Pith. sign in

REVIEW 4 major objections 6 minor 80 references

Joint Decision-Making in Robot Teleoperation: When are Two Heads Better Than One?

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Confidence sharing lets two operators beat the better one alone, even in active robot control.

desk verdict Promising MCS extension to teleoperation, but virtual dyads don't match on scene and the statistics need re-doing before the headline claim holds. read the letter →

arxiv 2503.15510 v1 pith:KL3CQJUG submitted 2025-01-28 cs.HC cs.RO

Duc-An Nguyen , Raunak Bhattacharyya , Clara Colombatto , Steve Fleming , Ingmar Posner , Nick Hawes This is my paper · ORCID
classification cs.HCcs.RO
keywords jointdecision-makingmaximumconfidenceslatingteleoperationhuman-robotinteractioncalibrationvirtualdyadsmetacognitiontwoheadsbetterthanone
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tests whether two human operators can make better decisions together than either can alone in a dynamic task: choosing which of two simulated mobile robots to operate, where the robots differ in hidden control delay. The authors ran an online experiment with 100 participants (80 retained after quality screening), each driving both robots for a short time and then selecting the lower-delay robot while rating confidence on a four-point scale. They formed virtual dyads by pairing participants offline and applied maximum confidence slating: on trials where paired operators disagreed, the joint decision was the choice of whoever reported higher confidence. The central finding is that these confidence-weighted joint decisions were more accurate than the choices of the better-performing individual in the dyad ($p<0.0001$), even though the task was active and time-pressured rather than passive and static. The paper also finds that the benefit shrinks as skill gaps widen, depends on how well calibrated each operator's confidence is, and is predicted by the dyad's overall confidence calibration.

What carries the argument

The load-bearing mechanism is maximum confidence slating (MCS), a decision rule that resolves disagreement between two operators by taking the choice of whichever operator reports higher confidence; the paper uses confidence ratings on a four-point Likert scale collected immediately after each robot-selection decision. MCS carries the argument because it converts each individual's metacognitive self-monitoring into a group decision without requiring discussion, and the virtual-dyad design lets the paper test the rule on thousands of pairings assembled from individual sessions. The second piece of machinery is the AUROC2 calibration measure, a type-2 ROC score that reflects how well an individual's (or a dyad's) confidence ratings separate correct from incorrect trials; the paper uses it both to classify operators as well-calibrated or poorly calibrated and to show that dyad-level calibration predicts the size of the MCS benefit.

What would settle it

Run the same two-robot selection task with operators who actually work as a pair in real time—able to talk or at least see each other's confidence before deciding—and compare their joint accuracy with the accuracy of the offline maximum-confidence rule on the same trials; if the interactive dyads fail to beat the better individual, or fall below offline MCS accuracy, the virtual-dyad premise is the part that gives way.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the "two heads are better than one" effect, previously shown in static perceptual and knowledge tasks, extends to spatiotemporal tasks with active operator control of robots. In a simulated teleoperation task where each participant drove two robots with different hidden control delays and chose which robot was more responsive, virtual dyads whose joint decisions followed maximum confidence slating (MCS) — selecting the higher-confidence member's choice whenever the two disagreed — achieved significantly higher accuracy than the highest-performing member of the pair alone, with a comparison across 3,160 dyads giving $p<0.0001$. Low-confidence choices were no better than random, larger differences in individual success rates reduced the accuracy gain (and the gain became negative beyond roughly an 8% gap), and confidence calibration regulated the benefit: above-average calibrators benefited stably, while below-average calibrators benefited when paired with differently calibrated partners. The paper further reports that dyadic confidence calibration, computed on the pooled confidence–correctness pairs of both members, correlated positively with accuracy gain.

Load-bearing premise

The load-bearing premise is that a virtual dyad—two people whose sessions were collected separately and paired afterwards, with no interaction—is a valid model of real two-person joint decision-making in this task; the paper itself notes that in real teams, social effects such as perceived competence and status can change how confidence is expressed and weighted.

Editorial extensions

If this is right

  • If the effect holds in practice, a two-operator teleoperation console could improve decision accuracy without extra training: when the operators disagree, follow whichever one is more confident.
  • The performance-discrepancy result implies that teams assembled for confidence pooling should be matched in skill; pairing a much stronger operator with a much weaker one can make the joint decision worse than the stronger operator alone.
  • Confidence calibration can act as a screening tool: operators whose confidence ratings track their accuracy well are safe to pool, while poorly calibrated operators benefit most when paired with someone whose calibration differs from theirs.
  • Low-confidence decisions are so unreliable that a joint teleoperation protocol should treat them as abstentions rather than votes.
  • Because the benefit is retained even without communication between the pair, a joint-decision aid could combine two operators' confidence ratings asynchronously, as long as both faced the same evidence.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next step the paper gestures toward is human-AI dyads; if MCS transfers, an AI assistant could ask a human operator for confidence and defer to the higher-confidence agent, making confidence a cheap coordination currency between human and machine.
  • Because the offline pairing removes real-time social pressure, the results bound what two people can gain purely from self-monitored confidence; live interaction could shrink or enlarge the gain depending on how status and perceived competence distort confidence reports.
  • A direct test of the virtual-dyad premise would run the same navigation task with physically paired operators who can talk, and compare their joint accuracy with the offline MCS prediction; a mismatch would localize where social interaction changes confidence weighting.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper reports a human-subject experiment (N=100, 80 after exclusions) in which participants teleoperate one of two simulated robots under different control delays and then choose which robot had the lower delay, rating their confidence on a 1–4 scale. The authors form virtual dyads by pairing participants' trials that share the same delay-pair condition, create a synthetic Maximum Confidence Slating (D-HC) agent that selects the more confident participant's choice on disagreement trials, and compare its accuracy to the better-performing individual (HP), a low-confidence agent, and a random agent. They report that D-HC significantly outperforms HP (t(3158)=8.36, p<0.0001), and they analyze how accuracy gains vary with skill difference and confidence calibration, alongside dyadic calibration estimates. The paper claims this is the first demonstration that 'two heads are better than one' in an active, spatiotemporal teleoperation task.

