Interconnects

Stop "reinventing" everything to "solve" alignment

Apr 17, 2024

Delve into integrating non-computing science into reinforcement learning for AI alignment. Explore social choice theory for diverse human feedback. Discover OLMo 1.7 7B model with good benchmarks and open design. Unveil insights on pluralistic alignment in AI systems for inclusivity.

07:32

Creator website

AI Summary

AI Chapters

Episode notes

Podcast summary created with Snipd AI

Quick takeaways

Integrating non-computing science elements in reinforcement learning broadens perspectives and enhances model accuracy.

Applying social choice theory to AI alignment improves transparency, minimizes biases, and fosters pluralistic model development.

Deep dives

Integration of non-computing science into reinforcement learning for human feedback

Integrating non-computing science elements into reinforcement learning from human feedback can help in achieving desired models. This inclusive approach, particularly in solving alignment issues, ensures a broader perspective beyond the Computer Science domain. By leveraging existing solutions from fields like economics and social sciences, the complexity of addressing human feedback in reinforcement learning can be effectively managed. The relevance of incorporating diverse opinions and methodologies from various disciplines is emphasized to enhance the transparency and efficacy of AI models.

Introduction

2min

Exploring the Integration of Science Fields for AI Alignment

2min

Exploring Various Aspects of Reinforcement Learning from Human Feedback and Pluralistic Alignment in AI Systems

3min

Integrating some non computing science into reinforcement learning from human feedback can give us the models we want.
This is AI generated audio with Python and 11Labs.
Source code: https://github.com/natolambert/interconnects-tools
Original post: https://www.interconnects.ai/p/reinventing-llm-alignment

0:00 Stop "reinventing" everything to "solve" AI alignment
2:19 Social Choice for AI Alignment: Dealing with Diverse Human Feedback
7:03 OLMo 1.7 7B: A truly open model with actually good benchmarks

Fig 1: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_013.png
Fig 2: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_015.png
Fig 3: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_018.png
Fig 4: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_024.png
Fig 5: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_027.png

Get full access to Interconnects at www.interconnects.ai/subscribe