pith. machine review for the scientific record. sign in

arxiv: 2602.06489 · v2 · submitted 2026-02-06 · 💻 cs.HC

Recognition: 2 theorem links

· Lean Theorem

Simulating Word Suggestion Usage in Mobile Typing to Guide Intelligent Text Entry Design

Yang Li , Anna Maria Feit
Authors on Pith no claims yet

Pith reviewed 2026-05-16 07:03 UTC · model grok-4.3

classification 💻 cs.HC
keywords word suggestionsmobile typingreinforcement learninguser simulationintelligent text entrycomputational modeldesign analysis
0
0 comments X

The pith

WSTypist, a reinforcement learning model, simulates how mobile typists decide to use word suggestions and supports what-if design testing.

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

The paper introduces WSTypist to model the high-level decisions people make about accepting word suggestions while typing on phones. It extends existing hierarchical typing models with three mechanisms: orthographic similarity checks, efficiency tradeoff evaluation, and individual preferences for AI help. This matters because improving suggestion systems currently requires long user studies to observe how behavior changes over time. The model reproduces varied human strategies, personal differences, and performance across systems, allowing designers to simulate adaptation to new interfaces or algorithms instead.

Core claim

WSTypist is a reinforcement learning-based model that simulates typists' integration of word suggestions into manual typing. It adds cognitive mechanisms for orthographic processes, efficiency gains assessment, and personal preference on AI support to prior hierarchical control models. Evaluations show the model produces human-like suggestion-use strategies, reproduces individual differences, and generalizes across systems. Four design cases demonstrate its use for simulating user adaptation to UI or algorithmic changes.

What carries the argument

WSTypist, a reinforcement learning agent that decides on word suggestion use by weighing orthographic similarity, efficiency benefits, and user-specific AI preferences inside a hierarchical typing control structure.

If this is right

  • The model reproduces diverse human-like strategies for using or ignoring word suggestions.
  • It captures and reproduces individual differences in suggestion usage across users.
  • It generalizes its predictions to different intelligent text entry systems.
  • It supports what-if analyses that forecast how users adapt to changes in UI layout or suggestion algorithms.
  • The approach reduces reliance on long-term user studies during the design process.

Where Pith is reading between the lines

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

  • Designers could run rapid iterations on new suggestion features by first testing predicted user responses in simulation.
  • Similar decision models might apply to other AI-assisted tools where users choose when to accept automated help.
  • Tuning the preference parameter could support creating more personalized text entry systems for different user groups.
  • Adding motor execution details to the model might improve accuracy for short typing sessions or error-prone scenarios.

Load-bearing premise

The three mechanisms of orthographic processing, efficiency assessment, and preference for AI support are sufficient to capture the main patterns in human decision-making about suggestion use.

What would settle it

Run a new study with users on an unseen suggestion interface and check whether the model's predicted changes in acceptance rates and typing speed match the actual observed adaptation over sessions.

Figures

Figures reproduced from arXiv: 2602.06489 by Anna Maria Feit, Yang Li.

Figure 1
Figure 1. Figure 1: Overall structure of the WSTypist model. We extend an existing hierarchical architecture to simulate suggestion use in mobile [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Effects of word length on suggestion usage by the agent, revealing a peak in selection rates for medium-length words and a [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Influence of suggestion accuracy on user behavior, including effects on Picked, Failed suggestions, Keystroke Savings, Gaze [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
read the original abstract

Intelligent text entry (ITE) methods, such as word suggestions, are widely used in mobile typing, yet improving ITE systems is challenging because the cognitive mechanisms behind suggestion use remain poorly understood, and evaluating new systems often requires long-term user studies to account for behavioral adaptation. We present WSTypist, a reinforcement learning-based model that simulates how typists integrate word suggestions into typing. We extend recent hierarchical control models of typing, by identifying and implementing important cognitive mechanisms that underlie the high-level decision-making for integrating word suggestions into manual typing: considering orthographic processes, assessing efficiency gains, and including personal preference on AI support. Our evaluations show that WSTypist simulates diverse human-like suggestion-use strategies, reproduces individual differences, and generalizes across different systems. Importantly, we demonstrate on four design cases how a computational rationality model can be used to inform what-if analyses during the design process, by simulating how users might adapt to changes in the UI or in the algorithmic support, reducing the need for long-term user studies.

