A Roadmap for Robust End-to-End Alignment
arXiv:1809.01036
Abstract
This paper discussed the {\it robust alignment} problem, that is, the problem of aligning the goals of algorithms with human preferences. It presented a general roadmap to tackle this issue. Interestingly, this roadmap identifies 5 critical steps, as well as many relevant aspects of these 5 steps. In other words, we have presented a large number of hopefully more tractable subproblems that readers are highly encouraged to tackle. Hopefully, this combination allows to better highlight the most pressing problems, how every expertise can be best used to, and how combining the solutions to subproblems might add up to solve robust alignment.
arXiv comment21 pages, 2 figures
References in corpus (14)
- Efficient Estimation of Word Representations in Vector Space
- Neural Architecture Search with Reinforcement Learning
- Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shift
- The Hidden Vulnerability of Distributed Learning in Byzantium
- Recurrent World Models Facilitate Policy Evolution
- When Will AI Exceed Human Performance? Evidence from AI Experts
- AI Safety Gridworlds
- A Game-Based Approximate Verification of Deep Neural Networks with Provable Guarantees
- Recent Advances in Algorithmic High-Dimensional Robust Statistics
- Supervising strong learners by amplifying weak experts
- Asynchronous Byzantine Machine Learning (the case of SGD)
- A Voting-Based System for Ethical Decision Making
- Dynamic Safe Interruptibility for Decentralized Multi-Agent Reinforcement Learning
- Scalable Byzantine Reliable Broadcast (Extended Version)