Significance. If the central result holds, the paper would make a useful contribution to human-robot interaction and joint decision-making by extending the Maximum Confidence Slating literature from static, passive tasks to active teleoperation, and by characterizing the role of skill similarity and confidence calibration in that setting. The authors use established metrics (MCS from Koriat, AUROC2 from Fleming and Lau) rather than inventing new ones, and the experimental design with a staircase procedure and a four-point confidence scale is reasonable for individual-level measurement. The headline effect is large and directionally plausible. However, the internal validity of the virtual-dyad construction is the load-bearing issue: as described, paired trials share only the delay condition, not the full spatiotemporal scene, which could create a scene-difficulty confound. The statistical analysis also treats overlapping dyads as independent, and at least one reported test statistic is internally inconsistent. These issues must be resolved before the central claim is established.

major comments (4)
  1. [III-B and IV] The virtual dyads are paired only on the delay-pair condition, not on the 24 scene conditions (6 initial robot poses × 4 doorway configurations, as defined in III-A). Each trial's scene is sampled randomly for each participant, so two paired participants may judge different spatiotemporal events. When D-HC selects the more confident participant's choice on a disagreement, confidence is then partly a function of the scene difficulty each participant happened to draw; a participant who received an easier scene is more likely to be both correct and confident. The claim that MCS beats the better individual on a spatiotemporal task requires both operators to perceive the same event. As reported, the virtual dyads never jointly observe a common trial. Please re-run the analysis matching trials on the full condition (pose and doorway) or demonstrate that scene matching is unnecessary; the Discussion's social-effect limitation does not address this item-level confound.
  2. [IV-A and all t-tests] All significance tests treat the 3,160 virtual dyads as independent observations (e.g., t(3158)=8.36 in Figure 3, t(1008)=7.1, t(694)=8.35 in IV-B). Each participant appears in 79 dyads, so the dyads are not independent; this clustering inflates the effective sample size and can produce artificially small p-values. Please re-analyze with a mixed-effects model that includes participant (or dyad-member) random effects, or with cluster-robust standard errors or a permutation test that resamples participants. Report effect sizes and confidence intervals, not only t statistics.
  3. [IV-C, Figure 8a] The text reports 'The performance of D-HC was significantly better than HP, (t(574) = 1.39, p < 0.0001).' This is internally inconsistent: with df=574, t=1.39 is not significant at any conventional level. This statistic is load-bearing for H3, which claims that well-calibrated pairs benefit from MCS. Please correct the value or the p-value, and if the corrected result is not significant, revise the claim accordingly.
  4. [IV (D-HC/D-LC construction)] The definition of D-HC and D-LC is incomplete for disagreement trials in which both participants give the same confidence rating. The text says 'the response of the participant who indicated higher confidence was selected,' but for ties there is no specified rule. Please specify whether ties are excluded, broken randomly, or assigned by some other rule, and report how many of the 1,932 disagreement trials involve ties. Without this, the main accuracy calculation is not fully reproducible.
minor comments (6)
  1. [IV-B] Figure 4 reports a regression coefficient r = −0.48 without stating the correlation type (Pearson or Spearman) or a confidence interval; please clarify.
  2. [IV-B] In the text near Figure 5, the mean skill level is given as 'success percentage > 67%' and '< 67%'; please state the exact mean value and include an error bar or confidence intervals in the figure.
  3. [IV-C, Figure 8a] The caption for Figure 8a says 'both well calibrated (288 pairs)' while the text says the group had 288 virtual dyads; please ensure the numbering is consistent across the three subplots and with the stated participant counts (25 well-calibrated and 26 poorly-calibrated participants cannot produce both 288 and 242 dyads unless some pairs are excluded; explain the construction).
  4. [IV-D] The dyadic AUROC2 is computed from the concatenated trials of the two participants after excluding trials with identical robot choices. Since the two participants' trials are not matched by scene, the resulting dyadic calibration measure may also be affected by the scene-matching issue raised above; please discuss this in the revised analysis.
  5. [Global] Multiple t-tests are presented without correction for multiple comparisons; given the number of pairwise contrasts in Figures 3, 8, and IV-B, a multiple-comparison adjustment or an explicit note about exploratory analyses would improve robustness.
  6. [VI] The heading 'A CKNOWLEDGEMENT' has a typo; it should be 'ACKNOWLEDGMENTS' or 'ACKNOWLEDGEMENT'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central MCS result is an empirical comparison of a fixed confidence heuristic against individual performance, not a derivation from fitted quantities.