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.

Referee Report

2 major / 2 minor

Summary. The paper introduces WSTypist, a reinforcement learning model that extends hierarchical control models of typing to simulate users' integration of word suggestions. It incorporates three cognitive mechanisms—orthographic processes, efficiency gains assessment, and personal preference on AI support—to model high-level decision-making. The central claims are that the model reproduces diverse human-like suggestion-use strategies, captures individual differences, generalizes across systems, and enables what-if design analyses in four cases to predict user adaptation to UI or algorithmic changes without long-term studies.

Significance. If the validation holds, the work provides a practical computational rationality framework for intelligent text entry design, allowing simulation of behavioral adaptation to reduce reliance on resource-intensive user studies. The explicit modeling of cognitive mechanisms and demonstration on design cases represent a constructive step toward predictive tools in HCI.

major comments (2)
  1. [Evaluations] Evaluations (as referenced in the abstract and §4–5): the claims that WSTypist 'simulates diverse human-like suggestion-use strategies, reproduces individual differences, and generalizes across different systems' rest on unspecified validation datasets, metrics, baselines, and error analysis. Without these details, it is impossible to assess whether the RL policy matches observed behaviors due to the claimed mechanisms or other latent factors.
  2. [Model mechanisms] Model mechanisms (§3): the assertion that orthographic processes, efficiency gains assessment, and personal preference on AI support are sufficient to capture high-level decision-making lacks ablation testing. If the hierarchical RL extension can reproduce the data through reward shaping or other parameters alone, the cognitive grounding and consequent generalization to UI/algorithm changes would be weakened.
minor comments (2)
  1. [Abstract] The abstract states that four design cases are demonstrated but provides no high-level description of what the cases involve (e.g., specific UI changes or algorithmic modifications). Adding one sentence would clarify the scope of the what-if analyses.
  2. [Model description] Notation for the RL reward weights (listed as free parameters) should be explicitly defined with symbols when first introduced to aid readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback and positive assessment of the work's potential. We address the two major comments below and will revise the manuscript to improve clarity and add supporting analyses where needed.

read point-by-point responses

  1. Referee: [Evaluations] Evaluations (as referenced in the abstract and §4–5): the claims that WSTypist 'simulates diverse human-like suggestion-use strategies, reproduces individual differences, and generalizes across different systems' rest on unspecified validation datasets, metrics, baselines, and error analysis. Without these details, it is impossible to assess whether the RL policy matches observed behaviors due to the claimed mechanisms or other latent factors.

    Authors: We appreciate the referee pointing this out. The evaluations in §4–5 are based on human typing datasets from prior studies involving mobile word suggestion interfaces (with details on participant numbers, suggestion conditions, and logging of acceptance rates), using metrics including suggestion usage frequency, typing speed (WPM), error rate, and strategy classification via clustering. Baselines include non-hierarchical RL variants and heuristic models from prior HCI literature, with error analysis via per-participant fits and statistical comparisons. We will revise to explicitly enumerate the datasets (with citations), metrics tables, baseline implementations, and error breakdowns at the start of §4 to make these elements unambiguous. revision: yes

  2. Referee: [Model mechanisms] Model mechanisms (§3): the assertion that orthographic processes, efficiency gains assessment, and personal preference on AI support are sufficient to capture high-level decision-making lacks ablation testing. If the hierarchical RL extension can reproduce the data through reward shaping or other parameters alone, the cognitive grounding and consequent generalization to UI/algorithm changes would be weakened.

