Transparent value alignment

L Sanneman, J Shah - Companion of the 2023 ACM/IEEE International …, 2023 - dl.acm.org
As robots become increasingly prevalent in our communities, aligning the values motivating
their behavior with human values is critical. However, it is often difficult or impossible for …

An Information Bottleneck Characterization of the Understanding-Workload Tradeoff in Human-Centered Explainable AI

L Sanneman, M Tucker, JA Shah - The 2024 ACM Conference on …, 2024 - dl.acm.org
Recent advances in artificial intelligence (AI) have underscored the need for explainable AI
(XAI) to support human understanding of AI systems. Consideration of human factors that …

An information bottleneck characterization of the understanding-workload tradeoff

L Sanneman, M Tucker, J Shah - arXiv preprint arXiv:2310.07802, 2023 - arxiv.org
Recent advances in artificial intelligence (AI) have underscored the need for explainable AI
(XAI) to support human understanding of AI systems. Consideration of human factors that …

[PDF][PDF] Closed-loop reasoning about counterfactuals to improve policy transparency

MS Lee, H Admoni, R Simmons - International Conference o …, 2023 - harplab.github.io
Explanations are a powerful way of increasing the transparency of complex AI policies. Such
explanations must not only be informative regarding the policy in question, but must also be …

Making AI Policies Transparent to Humans through Demonstrations

MS Lee - Proceedings of the AAAI Conference on Artificial …, 2024 - ojs.aaai.org
Demonstrations are a powerful way of increasing the transparency of AI policies to humans.
Though we can approximately model human learning from demonstrations as inverse …

Closed-loop Teaching via Demonstrations to Improve Policy Transparency

MS Lee, R Simmons, H Admoni - arXiv preprint arXiv:2406.11850, 2024 - arxiv.org
Demonstrations are a powerful way of increasing the transparency of AI policies. Though
informative demonstrations may be selected a priori through the machine teaching …

Understanding Robot Minds: Leveraging Machine Teaching for Transparent Human-Robot Collaboration Across Diverse Groups

SK Jayaraman, R Simmons, A Steinfeld… - arXiv preprint arXiv …, 2024 - arxiv.org
In this work, we aim to improve transparency and efficacy in human-robot collaboration by
developing machine teaching algorithms suitable for groups with varied learning …

[PDF][PDF] Adaptive group machine teaching for human group inverse reinforcement learning

SK Jayaraman, A Steinfeld, H Admoni, R Simmons - 2023 - researchgate.net
For safe and effective collaboration between a robot and a human group, the challenge
arises in teaching a diverse group of individuals about the robot's decision-making process …

Transparent Value Alignment: Foundations for Human-Centered Explainable AI in Alignment

L Sanneman - 2023 - dspace.mit.edu
Alignment of autonomous agents' values and objectives with those of humans can greatly
enhance these agents' ability to act flexibly to safely and reliably meet humans' goals across …

[PDF][PDF] Improving the Transparency of Agent Decision Making to Humans Using Demonstrations

MS Lee - 2024 - ri.cmu.edu
For intelligent agents (eg robots) to be seamlessly integrated into human society, humans
must be able to understand their decision making. For example, the decision making of …