full rationale

The paper's derivation chain is self-contained in the sense required for circularity analysis. The Maximum Confidence Slating (MCS) rule is adopted from prior literature (Koriat; Bahrami et al.) and is applied as a fixed, untrained per-trial selection rule: D-HC simply chooses the response of the participant who reported higher confidence. No parameter is fitted to the accuracy outcome, and the comparison against HP (the higher-performing individual) is not guaranteed by construction; it is a statistical result obtained from 3,160 virtual dyads and 1,932 disagreement trials. The AUROC2 calibration metric comes from external methodological work (Fleming and Lau, 2014; Koriat, 2012) and is not used to define the target accuracy advantage. The paper's self-citations, such as reference [8] used to justify exclusion criteria and metrics in Section III-C ('consistent with standard procedures in human self-confidence research'), support methodological choices and the broader research program but are not load-bearing for the headline claim. The one substantive concern is a validity threat rather than circularity: virtual dyads are paired by delay-pair condition only ('Participant 1 encountered 20 trials with the delay-pair condition... Participant 2 encountered 15 trials with the same delay condition'), so the two individuals may be judging different scenes and confidence may partially encode scene difficulty. This is a legitimate internal-validity caveat and is only partially addressed by the authors' admission that 'in real joint decision-making between two people social effects... can play a big role.' It does not, however, reduce the reported D-HC advantage to an identity or to a fitted-input prediction. Therefore no circular step is exhibited.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

No new physical or cognitive entities are postulated; D-HC, D-LC, D-Random, and virtual dyads are analytic constructs derived from existing participant data. The main analytical choices are hand-set thresholds for exclusion and grouping, and the load-bearing premise that offline virtual pairing represents real joint decision-making.

free parameters (5)
  • Exclusion accuracy threshold = 65%
    Participants with below 65% accuracy were excluded; this hand-chosen cut affects the participant pool and therefore the dyad accuracy distributions.
  • Confidence invariance exclusion threshold = over 95 of 100 trials
    Participants who used the same confidence rating in over 95 trials were excluded; chosen by the authors to filter inattentive responders.
  • AUROC2 grouping mean threshold = 0.6
    Participants were split into above/below mean calibration at AUROC2 = 0.6; the split point is data-driven and affects RQ3 analyses.
  • Calibration extremes thresholds = 0.65 and 0.55
    Top participants with AUROC2 > 0.65 and bottom with < 0.55 were selected for sub-analyses; cutoffs are hand-chosen.
  • Mean skill split threshold = 67% success rate
    Participants were split into above/below mean skill at 67% success; used for Figure 5 grouping.
assumptions (4)
  • domain assumption Participants' confidence ratings are truthful, comparable across individuals, and reflect internal monitoring of performance.
    The MCS rule and AUROC2 calibration assume confidence is a meaningful, comparable signal; stated throughout (e.g., Section I, III-D).
  • domain assumption Virtual dyads formed post hoc from non-interacting participants behave like interacting dyads for decision aggregation.
    All joint decisions are computed offline from separately collected trials; the authors acknowledge this in the Discussion limitations, and it is load-bearing for the 'two heads' claim.
  • domain assumption Standard t-test independence assumptions hold across pooled virtual dyads and disagreement trials.
    Each participant appears in 79 dyads, so trials are not independent; the paper does not use mixed effects or cluster-robust statistics.
  • domain assumption AUROC2 computed on concatenated decision-confidence pairs is a valid measure of dyadic calibration.
    Koriat's method is applied to pooled pairs; no validation that pooling preserves individual metacognitive properties.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Joint Decision-Making in Robot Teleoperation: When are Two Heads Better Than One?." pith.science (2026). https://pith.science/paper/KL3CQJUG

@misc{pith2026250315510,
  author       = {Pith},
  title        = {Pith review of: Joint Decision-Making in Robot Teleoperation: When are Two Heads Better Than One?},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KL3CQJUG}},
  note         = {Machine review of arXiv:2503.15510}
}
abstract

Operators working with robots in safety-critical domains have to make decisions under uncertainty, which remains a challenging problem for a single human operator. An open question is whether two human operators can make better decisions jointly, as compared to a single operator alone. While prior work has shown that two heads are better than one, such studies have been mostly limited to static and passive tasks. We investigate joint decision-making in a dynamic task involving humans teleoperating robots. We conduct a human-subject experiment with $N=100$ participants where each participant performed a navigation task with two mobiles robots in simulation. We find that joint decision-making through confidence sharing improves dyad performance beyond the better-performing individual (p<0.0001). Further, we find that the extent of this benefit is regulated both by the skill level of each individual, as well as how well-calibrated their confidence estimates are. Finally, we present findings on characterising the human-human dyad's confidence calibration based on the individuals constituting the dyad. Our findings demonstrate for the first time that two heads are better than one, even on a spatiotemporal task which includes active operator control of robots.

Figures

Figures reproduced from arXiv: 2503.15510 by the authors.

