Video descriptions of research papers relating to foundation models and scaling by @samuelalbanie.
1. On the Opportunities and Risks of Foundation Models
2. GPT-3: Language Models are Few-shot Learners
3. CLIP: Contrastive Language-Image Pre-training
4. DINO: Emerging properties in self-supervised vision transformers
5. Flamingo: a Visual Language Model for Few-Shot Learning
6. Codex: Evaluating Large Language Models Trained on Code
7. Flan-PaLM: Scaling Instruction-Finetuned Language Models
8. BLOOMZ & mT0: Crosslingual Generalization through Multitask Finetuning
9. BLOOM: A 176B-parameter Open-access Multilingual Language Model
10. GPT-4
Notes about content: As noted in the first lecture above, the topic, definition or indeed the name "Foundation Models" is not yet agreed upon in the research community (and may change). The collection of videos is a work in progress.