arxiv: 2605.12303 · v1 · submitted 2026-05-12 · 💻 cs.HC · cs.CV· cs.LG

Recognition: no theorem link

From Model Uncertainty to Human Attention: Localization-Aware Visual Cues for Scalable Annotation Review

Moussa Kassem Sbeyti , Joshua Holstein , Philipp Spitzer , Nadja Klein , Gerhard Satzger

Authors on Pith no claims yet

Pith reviewed 2026-05-13 03:52 UTC · model grok-4.3

classification 💻 cs.HC cs.CVcs.LG

keywords localization uncertaintyvisual cuesannotation reviewhuman-in-the-loopAI-assisted annotationbounding boxesmodel uncertaintyscalable labeling

0 comments

The pith

Showing localization uncertainty helps annotators create better labels faster by targeting likely mistakes.

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

This paper tests whether displaying a model's uncertainty about object locations in images can improve how humans review and correct AI-generated annotations. In experiments with 120 participants, providing these visual cues led to higher quality final labels and reduced overall time spent. The cues work by drawing attention to predictions that are likely spatially incorrect, allowing reviewers to skip over accurate ones. This matters because annotation is a major cost in building machine learning systems, and better guidance could reduce both errors and expenses in large labeling projects.

Core claim

The central finding is that localization-aware uncertainty visualizations in an annotation interface enable participants to achieve superior label accuracy while working more quickly overall. Box-level analysis shows that these cues successfully shift annotator focus toward high-uncertainty regions and away from well-predicted boxes.

What carries the argument

Localization uncertainty visualization, which highlights areas where the model is uncertain about the precise boundaries of detected objects.

If this is right

Higher quality annotations result from targeted review of uncertain predictions.
Annotators complete tasks faster when guided by uncertainty signals.
Attention is redirected from low-error to high-error predictions.
This establishes localization uncertainty as a practical tool for scalable annotation.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Integrating such cues into commercial annotation platforms could enhance productivity in real-world data labeling operations.
The method might extend to other spatial tasks such as keypoint detection or instance segmentation.
Combining uncertainty cues with other AI assistance techniques could further optimize human effort.
Validation in non-lab environments with expert annotators would strengthen the evidence for adoption.

Load-bearing premise

The model's localization uncertainty must accurately correspond to actual spatial errors in its predictions.

What would settle it

Compare the number of localization errors corrected by participants with and without the uncertainty cues, using a held-out ground truth to count differences in final label quality.

read the original abstract

High-quality labeled data is essential for training robust machine learning models, yet obtaining annotations at scale remains expensive. AI-assisted annotation has therefore become standard in large-scale labeling workflows. However, in tasks where model predictions carry two independent components, a class label and spatial boundaries, a model may classify an object with high confidence while mislocalizing it. Existing AI-assisted workflows offer annotators no signal about where spatial errors are most likely. Without such guidance, humans may systematically underinspect subtly misplaced boxes. We address this by studying the effect of visualizing spatial uncertainty via a purpose-built interface. In a controlled study with 120 participants, those receiving uncertainty cues achieve higher label quality while being faster overall. A box-level analysis confirms that the cues redirect annotator effort toward high-uncertainty predictions and away from well-localized boxes. These findings establish localization uncertainty as a lever to improve human-in-the-loop annotation. Code is available at https://mos-ks.github.io/MUHA/.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The 120-person study shows uncertainty cues can speed up and improve annotation review, but the paper skips the key check of whether those cues actually track real localization errors on the test images.

read the letter

The main takeaway is that visualizing localization uncertainty in an annotation interface led to better labels and faster work in a controlled study with 120 participants. The box-level analysis also shows annotators spent more time on high-uncertainty boxes and less on well-localized ones. That is a practical result for anyone running human-in-the-loop pipelines on object detection tasks where models get the class right but the box wrong.

Referee Report

2 major / 2 minor

Summary. The manuscript investigates the use of localization uncertainty visualizations in AI-assisted bounding box annotation interfaces. Through a controlled user study with 120 participants, it claims that providing visual cues for spatial uncertainty improves annotation quality while reducing overall time, with a box-level analysis showing that annotators redirect effort toward high-uncertainty predictions and away from well-localized ones. The work positions localization uncertainty as a practical lever for scalable human-in-the-loop labeling, distinct from class-confidence signals, and releases code for the interface.