Figure 1
Figure 1. Web based robot navigation simulator. The red and green [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Screenshots showing driving trials scenario and question queries (2a: environment setup; 2b: robot choices, 2c: user confidence levels) [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Accuracy in robot selection for high-confidence, higher [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (7 more)
Figure 5
Figure 5. Figure 5: Variation in accuracy gain as participants of different success [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Confidence calibration of two different participants. 6a: Plot of trial result vs. confidence rating. The line is the regression fit with 95% [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Variation in accuracy gain as participants of different perfor [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Choice accuracy (%) of different combinations of participants according to their confidence calibration: 8a: both well calibrated (288 [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: Variation in accuracy gain as participants of different confi [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 10
Figure 10. Figure 10: Trial results versus confidence ratings for example virtual [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]
Figure 11
Figure 11. Figure 11: Variation in dyadic confidence calibration based on how [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

80 extracted references · 74 canonical work pages

  1. [8]

    Towards intelligent decision support systems in robotics: Investigating the role of self-confidence calibration in joint decision-making

    Raunak Bhattacharyya, Duc An Nguyen, Clara Colom- batto, Stephen Fleming, Ingmar Posner, and Nick Hawes. Towards intelligent decision support systems in robotics: Investigating the role of self-confidence calibration in joint decision-making. In AAAI Spring Symposium Se- ries, 2024

    work page 2024

  2. [1]

    Towards metacog- nitive robot decision making for tool selection

    Ajith Anil Meera and Pablo Lanillos. Towards metacog- nitive robot decision making for tool selection. In International Workshop on Active Inference . Springer, 2023

    work page 2023

  3. [2]

    Inten- tional user adaptation to shared control assistance

    Reuben M Aronson and Elaine Schaertl Short. Inten- tional user adaptation to shared control assistance. In ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2024

    work page 2024

  4. [3]

    Opti- mally interacting minds

    Bahador Bahrami, Karsten Olsen, Peter E Latham, An- dreas Roepstorff, Geraint Rees, and Chris D Frith. Opti- mally interacting minds. Science, 329(5995):1081–1085, 2010

    work page 2010

  5. [4]

    Together, slowly but surely: The role of social interaction and feedback on the build-up of benefit in collective decision- making

    Bahador Bahrami, Karsten Olsen, Dan Bang, Andreas Roepstorff, Geraint Rees, and Chris Frith. Together, slowly but surely: The role of social interaction and feedback on the build-up of benefit in collective decision- making. Journal of Experimental Psychology: Human Perception and Performance, 38(1):3–8, February 2012. ISSN 1939-1277, 0096-1523. doi: 10.1...

    work page doi:10.1037/a0025708 2012

  6. [5]

    What failure in collective decision-making tells us about metacogni- tion

    Bahador Bahrami, Karsten Olsen, Dan Bang, Andreas Roepstorff, Geraint Rees, and Chris Frith. What failure in collective decision-making tells us about metacogni- tion. Philosophical Transactions of the Royal Society B: Biological Sciences , 367(1594):1350–1365, May 2012. ISSN 0962-8436, 1471-2970. doi: 10.1098/rstb.2011. 0420

    work page doi:10.1098/rstb.2011 2012

  7. [6]

    Latham, Jennifer Y .F

    Dan Bang, Riccardo Fusaroli, Kristian Tyl ´en, Karsten Olsen, Peter E. Latham, Jennifer Y .F. Lau, Andreas Roepstorff, Geraint Rees, Chris D. Frith, and Bahador Bahrami. Does interaction matter? testing whether a confidence heuristic can replace interaction in collective decision-making. Consciousness and Cognition , 26,

    work page

  8. [7]

    Dan Bang, Laurence Aitchison, Rani Moran, Santiago Herce Castanon, Banafsheh Rafiee, Ali Mahmoodi, Jen- nifer Y . F. Lau, Peter E. Latham, Bahador Bahrami, and Christopher Summerfield. Confidence matching in group decision-making. Nature Human Behaviour , 1(6):0117,

    work page

Show all 80 references
  1. [9]

    Target detection in video feeds with selected dyads and groups assisted by col- laborative brain-computer interfaces

    Saugat Bhattacharyya, Davide Valeriani, Caterina Cinel, Luca Citi, and Riccardo Poli. Target detection in video feeds with selected dyads and groups assisted by col- laborative brain-computer interfaces. In 2019 9th Inter- national IEEE/EMBS Conference on Neural Engineering (N...

    2019

  2. [10]

    Bayesian reinforcement learning for single- episode missions in partially unknown environments

    Matthew Budd, Paul Duckworth, Nick Hawes, and Bruno Lacerda. Bayesian reinforcement learning for single- episode missions in partially unknown environments. In Conference on Robot Learning (CoRL) , 2023

    2023

  3. [11]

    Toward developing hri metrics for teams: Pilot testing in the field

    Jennifer Burke, Matt Lineberry, Kevin S Pratt, Meng Taing, Robin Murphy, and Brian Day. Toward developing hri metrics for teams: Pilot testing in the field. Metrics for Human-Robot Interaction , 21, 2008

    2008

  4. [12]

    Yifan Cai, Abhinav Dahiya, Nils Wilde, and Stephen L. Smith. Scheduling operator assistance for shared auton- omy in multi-robot teams. (arXiv:2209.03458), 2022. URL http://arxiv.org/abs/2209.03458

    2022 arXiv

  5. [13]

    Jennifer Casper and Robin R. Murphy. Human-robot interactions during the robot-assisted urban search and rescue response at the world trade center. IEEE Trans- actions on Systems, Man, and Cybernetics, Part B (Cy- bernetics), 33(3):367–385, 2003

    2003

  6. [14]

    Supervisory con- trol of multiple robots: Effects of imperfect automation and individual differences

    Jessie YC Chen and Michael J Barnes. Supervisory con- trol of multiple robots: Effects of imperfect automation and individual differences. Human Factors, 54(2):157– 174, 2012

    2012

  7. [15]

    Experimental analysis of a variable autonomy framework for controlling a remotely operating mobile robot

    Manolis Chiou, Rustam Stolkin, Goda Bieksaite, Nick Hawes, Kimron L Shapiro, and Timothy S Harrison. Experimental analysis of a variable autonomy framework for controlling a remotely operating mobile robot. In IEEE/RSJ international conference on intelligent robots and systems...

