Robots That Know What to Ask: Recovering Misaligned Rewards through Targeted Explanations
Robots learning reward functions from demonstrations often suffer from underspecified features due to imperfect demonstrations. This paper proposes a framework that detects underspecified features by analyzing variation across demonstrations (low variation indicates well-specified, high variation indicates underspecified). The robot then explains its uncertainty in natural language and requests targeted corrective demonstrations. Evaluations in simulation and with a real Franka robot show that explanation-guided queries significantly improve reward recovery over random querying and passive data collection.
Article intelligence
Key points
- Imperfect demonstrations can lead to underspecified features and misaligned robot behavior at deployment.
- A method detects underspecified features by measuring variability across demonstrations.
- The robot explains which features it is uncertain about and asks for targeted demonstrations.
- Experiments show explanation-guided queries outperform random querying and passive data collection.
Why it matters
This matters because imperfect demonstrations can lead to underspecified features and misaligned robot behavior at deployment.
Technical impact
May affect research directions, evaluation methods, open-source reproduction, and productization paths.
[2605.22986] Robots That Know What to Ask: Recovering Misaligned Rewards through Targeted Explanations
[Submitted on 21 May 2026]
Title:Robots That Know What to Ask: Recovering Misaligned Rewards through Targeted Explanations
View a PDF of the paper titled Robots That Know What to Ask: Recovering Misaligned Rewards through Targeted Explanations, by Helena Merker and 2 other authors
View PDF HTML (experimental)
Abstract:Learning reward functions from demonstrations assumes that demonstrations provide adequate supervision over all features -- or task-relevant aspects of behavior. In practice, demonstrations are often imperfect: humans may under-emphasize certain features due to cognitive load or physical difficulty, or the training regime may fail to sufficiently cover all relevant situations. In either case, important features may be underspecified, leading to ambiguity in the learned reward function and misaligned behavior at deployment. We propose a framework that detects such underspecified features and actively solicits targeted corrective demonstrations. Our key insight is that demonstrations implicitly reveal which features are well specified: features that are consistently optimized show little variation across demonstrations, while features that are underspecified vary widely. We leverage this statistical signal to infer which features may have been insufficiently demonstrated. The robot then explains which features it is uncertain about in natural language and queries for demonstrations that explicitly address the identified gaps. We evaluate our approach in a simulated tabletop manipulation domain and in a user study with a real Franka robot. Targeted, explanation-guided queries significantly improve reward recovery compared to random querying and passive data collection, reducing ambiguity that would otherwise persist in learning from imperfect demonstrations.
Subjects:
Robotics (cs.RO); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC); Machine Learning (cs.LG)
Cite as: arXiv:2605.22986 [cs.RO]
(or arXiv:2605.22986v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2605.22986
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Nick Walker [view email] [v1] Thu, 21 May 2026 19:34:14 UTC (3,599 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Robots That Know What to Ask: Recovering Misaligned Rewards through Targeted Explanations, by Helena Merker and 2 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.RO
new | recent | 2026-05
Change to browse by:
cs cs.AI cs.HC cs.LG
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
Loading...
Data provided by:
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)