Significance. If the empirical results hold under scrutiny, the contribution is meaningful for human-AI collaboration in computer vision data pipelines. It supplies concrete evidence that targeted uncertainty cues can improve both efficiency and accuracy in annotation review, addressing a gap where standard confidence scores fail to flag localization errors. The availability of reproducible code strengthens the work by enabling direct follow-up and interface replication.

major comments (2)

Box-level analysis (and associated results): The manuscript reports that uncertainty cues redirect annotator effort but does not include a direct validation—such as Pearson correlation, AUC, or regression—between the model's localization uncertainty scores and actual spatial errors (e.g., 1-IoU deviation from ground truth) measured on the exact images shown to participants. Without this check, the causal link between the uncertainty signal and the observed quality/speed gains remains unestablished; improvements could stem from any salient visual highlighting rather than the specific uncertainty estimates.
User study reporting (results section): The abstract and study description claim higher label quality and faster performance for the uncertainty-cue condition, yet provide no statistical details (effect sizes, confidence intervals, p-values, exclusion criteria, or power analysis). This omission makes it impossible to assess the robustness of the 120-participant findings or to evaluate whether the box-level redirection effect is statistically reliable.

minor comments (2)

Abstract: The summary would benefit from one or two concrete quantitative outcomes (e.g., mean quality improvement or time reduction) to give readers an immediate sense of effect magnitude.
Interface description: Clarify how the uncertainty visualization is rendered (color mapping, opacity scaling, or contour style) and whether it was pilot-tested for perceptual salience independent of the uncertainty values themselves.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, which highlights opportunities to strengthen the empirical grounding of our results. We address each major comment below and will incorporate revisions to improve statistical transparency and validation of the uncertainty signal.

read point-by-point responses

Referee: Box-level analysis (and associated results): The manuscript reports that uncertainty cues redirect annotator effort but does not include a direct validation—such as Pearson correlation, AUC, or regression—between the model's localization uncertainty scores and actual spatial errors (e.g., 1-IoU deviation from ground truth) measured on the exact images shown to participants. Without this check, the causal link between the uncertainty signal and the observed quality/speed gains remains unestablished; improvements could stem from any salient visual highlighting rather than the specific uncertainty estimates.

Authors: We agree that a direct quantitative validation between the localization uncertainty scores and actual spatial errors would reinforce the specificity of the signal. Our box-level analysis demonstrates annotator redirection toward high-uncertainty boxes, but we will add a Pearson correlation (and optionally regression) between the model's uncertainty estimates and 1-IoU deviations from ground truth, computed on the precise images shown to participants. Ground truth is available from our quality metrics, so this analysis is feasible and will be reported in the revised results section. revision: yes
Referee: User study reporting (results section): The abstract and study description claim higher label quality and faster performance for the uncertainty-cue condition, yet provide no statistical details (effect sizes, confidence intervals, p-values, exclusion criteria, or power analysis). This omission makes it impossible to assess the robustness of the 120-participant findings or to evaluate whether the box-level redirection effect is statistically reliable.

Authors: We acknowledge that the current results section lacks the requested statistical details. In the revision we will report effect sizes (Cohen's d), 95% confidence intervals, p-values from the appropriate tests (t-tests or mixed-effects models), explicit exclusion criteria, and a post-hoc power analysis for the 120-participant sample. These additions will apply to both the overall condition comparisons and the box-level redirection findings. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical user study with no derivation chain

full rationale

The paper reports results from a controlled user study involving 120 participants comparing annotation interfaces with and without localization uncertainty cues. Claims rest on direct empirical measurements of label quality, speed, and effort redirection rather than any mathematical derivation, fitted parameters, or model predictions. No equations, self-definitional steps, or load-bearing self-citations appear in the provided abstract or described structure. The central findings are falsifiable via the study data itself and do not reduce to their inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim depends on the validity of the user-study design and the assumption that model-provided uncertainty maps correspond to real localization errors; these are not independently verified in the abstract.

axioms (1)

domain assumption Behavior of the 120 study participants is representative of annotators in real-world scalable labeling pipelines.
The abstract presents results from a controlled experiment but does not discuss how participant pool or task conditions map to production annotation settings.

pith-pipeline@v0.9.0 · 5487 in / 1230 out tokens · 119913 ms · 2026-05-13T03:52:00.206421+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

89 extracted references · 89 canonical work pages

[1]

Dropout as a

Gal, Yarin and Ghahramani, Zoubin , booktitle=. Dropout as a. 2016 , publisher=

work page 2016
[3]

Tan, Mingxing and Pang, Ruoming and Le, Quoc V , journal=

work page
[4]

Are we ready for autonomous driving?