    2016

  8. [16]

    Mixed-initiative variable autonomy for remotely operated mobile robots

    Manolis Chiou, Nick Hawes, and Rustam Stolkin. Mixed-initiative variable autonomy for remotely operated mobile robots. ACM Transactions on Human-Robot Interaction (THRI), 10(4):1–34, 2021

    2021

  9. [17]

    Robot-assisted nuclear disas- ter response: Report and insights from a field exercise

    Manolis Chiou, Georgios-Theofanis Epsimos, Grigoris Nikolaou, Pantelis Pappas, Giannis Petousakis, Stefan M¨uhl, and Rustam Stolkin. Robot-assisted nuclear disas- ter response: Report and insights from a field exercise. In IEEE/RSJ International Conference on Intelligent Robot...

    2022

  10. [18]

    Decision-making model for adaptive impedance control of teleoperation systems

    Javier Corredor, Jorge Sofrony, and Angelika Peer. Decision-making model for adaptive impedance control of teleoperation systems. IEEE Transactions on Haptics, 10(1):5–16, 2016

    2016

  11. [19]

    Abhinav Dahiya, Nima Akbarzadeh, Aditya Mahajan, and Stephen L. Smith. Scalable operator allocation for multi-robot assistance: A restless bandit approach. (arXiv:2111.06437), 2021. URL http://arxiv.org/abs/ 2111.06437

    2021 arXiv

  12. [20]

    Confidence in value-based choice

    Benedetto De Martino, Stephen M Fleming, Neil Garrett, and Raymond J Dolan. Confidence in value-based choice. Nature Neuroscience, 16(1):105–110, 2013. doi: 10.1038/nn.3279

    2013 doi

  13. [21]

    Robots need humans in the loop to improve the hopefulness of disaster survivors

    Lorin D Dole, David M Sirkin, Robin R Murphy, and Clifford I Nass. Robots need humans in the loop to improve the hopefulness of disaster survivors. In IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN) , 2015

    2015

  14. [22]

    A decomposition of uav-related situation awareness

    Jill L Drury, Laurel Riek, and Nathan Rackliffe. A decomposition of uav-related situation awareness. In ACM SIGCHI/SIGART Conference on Human-Robot In- teraction (HRI), 2006

    2006

  15. [23]

    Practi- cal, ethical, and overlooked: Teleoperated socially assis- tive robots in the quest for autonomy

    Saad Elbeleidy, Terran Mott, and Tom Williams. Practi- cal, ethical, and overlooked: Teleoperated socially assis- tive robots in the quest for autonomy. In ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 2022

    2022

  16. [24]

    Fleming and Nathaniel D

    Stephen M. Fleming and Nathaniel D. Daw. Self- evaluation of decision-making: A general bayesian framework for metacognitive computation. Psychological Review, 124(1):91–114, 2017

    2017

  17. [25]

    How to measure metacognition

    Stephen M Fleming and Hakwan C Lau. How to measure metacognition. Frontiers in Human Neuroscience, 8:443, 2014

    2014

  18. [26]

    Collaborative teleoperation via the internet

    Ken Goldberg, Billy Chen, Rory Solomon, Steve Bui, Bobak Farzin, Jacob Heitler, Derek Poon, and Gordon Smith. Collaborative teleoperation via the internet. In IEEE International Conference on Robotics and Automa- tion (ICRA), volume 2. IEEE, 2000

    2000

  19. [27]

    Collaborative online teleoperation with spatial dynamic voting and a human” tele-actor”

    Ken Goldberg, Dezhen Song, Y Khor, David Pescovitz, Anthony Levandowski, J Himmelstein, Janice Shih, An- namarie Ho, Eric Paulos, and J Donath. Collaborative online teleoperation with spatial dynamic voting and a human” tele-actor”. In IEEE International Conference on Robotics...

    2002

  20. [28]

    Probabilistic social learning improves the public’s judg- ments of news veracity

    Douglas Guilbeault, Samuel Woolley, and Joshua Becker. Probabilistic social learning improves the public’s judg- ments of news veracity. Plos One, 16(3):e0247487, 2021

    2021

  21. [29]

    Wisdom of crowds and collective decision-making in a survival situation with complex information integration

    Daisuke Hamada, Masataka Nakayama, and Jun Saiki. Wisdom of crowds and collective decision-making in a survival situation with complex information integration. Cognitive Research: Principles and Implications , 5(1): 1–15, 2020

    2020

  22. [30]

    Crossing re- ality: Comparing physical and virtual robot deixis

    Zhao Han, Yifei Zhu, Albert Phan, Fernando Sandoval Garza, Amia Castro, and Tom Williams. Crossing re- ality: Comparing physical and virtual robot deixis. In ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2023

    2023

  23. [31]

