CatBoost: unbiased boosting with categorical features
Influential paper introducing CatBoost, solving prediction shift and target leakage in gradient boosting through ordered boosting and oblivious trees.
Generalized Additive Models
Landmark statistical paper introducing GAMs, blending the interpretability of linear models with the flexibility of non-parametric splines.
InterpretML: A Unified Framework for Machine Learning Interpretability
Practical paper introducing EBM and InterpretML, delivering state-of-the-art glassbox models for high-stakes healthcare and financial lending decisions.
The Elements of Statistical Learning: Data Mining, Inference, and Prediction
Comprehensive 745-page bible of statistical machine learning covering linear models, kernel methods, tree ensembles, unsupervised learning, and model assessment.
Pattern Recognition and Machine Learning
The definitive Bayesian machine learning textbook introducing probabilistic graphical models, EM, variational inference, and MCMC.