Controllable Neural Text Generation - Lilian Weng

cnlg.md

Decoding Strategies

Common Decoding Methods

Greedy search

Beam search

Top-k sampling https://arxiv.org/abs/1805.04833

Nucleus sampling https://arxiv.org/abs/1904.09751

Penalized sampling https://arxiv.org/abs/1909.05858

Guided Decoding

https://www.aclweb.org/anthology/P17-4008/

https://arxiv.org/abs/1805.06087

https://arxiv.org/abs/1701.06549

https://arxiv.org/abs/1809.01215

https://arxiv.org/abs/1805.06087

https://arxiv.org/abs/2002.10375

https://arxiv.org/abs/2010.02650

Trainable Decoding

https://arxiv.org/abs/1702.02429

https://arxiv.org/abs/1605.03835

https://arxiv.org/abs/1906.09531

https://arxiv.org/abs/2004.11714

https://arxiv.org/abs/1906.03351

https://www.kdnuggets.com/2019/07/introduction-noise-contrastive-estimation.html

Smart Prompt Design

https://arxiv.org/abs/2010.15980

https://arxiv.org/abs/1911.12543

Gradient-based Search

http://ucinlp.github.io/autoprompt

https://arxiv.org/abs/1908.07125

https://arxiv.org/abs/2101.00190

https://github.com/THUDM/P-tuning

https://arxiv.org/abs/2104.08691

Heuristic-based Search

https://www.aclweb.org/anthology/P18-1079/

https://arxiv.org/abs/1911.12543

Fine-tuning

Conditional Training

https://arxiv.org/abs/1805.04833

https://www.aclweb.org/anthology/W18-1505/

https://github.com/salesforce/ctrl

RL Fine-tuning

https://arxiv.org/abs/1511.06732

https://arxiv.org/abs/1609.08144

https://arxiv.org/abs/1707.07402

https://arxiv.org/abs/1705.04304

https://arxiv.org/abs/1804.07036

https://arxiv.org/abs/1809.10736

https://arxiv.org/abs/1609.08144

https://arxiv.org/abs/1705.04304

RL Fine-tuning with Human Preferences

https://arxiv.org/abs/1706.03741

https://arxiv.org/abs/1904.13015

https://arxiv.org/abs/1909.01214

https://arxiv.org/abs/1909.08593

https://arxiv.org/abs/2009.01325

Guided Fine-tuning with Steerable Layer

https://arxiv.org/abs/1612.00005

https://arxiv.org/abs/1912.02164

https://arxiv.org/abs/1912.13503

https://arxiv.org/abs/1709.07871

https://arxiv.org/abs/2006.16823

https://arxiv.org/abs/2009.06367

Distributional Approach

https://arxiv.org/abs/2012.11635

Unlikelihood Training

https://arxiv.org/abs/1908.04319