    Three heads are better than two: Comparing learning properties and performances across individuals, dyads, and triads through a computational approach

    Tsutomu Harada. Three heads are better than two: Comparing learning properties and performances across individuals, dyads, and triads through a computational approach. PLOS ONE , 16, 2021. doi: 10.1371/journal. pone.0252122

    2021 doi

  24. [32]

    The strands project: Long-term autonomy in everyday environments

    Nick Hawes, Christopher Burbridge, Ferdian Jovan, Lars Kunze, Bruno Lacerda, Lenka Mudrova, Jay Young, Jeremy Wyatt, Denise Hebesberger, Tobias Kortner, et al. The strands project: Long-term autonomy in everyday environments. IEEE Robotics & Automation Magazine , 24(3):146–156, 2017

    2017

  25. [33]

    Improving collocated robot teleoperation with augmented reality

    Hooman Hedayati, Michael Walker, and Daniel Szafir. Improving collocated robot teleoperation with augmented reality. In ACM/IEEE International Conference on Human-Robot Interaction (HRI) , 2018

    2018

  26. [34]

    Whose job is it anyway? a study of human-robot inter- action in a collaborative task

    Pamela J Hinds, Teresa L Roberts, and Hank Jones. Whose job is it anyway? a study of human-robot inter- action in a collaborative task. Human-Computer Interac- tion, 2004

    2004

  27. [35]

    Cost-based antic- ipatory action selection for human–robot fluency

    Guy Hoffman and Cynthia Breazeal. Cost-based antic- ipatory action selection for human–robot fluency. IEEE Transactions on Robotics , 23(5):952–961, 2007

    2007

  28. [36]

    Fleet-dagger: Interactive robot fleet learning with scalable human supervision

    Ryan Hoque, Lawrence Yunliang Chen, Satvik Sharma, Karthik Dharmarajan, Brijen Thananjeyan, Pieter Abbeel, and Ken Goldberg. Fleet-dagger: Interactive robot fleet learning with scalable human supervision. In Conference on Robot Learning (CoRL) , 2023

    2023

  29. [37]

    Human-automation coordination in multi-uav control

    Thomas Hughes. Human-automation coordination in multi-uav control. In AIAA Guidance, Navigation and Control Conference and Exhibit , page 6315, 2008

    2008

  30. [38]

    Traversing supervisor problem: An approximately opti- mal approach to multi-robot assistance

    Tianchen Ji, Roy Dong, and Katherine Driggs-Campbell. Traversing supervisor problem: An approximately opti- mal approach to multi-robot assistance. In Robotics: Science and Systems XVIII . Robotics: Science and Sys- tems Foundation, 2022. ISBN 978-0-9923747-8-5. doi: 10.15607/...

    2022 doi

  31. [39]

    Traversing supervisor problem: An approximately opti- mal approach to multi-robot assistance

    Tianchen Ji, Roy Dong, and Katherine Driggs-Campbell. Traversing supervisor problem: An approximately opti- mal approach to multi-robot assistance. In Robotics: Science and Systems (RSS) , 2022

    2022

  32. [40]

    When are two heads better than one and why? Science, 336(6079):360–362, 2012

    Asher Koriat. When are two heads better than one and why? Science, 336(6079):360–362, 2012

    2012

  33. [41]

    When two heads are better than one and when they can be worse: The amplification hypothesis

    Asher Koriat. When two heads are better than one and when they can be worse: The amplification hypothesis. Journal of Experimental Psychology: General , 144(5): 934–950, 2015. doi: 10.1037/xge0000092

    2015 doi

  34. [42]

    Boosting medical diagnostics by pooling independent judgments

    Ralf HJM Kurvers, Stefan M Herzog, Ralph Hertwig, Jens Krause, Patricia A Carney, Andy Bogart, Giuseppe Argenziano, Iris Zalaudek, and Max Wolf. Boosting medical diagnostics by pooling independent judgments. Proceedings of the National Academy of Sciences , 113 (31):8777–8782, 2016

    2016

  35. [43]

    Development of the human interactive auton- omy for the shared teleoperation of mobile robots

    Kwang-Hyun Lee, Usman Mehmood, and Jee-Hwan Ryu. Development of the human interactive auton- omy for the shared teleoperation of mobile robots. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , 2016

    2016

  36. [44]

    Transformed up-down methods in psy- choacoustics

    HCCH Levitt. Transformed up-down methods in psy- choacoustics. The Journal of the Acoustical society of America, 49(2B):467–477, 1971

    1971

  37. [45]

    Efficient learning of safe driving policy via human-ai copilot optimization

    Quanyi Li, Zhenghao Peng, and Bolei Zhou. Efficient learning of safe driving policy via human-ai copilot optimization. (arXiv:2202.10341), 2022. URL http: //arxiv.org/abs/2202.10341

    2022 arXiv

  38. [46]

    Role adaptation of human and robot in collaborative tasks

    Yanan Li, Keng Peng Tee, Wei Liang Chan, Rui Yan, Yuanwei Chua, and Dilip Kumar Limbu. Role adaptation of human and robot in collaborative tasks. In IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2015

    2015

  39. [47]

    What affects human decision making in human–robot collaboration?: A scoping review

    Yuan Liu, Glenda Caldwell, Markus Rittenbruch, M ¨uge Belek Fialho Teixeira, Alan Burden, and Matthias Guertler. What affects human decision making in human–robot collaboration?: A scoping review. Robotics, 13, 2024