    Authors: We agree that explicit ablation would strengthen the cognitive claims. The three mechanisms are motivated by established cognitive models of typing and decision-making (cited in §3), and the full model is required to match the observed diversity of strategies and individual differences. To address the concern directly, we will add ablation experiments in a new subsection of §4, systematically disabling each mechanism (e.g., removing orthographic simulation or preference weighting) and demonstrating degraded fits to human data. This will support that the mechanisms, rather than generic reward shaping, drive the results and enable the what-if design analyses. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper constructs WSTypist by extending prior hierarchical control models of typing with three explicitly identified cognitive mechanisms (orthographic processes, efficiency gains assessment, and personal preference on AI support) implemented within a reinforcement learning framework. Evaluations then test whether the resulting model reproduces observed human strategies, individual differences, and generalization across systems, followed by what-if simulations on design cases. No equations, parameter-fitting steps, or self-citations are shown to reduce the central outputs to the inputs by construction; the mechanisms are added as independent modeling choices rather than being defined in terms of the target behaviors they are meant to explain. The derivation therefore remains self-contained and externally falsifiable against human data.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on extending prior hierarchical typing models with three domain-specific cognitive mechanisms whose sufficiency is assumed rather than derived; no free parameters or invented entities are explicitly named in the abstract, but reward weights in the RL component are implied to be tuned to human data.

free parameters (1)
  • RL reward weights for efficiency and preference
    Must be set or fitted to reproduce human suggestion-use strategies and individual differences.
axioms (1)
  • domain assumption Users integrate word suggestions via orthographic processes, efficiency gain assessment, and personal preference on AI support
    These three mechanisms are identified and implemented as the key additions to the hierarchical control model.

pith-pipeline@v0.9.0 · 5474 in / 1322 out tokens · 35918 ms · 2026-05-16T07:03:02.360194+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

81 extracted references · 81 canonical work pages · 3 internal anchors

  1. [1]

    Ohoud Alharbi, Wolfgang Stuerzlinger, and Felix Putze. 2020. The effects of predictive features of mobile keyboards on text entry speed and errors. Proceedings of the ACM on Human-Computer Interaction4, ISS (2020), 1–16. https://doi.org/10.1145/3427311

    work page doi:10.1145/3427311 2020

  2. [2]

    Anna Anastaseni, Quentin Roy, Cyril Perret, Antonio Romano, and Sonia Kandel. 2025. What smartphones change about writing: The impact of word suggestions on orthographic processing.Cognitive Neuropsychology(2025), 1–26

    work page 2025

  3. [3]

    John R Anderson, Daniel Bothell, Michael D Byrne, Scott Douglass, Christian Lebiere, and Yulin Qin. 2004. An integrated theory of the mind. Psychological review111, 4 (2004), 1036

    work page 2004

  4. [4]

    Alan Baddeley. 2012. Working memory: Theories, models, and controversies.Annual review of psychology63, 1 (2012), 1–29

    work page 2012

  5. [5]

    Nikola Banovic, Ticha Sethapakdi, Yasasvi Hari, Anind K Dey, and Jennifer Mankoff. 2019. The limits of expert text entry speed on mobile keyboards with autocorrect. InProceedings of the 21st International Conference on Human-Computer Interaction with Mobile Devices and Services. 1–12. https://doi.org/10.1145/3338286.3340126

    work page doi:10.1145/3338286.3340126 2019

  6. [6]

    Yoshua Bengio, Jérôme Louradour, Ronan Collobert, and Jason Weston. 2009. Curriculum learning. InProceedings of the 26th annual international conference on machine learning. 41–48

    work page 2009

  7. [7]

    Anol Bhattacherjee. 2001. Understanding information systems continuance: an expectation-confirmation model.MIS Q.25, 3 (Sept. 2001), 351–370. doi:10.2307/3250921

    work page doi:10.2307/3250921 2001

  8. [8]