Geiger, Andreas and Lenz, Philip and Urtasun, Raquel , journal=. Are we ready for autonomous driving?

work page
[5]

HCOMP@ AAAI , volume=

Crowdsourcing annotations for visual object detection , author=. HCOMP@ AAAI , volume=

work page
[6]

Efficiently scaling up crowdsourced video annotation:

Vondrick, Carl and Patterson, Donald and Ramanan, Deva , journal=. Efficiently scaling up crowdsourced video annotation:. 2013 , publisher=

work page 2013
[7]

European Conference on Computer Vision (

Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll. European Conference on Computer Vision (

work page
[8]

2015 , publisher=

Russakovsky, Olga and Deng, Jia and Su, Hao and Krause, Jonathan and Satheesh, Sanjeev and Ma, Sean and Huang, Zhiheng and Karpathy, Andrej and Khosla, Aditya and Bernstein, Michael and others , journal=. 2015 , publisher=

work page 2015
[9]

Data annotation quality problems in

Saeeda, Hina and Johansson, Tommy and Mohamad, Mazen and Knauss, Eric , journal=. Data annotation quality problems in

work page
[10]

Information Systems , volume=

The effects of data quality on machine learning performance on tabular data , author=. Information Systems , volume=. 2025 , publisher=

work page 2025
[11]

Information and Software Technology , volume=

A survey on dataset quality in machine learning , author=. Information and Software Technology , volume=. 2023 , publisher=

work page 2023
[12]

What uncertainties do we need in

Kendall, Alex and Gal, Yarin , journal=. What uncertainties do we need in

work page
[13]

Advances in Neural Information Processing Systems , volume=

Simple and scalable predictive uncertainty estimation using deep ensembles , author=. Advances in Neural Information Processing Systems , volume=

work page
[14]

Artificial Intelligence Review , volume=

A survey of uncertainty in deep neural networks , author=. Artificial Intelligence Review , volume=. 2023 , publisher=

work page 2023
[15]

Nature Reviews Neuroscience , volume=

Computational modelling of visual attention , author=. Nature Reviews Neuroscience , volume=. 2001 , publisher=

work page 2001
[16]

Cognitive load during problem solving:

Sweller, John , journal=. Cognitive load during problem solving:. 1988 , publisher=

work page 1988
[17]

Guided search 6.0:

Wolfe, Jeremy M , journal=. Guided search 6.0:. 2021 , publisher=

work page 2021
[18]

Overcoming the limitations of localization uncertainty:

Kassem Sbeyti, Moussa and Karg, Michelle and Wirth, Christian and Nowzad, Azarm and Albayrak, Sahin , journal=. Overcoming the limitations of localization uncertainty:

work page
[19]

Uncertainty in

Cost-sensitive uncertainty-based failure recognition for object detection , author=. Uncertainty in

work page
[21]

Proceedings of the

Active learning for deep object detection via probabilistic modeling , author=. Proceedings of the

work page
[22]

Nature Communications , volume=

Annotation-efficient deep learning for automatic medical image segmentation , author=. Nature Communications , volume=. 2021 , publisher=

work page 2021
[23]

Nature Communications , volume=

A human-machine collaborative approach measures economic development using satellite imagery , author=. Nature Communications , volume=. 2023 , publisher=

work page 2023
[24]

Can you trust your model's uncertainty?

Ovadia, Yaniv and Fertig, Emily and Ren, Jie and Nado, Zachary and Sculley, David and Nowozin, Sebastian and Dillon, Joshua and Lakshminarayanan, Balaji and Snoek, Jasper , journal=. Can you trust your model's uncertainty?

work page
[25]

The speed-accuracy tradeoff:

Heitz, Richard P , journal=. The speed-accuracy tradeoff:. 2014 , publisher=

work page 2014
[26]

Automation bias:

Goddard, Kate and Roudsari, Abdul and Wyatt, Jeremy C , journal=. Automation bias:. 2012 , publisher=

work page 2012
[27]

When combinations of humans and

Vaccaro, Michelle and Almaatouq, Abdullah and Malone, Thomas , journal=. When combinations of humans and. 2024 , publisher=

work page 2024
[28]