    2024

  40. [48]

    Tutelage and socially guided robot learning

    Andrea Lockerd and Cynthia Breazeal. Tutelage and socially guided robot learning. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2004

    2004

  41. [49]

    Learning to make collective decisions: The impact of confidence escalation

    Ali Mahmoodi, Dan Bang, Majid Nili Ahmadabadi, and Bahador Bahrami. Learning to make collective decisions: The impact of confidence escalation. PLoS ONE, 8(12): e81195, 2013. doi: 10.1371/journal.pone.0081195

    2013 doi

  42. [50]

    Optimal group decision: A matter of confidence calibration

    S ´ebastien Massoni and Nicolas Roux. Optimal group decision: A matter of confidence calibration. Journal of Mathematical Psychology, 79:121–130, 2017

    2017

  43. [51]

    Modeling user expertise for choosing levels of shared autonomy

    Lauren Milliken and Geoffrey A Hollinger. Modeling user expertise for choosing levels of shared autonomy. In IEEE International Conference on Robotics and Au- tomation (ICRA). IEEE, 2017

    2017

  44. [52]

    Monferrer and D

    A. Monferrer and D. Bonyuet. Cooperative robot teleop- eration through virtual reality interfaces. In Proceedings Sixth International Conference on Information Visualisa- tion, 2002. doi: 10.1109/IV .2002.1028783

    2002 arXiv

  45. [53]

    Human-robot interaction in rescue robotics

    Robin R Murphy. Human-robot interaction in rescue robotics. IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews) , 34(2): 138–153, 2004

    2004

  46. [54]

    Robot team teleoperation for cooperative manipulation using wearable haptics

    Selma Musi ´c, Gionata Salvietti, Francesco Chinello, Domenico Prattichizzo, Sandra Hirche, et al. Robot team teleoperation for cooperative manipulation using wearable haptics. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2017

    2017

  47. [55]

    Emergency response to the nuclear accident at the Fukushima Daiichi Nuclear Power Plants using mobile rescue robots

    Keiji Nagatani, Seiga Kiribayashi, Yoshito Okada, Kazuki Otake, Kazuya Yoshida, Satoshi Tadokoro, Takeshi Nishimura, Tomoaki Yoshida, Eiji Koyanagi, Mineo Fukushima, et al. Emergency response to the nuclear accident at the Fukushima Daiichi Nuclear Power Plants using mobile re...

    2013

  48. [56]

    Models of trust in human control of swarms with varied levels of autonomy

    Changjoo Nam, Phillip Walker, Huao Li, Michael Lewis, and Katia Sycara. Models of trust in human control of swarms with varied levels of autonomy. IEEE Transac- tions on Human-Machine Systems, 50(3):194–204, 2019

    2019

  49. [57]

    Model structure identification of hybrid dynamical systems based on unsupervised clustering and variable selection

    Duc-An Nguyen, Jude Nwadiuto, Hiroyuki Okuda, and Tatsuya Suzuki. Model structure identification of hybrid dynamical systems based on unsupervised clustering and variable selection. IFAC-PapersOnLine, 53(2):1090– 1095, 2020

    2020

  50. [58]

    Modeling car-following behavior in downtown area based on unsupervised clustering and variable selection method

    Duc-An Nguyen, Jude Nwadiuto, Hiroyuki Okuda, and Tatsuya Suzuki. Modeling car-following behavior in downtown area based on unsupervised clustering and variable selection method. In IEEE International Con- ference on Systems, Man, and Cybernetics (SMC) , pages 3714–3720. IEEE, 2020

    2020

  51. [59]

    The perceptual and social components of metacognition

    Niccolo Pescetelli, Geraint Rees, and Bahador Bahrami. The perceptual and social components of metacognition. Journal of Experimental Psychology: General , 145(8): 949–965, 2016. doi: 10.1037/xge0000180

    2016 doi

  52. [60]

    Pre: Vision-language prompt learning with reparameterization encoder

    Thi Minh Anh Pham, An Duc Nguyen, Cephas Svosve, Vasileios Argyriou, and Georgios Tzimiropoulos. Pre: Vision-language prompt learning with reparameterization encoder. arXiv preprint arXiv:2309.07760 , 2023

    2023 arXiv

  53. [61]

    Pre: Vision-language prompt learning with reparameterization encoder

    Thi Minh Anh Pham, Duc-An Nguyen, Cephas Svosve, Vasileios Argyriou, and Georgios Tzimiropoulos. Pre: Vision-language prompt learning with reparameterization encoder. In International Conference on Learning Rep- resentations (ICLR 2024) , 2024

    2024

  54. [62]

    Physical collab- oration of human-human and human-robot teams

    Kyle B Reed and Michael A Peshkin. Physical collab- oration of human-human and human-robot teams. IEEE Transactions on Haptics , 1(2):108–120, 2008

    2008

  55. [63]

    Mechanisms for collaborative teleoperation with a team of cooperative robots

    O Reinoso, Arturo Gil, Luis Pay ´a, and Miguel Juli ´a. Mechanisms for collaborative teleoperation with a team of cooperative robots. Industrial Robot: An International Journal, 35(1):27–36, 2008

    2008

  56. [64]