    Marcel Binz and Eric Schulz. 2023. Turning large language models into cognitive models.arXiv preprint arXiv:2306.03917(2023)

    work page arXiv 2023

  9. [9]

    Daniel Buschek. 2024. Collage is the New Writing: Exploring the Fragmentation of Text and User Interfaces in AI Tools. InProceedings of the 2024 ACM Designing Interactive Systems Conference (DIS ’24). 2719–2737. https://doi.org/10.1145/3643834.3660681

    work page doi:10.1145/3643834.3660681 2024

  10. [10]

    Daniel Buschek, Martin Zürn, and Malin Eiband. 2021. The Impact of Multiple Parallel Phrase Suggestions on Email Input and Composition Behaviour of Native and Non-Native English Writers. InProceedings of the 2021 CHI Conference on Human Factors in Computing Systems(Yokohama, Japan)(CHI ’21). Association for Computing Machinery, New York, NY, USA, Article ...

    work page doi:10.1145/3411764.3445372 2021

  11. [11]

    Suyog Chandramouli, Danqing Shi, Aini Putkonen, Sebastiaan De Peuter, Shanshan Zhang, Jussi Jokinen, Andrew Howes, and Antti Oulasvirta

    work page

  12. [12]

    https: //doi.org/10.1007/s42113-024-00208-6

    A Workflow for Building Computationally Rational Models of Human Behavior.Computational Brain & Behavior(2024), 1–21. https: //doi.org/10.1007/s42113-024-00208-6

    work page doi:10.1007/s42113-024-00208-6 2024

  13. [13]

    Mia Xu Chen, Benjamin N Lee, Gagan Bansal, Yuan Cao, Shuyuan Zhang, Justin Lu, Jackie Tsay, Yinan Wang, Andrew M Dai, Zhifeng Chen, et al

    work page

  14. [14]

    InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining

    Gmail smart compose: Real-time assisted writing. InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. 2287–2295

    work page

  15. [15]

    John J Darragh, Ian H Witten, and Mark L. James. 1990. The reactive keyboard: A predictive typing aid.Computer23, 11 (1990), 41–49

    work page 1990

  16. [16]

    François Delebecque, Claude Gomez, Maurice Goursat, Ramine Nikoukhah, Serge Steer, and Jean-Philippe Chancelier. 1994. Scilab. https: //www.scilab.org/

    work page 1994

  17. [17]

    2018.Assignment Problems for Optimizing Text Input

    Anna Maria Feit. 2018.Assignment Problems for Optimizing Text Input. Ph. D. Dissertation. Aalto University. http://urn.fi/URN:ISBN:978-952-60- 8016-1

    work page 2018

  18. [18]

    Anna Maria Feit and Antti Oulasvirta. 2014. Pianotext: Redesigning the piano keyboard for text entry. InProceedings of the 2014 conference on Designing interactive systems. 1045–1054

    work page 2014

  19. [19]

    Anna Maria Feit, Daryl Weir, and Antti Oulasvirta. 2016. How We Type: Movement Strategies and Performance in Everyday Typing. InProceedings of the 2016 CHI Conference on Human Factors in Computing Systems(San Jose, California, USA)(CHI ’16). Association for Computing Machinery, New York, NY, USA, 4262–4273. doi:10.1145/2858036.2858233

    work page doi:10.1145/2858036.2858233 2016

  20. [20]

    Andrew Fowler, Kurt Partridge, Ciprian Chelba, Xiaojun Bi, Tom Ouyang, and Shumin Zhai. 2015. Effects of language modeling and its personalization on touchscreen typing performance. InProceedings of the 33rd annual ACM conference on human factors in computing systems. 649–658. https: //doi.org/10.1145/2702123.2702503

    work page doi:10.1145/2702123.2702503 2015

  21. [21]

    Nestor Garay-Vitoria and Julio Abascal. 2006. Text prediction systems: a survey.Universal Access in the Information Society4 (2006), 188–203. doi:10.1007/s10209-005-0005-9

    work page doi:10.1007/s10209-005-0005-9 2006

  22. [22]