Uncertainty as a form of transparency:

Bhatt, Umang and Antor. Uncertainty as a form of transparency:. Proceedings of the 2021

work page 2021
[29]

Aleatoric and epistemic uncertainty in machine learning:

H. Aleatoric and epistemic uncertainty in machine learning:. Machine Learning , volume=. 2021 , publisher=

work page 2021
[30]

Proceedings of the International Conference on Machine Learning (

On calibration of modern neural networks , author=. Proceedings of the International Conference on Machine Learning (

work page
[31]

Explainable

Senoner, Julian and Schallmoser, Simon and Kratzwald, Bernhard and Feuerriegel, Stefan and Netland, Torbj. Explainable. Scientific Reports , volume=. 2024 , publisher=

work page 2024
[32]

Balancing the unknown:

Holstein, Joshua and B\". Balancing the unknown:. ACM Transactions on Computer-Human Interaction , volume=. 2025 , publisher=

work page 2025
[33]

Nature Machine Intelligence , volume=

Labelling instructions matter in biomedical image analysis , author=. Nature Machine Intelligence , volume=. 2023 , publisher=

work page 2023
[34]

Asian Conference on Computer Vision (

Localization-aware active learning for object detection , author=. Asian Conference on Computer Vision (

work page
[35]

Proceedings of the

Plug and play active learning for object detection , author=. Proceedings of the

work page
[36]

Confident learning:

Northcutt, Curtis and Jiang, Lu and Chuang, Isaac , journal=. Confident learning:

work page
[37]

Nature Communications , volume=

Active label cleaning for improved dataset quality under resource constraints , author=. Nature Communications , volume=. 2022 , publisher=

work page 2022
[38]

Learning from noisy labels with deep neural networks:

Song, Hwanjun and Kim, Minseok and Park, Dongmin and Shin, Yooju and Lee, Jae-Gil , journal=. Learning from noisy labels with deep neural networks:. 2022 , publisher=

work page 2022
[39]

A survey on autonomous driving datasets:

Liu, Mingyu and Yurtsever, Ekim and Fossaert, Jonathan and Zhou, Xingcheng and Zimmer, Walter and Cui, Yuning and Zagar, Bare Luka and Knoll, Alois C , journal=. A survey on autonomous driving datasets:. 2024 , publisher=

work page 2024
[40]

No need to sacrifice data quality for quantity:

Klugmann, Christopher and Mahmood, Rafid and Hegde, Guruprasad and Kale, Amit and Kondermann, Daniel , journal=. No need to sacrifice data quality for quantity:

work page
[41]

Transactions on Machine Learning Research , year=

Building blocks for robust and effective semi-supervised real-world object detection , author=. Transactions on Machine Learning Research , year=

work page
[42]

Williams, Alex C and Bai, Min and Buck, Jonathan and McKinney, Tristan J and Rechkemmer, Amy and Kalyanaraman, Koushik and Lease, Matthew and Haffner, Patrick and Zhou, Xiong and Li, Erran , journal=

work page
[44]

2026 , howpublished=

Streamlit , author=. 2026 , howpublished=

work page 2026
[45]

Proceedings of the 25th International Conference on Pattern Recognition (

Iterative bounding box annotation for object detection , author=. Proceedings of the 25th International Conference on Pattern Recognition (

work page
[46]

Proceedings of the

Extreme clicking for efficient object annotation , author=. Proceedings of the

work page
[47]

It's not just labeling:

Li, Baichuan and Powell, Larry and Hammond, Tracy , journal=. It's not just labeling:

work page
[48]

European Journal of Information Systems , volume=

Complementarity in human-AI collaboration: Concept, sources, and evidence , author=. European Journal of Information Systems , volume=. 2025 , publisher=

work page 2025
[49]

Proceedings of the AAAI Conference on Human Computation and Crowdsourcing , volume=

A taxonomy of human and ML strengths in decision-making to investigate human-ML complementarity , author=. Proceedings of the AAAI Conference on Human Computation and Crowdsourcing , volume=

work page
[50]

Scientific reports , volume=

Efficient automated error detection in medical data using deep-learning and label-clustering , author=. Scientific reports , volume=. 2023 , publisher=

work page 2023
[51]

& Urtasun, R

Geiger, A., Lenz, P. & Urtasun, R. Are we ready for autonomous driving? The KITTI vision benchmark suite. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 3354--3361 (2012)

work page 2012
[52]

