BERT: Pre-training of Deep Bidirectional Transformers (2018)
Showed how pre-training transfers across nearly every language task.
- Category
- Artificial Intelligence
- Section
- Renowned AI Publications & Articles
- Tag
- Paper
Why it's in the directory
BERT: Pre-training of Deep Bidirectional Transformers (2018) sits in Renowned AI Publications & Articles — The papers that shaped modern artificial intelligence — every link opens the paper itself. It earns a place here because it is a primary, actively maintained source rather than a rehost of someone else's data, and because it stays useful whether you are an enthusiast, a student or working the line.
Related links
All Artificial IntelligenceAttention Is All You Need (2017)
Introduced the Transformer, the architecture behind today's language models.
ImageNet Classification with Deep CNNs (AlexNet, 2012)
The result that started the deep-learning era in computer vision.
Deep Residual Learning (ResNet, 2015)
Made very deep networks trainable; still the backbone of many vision systems.
Generative Adversarial Networks (2014)
The original GAN paper that opened up modern image generation.
Language Models are Few-Shot Learners (GPT-3, 2020)
Demonstrated that scale alone unlocks learning from a handful of examples.
Chain-of-Thought Prompting (2022)
Showed that asking a model to reason step by step sharply improves accuracy.