    Nestor Garay-Vitoria and Julio Gonzalez-Abascal. 1997. Intelligent word-prediction to enhance text input rate (a syntactic analysis-based word- prediction aid for people with severe motor and speech disability). InProceedings of the 2nd international conference on Intelligent user interfaces. 241–244

    work page 1997

  23. [23]

    Shaona Ghosh and Per Ola Kristensson. 2017. Neural networks for text correction and completion in keyboard decoding.arXiv preprint arXiv:1709.06429(2017). https://doi.org/10.48550/arXiv.1709.06429

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1709.06429 2017

  24. [24]

    Steven M Goodman, Erin Buehler, Patrick Clary, Andy Coenen, Aaron Donsbach, Tiffanie N Horne, Michal Lahav, Robert MacDonald, Rain Breaw Michaels, Ajit Narayanan, et al. 2024. LaMPost: AI writing assistance for adults with dyslexia using large language models.Commun. ACM67, 9 (2024), 80–89

    work page 2024

  25. [25]

    Yves Guiard and Olivier Rioul. 2015. A mathematical description of the speed/accuracy trade-off of aimed movement. InProceedings of the 2015 British HCI Conference. 91–100

    work page 2015

  26. [26]

    Andrew Hard, Kanishka Rao, Rajiv Mathews, Swaroop Ramaswamy, Françoise Beaufays, Sean Augenstein, Hubert Eichner, Chloé Kiddon, and Daniel Ramage. 2018. Federated learning for mobile keyboard prediction.arXiv preprint arXiv:1811.03604(2018). https://doi.org/10.48550/arXiv.1811.03604 Manuscript submitted to ACM Simulating Word Suggestion Usage 19

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1811.03604 2018

  27. [27]

    John R. Hayes. 2012. Modeling and Remodeling Writing.Written Communication29, 3 (2012), 369–388. arXiv:https://doi.org/10.1177/0741088312451260 doi:10.1177/0741088312451260

    work page doi:10.1177/0741088312451260 2012

  28. [28]

    Morten Hertzum and Kasper Hornbæk. 2023. Frustration: Still a Common User Experience.ACM Trans. Comput.-Hum. Interact.30, 3, Article 42 (June 2023), 26 pages. doi:10.1145/3582432

    work page doi:10.1145/3582432 2023

  29. [29]

    Lorenz Hetzel, John Dudley, Anna Maria Feit, and Per Ola Kristensson. 2021. Complex Interaction as Emergent Behaviour: Simulating Mid-Air Virtual Keyboard Typing using Reinforcement Learning.IEEE Transactions on Visualization and Computer Graphics27, 11 (2021), 4140–4149. doi:10.1109/TVCG.2021.3106494

    work page doi:10.1109/tvcg.2021.3106494 2021

  30. [30]

    Sepp Hochreiter and Jürgen Schmidhuber. 1997. Long short-term memory.Neural computation9, 8 (1997), 1735–1780

    work page 1997

  31. [31]

    SB Hutton. 2019. Eye tracking methodology. InEye movement research: An introduction to its scientific foundations and applications. Springer, 277–308

    work page 2019

  32. [32]

    Xinhui Jiang, Yang Li, Jussi PP Jokinen, Viet Ba Hirvola, Antti Oulasvirta, and Xiangshi Ren. 2020. How we type: Eye and finger movement strategies in mobile typing. InProceedings of the 2020 CHI conference on human factors in computing systems. 1–14. https://doi.org/10.1145/3313831.3376711

    work page doi:10.1145/3313831.3376711 2020

  33. [33]

    Jussi Jokinen, Aditya Acharya, Mohammad Uzair, Xinhui Jiang, and Antti Oulasvirta. 2021. Touchscreen typing as optimal supervisory control. In Proceedings of the 2021 CHI conference on human factors in computing systems. 1–14. https://doi.org/10.1145/3411764.3445483

    work page doi:10.1145/3411764.3445483 2021

  34. [34]