, author Schallmoser, S

author Senoner, J. , author Schallmoser, S. , author Kratzwald, B. , author Feuerriegel, S. & author Netland, T. title Explainable AI improves task performance in human-- AI collaboration . journal Scientific Reports volume 14 , pages 31150 ( year 2024 )

work page 2024
[53]

, author Almaatouq, A

author Vaccaro, M. , author Almaatouq, A. & author Malone, T. title When combinations of humans and AI are useful: A systematic review and meta-analysis . journal Nature Human Behaviour volume 8 , pages 2293--2303 ( year 2024 )

work page 2024
[54]

author Holstein, J. et al. title Balancing the unknown: E xploring human reliance on AI advice under aleatoric and epistemic uncertainty . journal ACM Transactions on Computer-Human Interaction volume 32 , pages 64 ( year 2025 )

work page 2025
[55]

author Bhatt, U. et al. title Uncertainty as a form of transparency: M easuring, communicating, and using uncertainty . journal Proceedings of the 2021 AAAI/ACM Conference on AI , Ethics, and Society pages 401--413 ( year 2021 )

work page 2021
[56]

, author Pleiss, G

author Guo, C. , author Pleiss, G. , author Sun, Y. & author Weinberger, K. Q. title On calibration of modern neural networks . journal Proceedings of the International Conference on Machine Learning ( ICML ) pages 1321--1330 ( year 2017 )

work page 2017
[57]

, author Pritzel, A

author Lakshminarayanan, B. , author Pritzel, A. & author Blundell, C. title Simple and scalable predictive uncertainty estimation using deep ensembles . journal Advances in Neural Information Processing Systems volume 30 ( year 2017 )

work page 2017
[58]

, author Elezi, I

author Choi, J. , author Elezi, I. , author Lee, H.-J. , author Farabet, C. & author Alvarez, J. M. title Active learning for deep object detection via probabilistic modeling . journal Proceedings of the IEEE/CVF International Conference on Computer Vision ( ICCV ) pages 10264--10273 ( year 2021 )

work page 2021
[59]

, author Karg, M

author Kassem Sbeyti, M. , author Karg, M. , author Wirth, C. , author Klein, N. & author Albayrak, S. title Cost-sensitive uncertainty-based failure recognition for object detection . journal Uncertainty in A rtificial I ntelligence pages 1890--1900 ( year 2024 )

work page 1900
[60]

author Lin, T.-Y. et al. title Microsoft COCO : C ommon objects in context . journal European Conference on Computer Vision ( ECCV ) pages 740--755 ( year 2014 )

work page 2014
[61]

author Russakovsky, O. et al. title ImageNet large scale visual recognition challenge . journal International Journal of Computer Vision volume 115 , pages 211--252 ( year 2015 )

work page 2015
[62]

, author Jiang, L

author Northcutt, C. , author Jiang, L. & author Chuang, I. title Confident learning: E stimating uncertainty in dataset labels . journal Journal of Artificial Intelligence Research volume 70 , pages 1373--1411 ( year 2021 )

work page 2021
[63]

, author Kim, M

author Song, H. , author Kim, M. , author Park, D. , author Shin, Y. & author Lee, J.-G. title Learning from noisy labels with deep neural networks: A survey . journal IEEE Transactions on Neural Networks and Learning Systems volume 34 , pages 8135--8153 ( year 2022 )

work page 2022
[64]

author Nguyen, T. et al. title Efficient automated error detection in medical data using deep-learning and label-clustering . journal Scientific reports volume 13 , pages 19587 ( year 2023 )

work page 2023
[65]

, author Klein, N

author Kassem Sbeyti, M. , author Klein, N. , author Nowzad, A. , author Sivrikaya, F. & author Albayrak, S. title Building blocks for robust and effective semi-supervised real-world object detection . journal Transactions on Machine Learning Research ( year 2025 )

work page 2025
[66]

, author Johansson, T

author Saeeda, H. , author Johansson, T. , author Mohamad, M. & author Knauss, E. title Data annotation quality problems in AI -enabled perception system development . journal arXiv preprint arXiv:2511.16410 ( year 2025 )

work page arXiv 2025
[67]

author Zhan, X. et al. title A comparative survey of deep active learning . journal arXiv preprint arXiv:2203.13450 ( year 2022 )

work page arXiv 2022
[68]

