See also Sparsity in Neural Networks.

• General
  • Cheng et al 2017 - A Survey of Model Compression and Acceleration for Deep Neural Networks
  • Hoefler et al 2021 - Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks
  • 2023 - Model Compression for Deep Neural Networks: A Survey
  • See chapter 4 of 2023 - Efficient Neural Networks for Tiny Machine Learning: A Comprehensive Review
• LLMs and Transformers
  • Xu & McAuley et al 2022 - A Survey on Model Compression and Acceleration for Pretrained Language Models
  • Zhu et al 2023 - A Survey on Model Compression for Large Language Models
  • Tang et al 2024 - A Survey on Transformer Compression
• Pruning
  • Blalock et al 2020 - What is the State of Neural Network Pruning?
  • Hoefler et al 2021 - Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks
• Distillation
  • Xu et al 2024 - A Survey on Knowledge Distillation of Large Language Models

• Li et al 2020 - Train Large, Then Compress: Rethinking Model Size for Efficient Training and Inference of Transformers Related: Scaling laws
• Ganesh et al 2020 - Compressing Large-Scale Transformer-Based Models: A Case Study on BERT

See also Sparsity.

• LeCun et al 1990 - Optimal Brain Damage
• Giles & Omlin 1994 - Pruning Recurrent Neural Networks for Improved Generalization Performance Old 1990s paper. Prunes nodes by magnitude of total incoming weights, rather than pruning edges like more recent work does.
• See et al 2016 - Compression of Neural Machine Translation Models via Pruning
• Frankle & Carbin 2019 - The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks After training, only a subset of the edges in the network matter. The rest can be pruned (take only the edges with the highest weights) before training if you keep the same initialization weights.
• LayerDrop: Fan et al 2019 - Reducing Transformer Depth on Demand with Structured Dropout Regularize networks to allow extraction of smaller networks of any depth at test time without needing to finetune. As easy to implement as dropout.
• Sanh et al 2020 - Movement Pruning: Adaptive Sparsity by Fine-Tuning Works for transfer-learning of pre-trained models
• Lagunas et al 2021 - Block Pruning For Faster Transformers
• LLM Pruning
  • See overview in section 2.1.2 of Wan 2024 or section 4 of Zhu 2023.
  • Frantar & Alistarh 2023 - SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot
  • Ma et al 2023 - LLM-Pruner: On the Structural Pruning of Large Language Models github
  • Sun et al 2023 - A Simple and Effective Pruning Approach for Large Language Models
  • SIMPLE: Tao et al 2023 - Structured Pruning for Efficient Generative Pre-trained Language Models
  • Xia et al 2023 - Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning
  • Kim et al 2024 - Shortened LLaMA: Depth Pruning for Large Language Models with Comparison of Retraining Methods
  • Men et al 2024 - ShortGPT: Layers in Large Language Models are More Redundant Than You Expect

• Shen et al 2019 - Q-BERT: Hessian Based Ultra Low Precision Quantization of BERT
• Aji & Heafield 2020 - Compressing Neural Machine Translation Models with 4-bit Precision
• Kim et al 2020 - I-BERT: Integer-only BERT Quantization
• Empirical Studies
  • Zheng et al 2025 - An Empirical Study of Qwen3 Quantization
  • Liu et al 2024 - Do Emergent Abilities Exist in Quantized Large Language Models: An Empirical Study
  • Liu et al 2025 - Quantization Hurts Reasoning? An Empirical Study on Quantized Reasoning Models

• Zhou et al 2016 - DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
• Hubara et al 2018 - Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
• Hubara et al 2018 - Quantized Back-Propagation: Training Binarized Neural Networks with Quantized Gradients Interesting, but was rejected, see here.
• 2019 - Q8BERT: Quantized 8Bit BERT Software
• Chmiel et al 2021 2021 - Logarithmic Unbiased Quantization: Simple 4-bit Training in Deep Learning
• Sun et al 2025 - Scaling Laws for Floating Point Quantization Training

• Courbariaux et al 2016 - Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1
• Rastegari et al 2016 - XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks
• Anderson & Berg 2017 - The High-Dimensional Geometry of Binary Neural Networks
• Liu et al 2018 - Bi-Real Net: Enhancing the Performance of 1-bit CNNs With Improved Representational Capability and Advanced Training Algorithm Binarizes the weights and activations of the CNN, but includes real residual connections between layers to increase expressivity of the networks
• Li & Su 2020 - Accelerating Binarized Neural Networks via Bit-Tensor-Cores in Turing GPUs
• Kurtz & Bah 2020 - An Integer Programming Approach to Deep Neural Networks with Binary Activation Functions
• Xu et al 2021 - Learning Frequency Domain Approximation for Binary Neural Networks

See also Knowledge Distillation.

• Sanh et al 2019 - DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter
• Zhao et al 2021 - Extremely Small BERT Models from Mixed-Vocabulary Training Most of BERT's memory footprint comes from the input embeddings. This method uses model distillation to compress BERT large by an order of magnitude better than distilled models.

• HashedNets: Chen et al 2015 - Compressing Neural Networks with the Hashing Trick Randomly shares weights in the neural network by using hashing

• MLSys

• Daniel Soudry - Quantization and binarized NNs

• Conditional Computation Invokes certain parts of the network for each instance, early exit, etc
• Edge Computing
• Efficient NNs
• Knowledge Distillation
• Sparsity in Neural Networks