    Jussi PP Jokinen, Sayan Sarcar, Antti Oulasvirta, Chaklam Silpasuwanchai, Zhenxin Wang, and Xiangshi Ren. 2017. Modelling learning of new keyboard layouts. InProceedings of the 2017 CHI conference on human factors in computing systems. 4203–4215. https://doi.org/10.1145/3025453.3025580

    work page doi:10.1145/3025453.3025580 2017

  35. [35]

    Leslie Pack Kaelbling, Michael L Littman, and Anthony R Cassandra. 1998. Planning and acting in partially observable stochastic domains.Artificial intelligence101, 1-2 (1998), 99–134

    work page 1998

  36. [36]

    Sonia Kandel. 2023. Written production: The APOMI model of word writing: Anticipatory processing of orthographic and motor information. In Language production. Routledge, 209–232

    work page 2023

  37. [37]

    Davis E Kieras and Davis E Meyer. 1997. An overview of the EPIC architecture for cognition and performance with application to human-computer interaction.Human–Computer Interaction12, 4 (1997), 391–438

    work page 1997

  38. [38]

    Heidi Horstmann Koester and Simon Levine. 1996. Effect of a word prediction feature on user performance.Augmentative and alternative communication12, 3 (1996), 155–168

    work page 1996

  39. [39]

    Heidi Horstmann Koester and Simon Levine. 1997. Keystroke-level models for user performance with word prediction.Augmentative and Alternative Communication13, 4 (1997), 239–257

    work page 1997

  40. [40]

    Heidi Horstmann Koester and Simon Levine. 1998. Model simulations of user performance with word prediction.Augmentative and Alternative Communication14, 1 (1998), 25–36

    work page 1998

  41. [41]

    Heidi Horstmann Koester and Simon P Levine. 2002. Modeling the speed of text entry with a word prediction interface.IEEE transactions on rehabilitation engineering2, 3 (2002), 177–187

    work page 2002

  42. [42]

    Per Ola Kristensson. 2009. Five challenges for intelligent text entry methods.AI Magazine30, 4 (2009), 85–85. https://doi.org/10.1609/aimag.v30i4.2269

    work page doi:10.1609/aimag.v30i4.2269 2009

  43. [43]

    Per Ola Kristensson. 2018. Statistical Language Processing for Text Entry.Computational Interaction(2018), 43–64. https://doi.org/10.1093/oso/ 9780198799603.003.0003

    work page doi:10.1093/oso/ 2018

  44. [44]

    Per Ola Kristensson and Thomas Müllners. 2021. Design and analysis of intelligent text entry systems with function structure models and envelope analysis. InProceedings of the 2021 CHI Conference on Human Factors in Computing Systems. 1–12. https://doi.org/10.1145/3411764.3445566

    work page doi:10.1145/3411764.3445566 2021

  45. [45]

    Per-Ola Kristensson and Shumin Zhai. 2004. SHARK2: a large vocabulary shorthand writing system for pen-based computers. InProceedings of the 17th Annual ACM Symposium on User Interface Software and Technology (UIST ’04). 43–52. doi:10.1145/1029632.1029640

    work page doi:10.1145/1029632.1029640 2004

  46. [46]

    Po-Ming Lee, Wei-Hsuan Tsui, and Tzu-Chien Hsiao. 2014. The influence of emotion on keyboard typing: an experimental study using visual stimuli. Biomedical engineering online13, 1 (2014), 81

    work page 2014

  47. [47]

    Po-Ming Lee, Wei-Hsuan Tsui, and Tzu-Chien Hsiao. 2015. The influence of emotion on keyboard typing: An experimental study using auditory stimuli.PloS one10, 6 (2015), e0129056

    work page 2015

  48. [48]