, author Lee, T.-Y

author Kao, C.-C. , author Lee, T.-Y. , author Sen, P. & author Liu, M.-Y. title Localization-aware active learning for object detection . journal Asian Conference on Computer Vision ( ACCV ) pages 506--522 ( year 2018 )

work page 2018
[69]

, author Huang, L

author Yang, C. , author Huang, L. & author Crowley, E. J. title Plug and play active learning for object detection . journal Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition ( CVPR ) pages 17784--17793 ( year 2024 )

work page 2024
[70]

& author Gal, Y

author Kendall, A. & author Gal, Y. title What uncertainties do we need in B ayesian deep learning for computer vision? journal Advances in Neural Information Processing Systems volume 30 ( year 2017 )

work page 2017
[71]

& author Waegeman, W

author H \"u llermeier, E. & author Waegeman, W. title Aleatoric and epistemic uncertainty in machine learning: A n introduction to concepts and methods . journal Machine Learning volume 110 , pages 457--506 ( year 2021 )

work page 2021
[72]

, author Karg, M

author Kassem Sbeyti, M. , author Karg, M. , author Wirth, C. , author Nowzad, A. & author Albayrak, S. title Overcoming the limitations of localization uncertainty: E fficient and exact non-linear post-processing and calibration . journal Joint European Conference on Machine Learning and Knowledge Discovery in Databases ( ECML PKDD ) pages 52--68 ( year 2023 )

work page 2023
[73]

author Williams, A. C. et al. title Snapper : A ccelerating bounding box annotation in object detection tasks with find-and-snap tooling . journal Proceedings of the 29th International Conference on Intelligent User Interfaces ( IUI ) pages 471--488 ( year 2024 )

work page 2024
[74]

author Papadopoulos, D. P. , author Uijlings, J. R. , author Keller, F. & author Ferrari, V. title Extreme clicking for efficient object annotation . journal Proceedings of the IEEE International Conference on Computer Vision ( ICCV ) pages 4930--4939 ( year 2017 )

work page 2017
[75]

, author Powell, L

author Li, B. , author Powell, L. & author Hammond, T. title It's not just labeling: A research on LLM generated feedback interpretability and image labeling sketch features . journal arXiv preprint arXiv:2505.19419 ( year 2025 )

work page arXiv 2025
[76]

& author Huttunen, H

author Adhikari, B. & author Huttunen, H. title Iterative bounding box annotation for object detection . journal Proceedings of the 25th International Conference on Pattern Recognition ( ICPR ) pages 4040--4046 ( year 2021 )

work page 2021
[77]

author Henley, A. Z. & author Piorkowski, D. title Supporting annotators with affordances for efficiently labeling conversational data . journal arXiv preprint arXiv:2403.07762 ( year 2024 )

work page arXiv 2024
[78]

u hl, N. , author V \

author Hemmer, P. , author Schemmer, M. , author K \"u hl, N. , author V \"o ssing, M. & author Satzger, G. title Complementarity in human-ai collaboration: Concept, sources, and evidence . journal European Journal of Information Systems volume 34 , pages 979--1002 ( year 2025 )

work page 2025
[79]

& author Koch, C

author Itti, L. & author Koch, C. title Computational modelling of visual attention . journal Nature Reviews Neuroscience volume 2 , pages 194--203 ( year 2001 )

work page 2001
[80]

author Wolfe, J. M. title Guided search 6.0: A n updated model of visual search . journal Psychonomic Bulletin & Review volume 28 , pages 1060--1092 ( year 2021 )

work page 2021
[81]

title Cognitive load during problem solving: E ffects on learning

author Sweller, J. title Cognitive load during problem solving: E ffects on learning . journal Cognitive Science volume 12 , pages 257--285 ( year 1988 )

work page 1988
[82]

author Heitz, R. P. title The speed-accuracy tradeoff: H istory, physiology, methodology, and behavior . journal Frontiers in Neuroscience volume 8 , pages 150 ( year 2014 )

work page 2014
[83]

, author Leqi, L

author Rastogi, C. , author Leqi, L. , author Holstein, K. & author Heidari, H. title A taxonomy of human and ml strengths in decision-making to investigate human-ml complementarity . journal Proceedings of the AAAI Conference on Human Computation and Crowdsourcing volume 11 , pages 127--139 ( year 2023 )

work page 2023

Showing first 80 references.