Skip to content

Papers

A comprehensive, organized collection of foundational and cutting-edge papers in machine learning, deep learning, and generative AI. Papers are organized into 13 chapters by topic and era.

Chapters

00. Classic Foundations (Pre-2012)

Foundational papers in machine learning, information retrieval, and neural networks.

Topics: Word embeddings, ML algorithms, information retrieval, sequence modeling, computer vision, Bayesian methods

Key Papers: Backpropagation, SVM, Decision Trees, LeNet, TF-IDF, PageRank

01. Deep Learning Era (2012-2016)

The revolution that transformed AI through deep neural networks at scale.

Topics: Image classification, object detection, training fundamentals, generative models, reinforcement learning, meta-learning

Key Papers: AlexNet, VGGNet, ResNet, R-CNN, YOLO, GANs, DQN, AlphaGo

02. Transformer Architectures (2017-2023)

The foundational transformer architecture and efficient variants.

Topics: Core transformer, BERT, GPT, position embeddings, efficient attention

Key Papers: Attention Is All You Need, BERT, GPT-2, T5, FlashAttention, RoPE, ALiBi

03. Large Language Models (2020-2024)

Scaling transformers to billions of parameters.

Topics: GPT series, open-source models, alignment approaches

Key Papers: GPT-3, GPT-4, LLaMA, Llama 2, Mistral, Constitutional AI

04. Training Techniques & Optimization (2019-2023)

Methods for efficiently training and adapting large language models.

Topics: Parameter-efficient fine-tuning, supervised fine-tuning, alignment, optimization

Key Papers: LoRA, QLoRA, Adapters, Prefix Tuning, SFT, RLHF, DPO, Flan Collection

05. Inference & Deployment (2017-2024)

Techniques for efficient inference, quantization, and serving LLMs.

Topics: Quantization, decoding strategies, serving, memory optimization

Key Papers: GPTQ, AWQ, Speculative Decoding, Medusa, vLLM

06. Retrieval & Knowledge Integration (2019-2023)

Methods for augmenting LLMs with external knowledge.

Topics: RAG, dense retrieval, embeddings, knowledge bases

Key Papers: RAG, DPR, Sentence-BERT, When Not to Trust LMs

07. Multimodal & Vision (2017-2023)

Vision models and vision-language models for multimodal AI.

Topics: Efficient vision architectures, object detection, vision transformers, vision-language models

Key Papers: MobileNet, EfficientNet, ViT, CLIP, LLaVA, GPT-4V

08. Reasoning & Interpretability (2013-2023)

Methods for improving model reasoning and understanding decisions.

Topics: Chain-of-thought, tree of thoughts, interpretability, attention analysis

Key Papers: CoT Prompting, Tree of Thoughts, Self-Consistency, Attention Is Not Explanation

09. Safety & Alignment (2013-2023)

Techniques for ensuring AI systems are safe and aligned.

Topics: Safety evaluation, adversarial robustness, jailbreaks, prompt injection

Key Papers: Holistic Evaluation, TruthfulQA, Adversarial Examples, Prompt Injection

10. Benchmarking & Evaluation (2004-2023)

Benchmark datasets and evaluation frameworks.

Topics: NLU benchmarks, knowledge benchmarks, evaluation metrics

Key Papers: GLUE, SuperGLUE, MMLU, MT-Bench, BERTScore, ROUGE

11. Prompting & In-Context Learning (2020-2021)

Techniques for leveraging prompts and in-context learning.

Topics: Few-shot learning, prompt-based learning, zero-shot transfer

Key Papers: Few-Shot Learners, Prompt Learning

12. Emerging Topics (2015-2023)

Cutting-edge research directions and future opportunities.

Topics: Long context, agent systems, knowledge distillation

Key Papers: LongNet, ReAct, Generative Agents, Knowledge Distillation

How to Use This Reference

By Goal

Goal Path
Understand AI fundamentals 00. Classic Foundations
Learn deep learning 01. Deep Learning Era
Build LLM applications 02. Transformers → 04. Training
Deploy LLMs efficiently 05. Inference & Deployment
Add knowledge to LLMs 06. Retrieval & Knowledge
Multimodal applications 07. Multimodal & Vision
Improve model reasoning 08. Reasoning & Interpretability
Ensure safety & alignment 09. Safety & Alignment
Evaluate models 10. Benchmarking & Evaluation
Explore cutting edge 12. Emerging Topics

By Timeline

  1. Pre-2012: Classic Foundations
  2. 2012-2016: Deep Learning Era
  3. 2017-2019: Transformer Architectures + Early LLMs
  4. 2020-2023: LLMs, Training, Multimodal
  5. 2023-2024: Emerging Topics

Tips for Reading Papers

  1. Read the abstract first to understand the contribution
  2. Look at figures and tables before diving into text
  3. Check related work for context with other papers
  4. Understand the problem statement before the solution
  5. Focus on key insights, not every technical detail
  6. Follow the citation chain to understand the research lineage

Statistics

  • Total Papers: 152
  • Time Span: 1972-2024 (52 years)
  • Chapters: 13
  • Distribution:
  • Classic Foundations: 13 papers
  • Deep Learning Era: 39 papers
  • Transformer Architectures: 14 papers
  • Large Language Models: 9 papers
  • Training Techniques: 21 papers
  • Inference & Deployment: 9 papers
  • Retrieval & Knowledge: 7 papers
  • Multimodal & Vision: 13 papers
  • Reasoning & Interpretability: 6 papers
  • Safety & Alignment: 6 papers
  • Benchmarking & Evaluation: 5 papers
  • Prompting & In-Context Learning: 2 papers
  • Emerging Topics: 5 papers
  • Topics Covered: Foundations, Deep Learning, Transformers, LLMs, Training, Inference, Retrieval, Multimodal, Reasoning, Safety, Evaluation, Prompting, Emerging

Last updated: 2026-08-27

Ready to start? Choose a chapter above or use the goal-based navigation table!