Resources

3.1 - Limitations and Challenges with AI Language Models

Proposed Readings:

• Deng, L., & Yu, D. (Eds.). (2014). Deep learning: Methods and applications. Now Publishers Inc. (Chapter 6: Language Modeling and Text Generation) (Chapter 7: Speech Recognition and Synthesis)

• Brown, T. B., et al. (2020). Language models are few-shot learners. OpenAI Blog. Retrieved from https://openai.com/blog/language-models-are-few-shot-learners/

Proposed Media:

• OpenAI. (2021). AI Safety: A Look Into The Future [Video]. YouTube. Retrieved from https://youtu.be/v_f7WqMuoPk

3.2 - Examining Bias

Proposed Readings:

• Hutchinson, B. (2020). The AI language model that prefers "man" as a doctor. OpenAI Blog. Retrieved from https://openai.com/blog/gpt-3-5-turbo/

Proposed Media:

• AI Alignment Podcast. (2021). Debiasing Language Models [Video]. YouTube. Retrieved from https://youtu.be/edMxDn3DTwQ

3.3 - Challenges in Complex Queries

Proposed Readings:

• Jurafsky, D., & Martin, J. H. (2020). Speech and language processing (3rd ed.). Pearson. (Chapter 23: Semantics: Representing Word Meaning) (Chapter 27: Question Answering and Dialog Systems)

• OpenAI. (2021). How ChatGPT Works - The Magic of Language Models. Retrieved from https://openai.com/blog/how-chatgpt-works/

Proposed Media:

• OpenAI. (2021). GPT-3: More Than a Model [Video]. YouTube. Retrieved from https://youtu.be/5O0ptv4Lj98

3.4 - AI Safety

Proposed Readings:

• Bostrom, N. (2014). Superintelligence: Paths, dangers, strategies. Oxford University Press. (Chapter 8: The strategic picture) (Chapter 9: Multipolar scenarios)

Proposed Media:

• Future of Life Institute. (2018). OpenAI Scholars - Safety Team Chat [Video]. YouTube. Retrieved from https://youtu.be/La9r8u1mGhI