    Allen, and Stephen M

    Marion Rouault, Andrew McWilliams, Micah G. Allen, and Stephen M. Fleming. Human metacognition across domains: Insights from individual differences and neu- roimaging. Personality Neuroscience, 1:e17, 2018. doi: 10.1017/pen.2018.16

    2018 doi

  57. [65]

    Psychiatric symptom dimensions are associated with dissociable shifts in metacognition but not task performance

    Marion Rouault, Tricia Seow, Claire M Gillan, and Stephen M Fleming. Psychiatric symptom dimensions are associated with dissociable shifts in metacognition but not task performance. Biological psychiatry, 84(6): 443–451, 2018

    2018

  58. [66]

    Forming global estimates of self-performance from local confidence

    Marion Rouault, Peter Dayan, and Stephen M Fleming. Forming global estimates of self-performance from local confidence. Nature communications, 10(1):1141, 2019

    2019

  59. [67]

    Flexible cooperation between human and robot by interpreting human inten- tion from gaze information

    Kenji Sakita, Koichi Ogawara, Shinji Murakami, Kentaro Kawamura, and Katsushi Ikeuchi. Flexible cooperation between human and robot by interpreting human inten- tion from gaze information. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2004

    2004

  60. [68]

    Evaluation of human-robot interaction awareness in search and rescue

    Jean Scholtz, Jeff Young, Jill L Drury, and Holly A Yanco. Evaluation of human-robot interaction awareness in search and rescue. In IEEE International Conference on Robotics and Automation (ICRA) , volume 3. IEEE, 2004

    2004

  61. [69]

    A collaborative-shared control sys- tem with safe obstacle avoidance capability

    Jian Shen, Javier Ibanez-Guzman, Teck Chew Ng, and Boon Seng Chew. A collaborative-shared control sys- tem with safe obstacle avoidance capability. In IEEE Conference on Robotics, Automation and Mechatronics (ICARM)., volume 1, pages 119–123. IEEE, 2004

    2004

  62. [70]

    Costume vs

    Sichao Song, Jun Baba, Junya Nakanishi, Yuichiro Yoshikawa, and Hiroshi Ishiguro. Costume vs. wizard of oz vs. telepresence: how social presence forms of tele-operated robots influence customer behavior. In ACM/IEEE International Conference on Human-Robot Interaction (HRI). IEEE, 2022

    2022

  63. [71]

    wait, can you move the robot?

    Brett Stoll, Samantha Reig, Lucy He, Ian Kaplan, Malte F. Jung, and Susan R. Fussell. “wait, can you move the robot?”: Examining telepresence robot use in collaborative teams. In ACM/IEEE International Confer- ence on Human-Robot Interaction (HRI) , 2018

    2018

  64. [72]

    Pre- dicting human performance during teleoperation

    Justin Storms, Steven V ozar, and Dawn Tilbury. Pre- dicting human performance during teleoperation. In ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2014

    2014

  65. [73]

    Teleoperated or autonomous?: How to produce a robot operator’s pseudo presence in hri

    Kazuaki Tanaka, Naomi Yamashita, Hideyuki Nakanishi, and Hiroshi Ishiguro. Teleoperated or autonomous?: How to produce a robot operator’s pseudo presence in hri. In ACM/IEEE International Conference on Human- Robot Interaction (HRI) , 2016. doi: 10.1109/HRI.2016. 7451744

    2016 doi

  66. [74]

    Enabling effective human-robot interaction using perspective-taking in robots

    J Gregory Trafton, Nicholas L Cassimatis, Magdalena D Bugajska, Derek P Brock, Farilee E Mintz, and Alan C Schultz. Enabling effective human-robot interaction using perspective-taking in robots. IEEE Transactions on Systems, Man, and Cybernetics-Part A: Systems and Humans, 35(...

    2005

  67. [75]

    Tsui, Munjal Desai, Holly A

    Katherine M. Tsui, Munjal Desai, Holly A. Yanco, and Chris Uhlik. Exploring use cases for telepresence robots. In ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2011

    2011

  68. [76]

    Tsui, Stephen V on Rump, Hiroshi Ishiguro, Leila Takayama, and Peter Vicars

    Katherine M. Tsui, Stephen V on Rump, Hiroshi Ishiguro, Leila Takayama, and Peter Vicars. Robots in the loop: Telepresence robots in everyday life. In ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2012

    2012

  69. [77]

    Carter, Braden McDor- man, Jodi Forlizzi, Aaron Steinfeld, and Scott E

    Marynel V ´azquez, Elizabeth J. Carter, Braden McDor- man, Jodi Forlizzi, Aaron Steinfeld, and Scott E. Hud- son. Towards robot autonomy in group conversations: Understanding the effects of body orientation and gaze. In ACM/IEEE International Conference on Human-Robot Interact...

    2017

  70. [78]

    Walker, Hooman Hedayati, and Daniel Szafir

    Michael E. Walker, Hooman Hedayati, and Daniel Szafir. Robot teleoperation with augmented reality virtual sur- rogates. In ACM/IEEE International Conference on Human-Robot Interaction (HRI) , 2019. doi: 10.1109/ HRI.2019.8673306

    2019

  71. [2014]

    doi: 10.1016/j.concog.2014.02.002

    2014 doi

  72. [2017]

    doi: 10.1038/s41562-017-0117

    ISSN 2397-3374. doi: 10.1038/s41562-017-0117

    doi

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.