    Florian Lehmann, Itto Kornecki, Daniel Buschek, and Anna Maria Feit. 2023. Typing Behavior is About More than Speed: Users’ Strategies for Choosing Word Suggestions Despite Slower Typing Rates.Proceedings of the ACM on Human-Computer Interaction7, MHCI (2023), 1–26. https://doi.org/10.1145/3604276

    work page doi:10.1145/3604276 2023

  49. [49]

    Katri Leino, Markku Laine, Mikko Kurimo, and Antti Oulasvirta. 2024. Mobile Typing with Intelligent Text Entry: A Large-Scale Dataset and Results. Preprint(2024). https://doi.org/10.21203/rs.3.rs-4654512/v1

    work page doi:10.21203/rs.3.rs-4654512/v1 2024

  50. [50]

    Luis A Leiva, Sunjun Kim, Wenzhe Cui, Xiaojun Bi, and Antti Oulasvirta. 2021. How we swipe: A large-scale shape-writing dataset and empirical findings. InProceedings of the 23rd International Conference on Mobile Human-Computer Interaction. 1–13. https://doi.org/10.1145/3447526.3472059

    work page doi:10.1145/3447526.3472059 2021

  51. [51]

    Yang Li and Anna Maria Feit. 2025. How We Type with Word Suggestions: Understanding Visual Attention and Checking Behavior during Mobile Text Input.Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies9, 3 (2025), 1–29

    work page 2025

  52. [52]

    Jianhua Lin. 2002. Divergence measures based on the Shannon entropy.IEEE Transactions on Information theory37, 1 (2002), 145–151

    work page 2002

  53. [53]

    Susan Lin, Jeremy Warner, JD Zamfirescu-Pereira, Matthew G Lee, Sauhard Jain, Shanqing Cai, Piyawat Lertvittayakumjorn, Michael Xuelin Huang, Shumin Zhai, Bjoern Hartmann, et al. 2024. Rambler: Supporting Writing With Speech via LLM-Assisted Gist Manipulation. InProceedings of the 2024 CHI Conference on Human Factors in Computing Systems. 1–19

    work page 2024

  54. [54]

    Gordon D Logan. 2018. Automatic control: How experts act without thinking.Psychological Review125, 4 (2018), 453. Manuscript submitted to ACM 20 Li and Feit

    work page 2018

  55. [55]

    I Scott MacKenzie and R William Soukoreff. 2003. Phrase sets for evaluating text entry techniques. InCHI’03 extended abstracts on Human factors in computing systems. 754–755

    work page 2003

  56. [56]

    Scott MacKenzie and Shawn X

    I. Scott MacKenzie and Shawn X. Zhang. 1999. The design and evaluation of a high-performance soft keyboard. InProceedings of the SIGCHI Conference on Human Factors in Computing Systems(Pittsburgh, Pennsylvania, USA)(CHI ’99). Association for Computing Machinery, New York, NY, USA, 25–31. doi:10.1145/302979.302983

    work page doi:10.1145/302979.302983 1999

  57. [57]

    I Scott MacKenzie and Shawn X Zhang. 1999. The design and evaluation of a high-performance soft keyboard. InProceedings of the SIGCHI conference on Human Factors in Computing Systems. 25–31

    work page 1999

  58. [58]

    Roderick Murray-Smith, Antti Oulasvirta, Andrew Howes, Jörg Müller, Aleksi Ikkala, Miroslav Bachinski, Arthur Fleig, Florian Fischer, and Markus Klar. 2022. What simulation can do for HCI research.Interactions29, 6 (2022), 48–53

    work page 2022

  59. [59]

    Thuy Ngoc Nguyen, Kasturi Jamale, and Cleotilde Gonzalez. 2024. Predicting and understanding human action decisions: Insights from large language models and cognitive instance-based learning. InProceedings of the AAAI Conference on Human Computation and Crowdsourcing, Vol. 12. 126–136

    work page 2024

  60. [60]

    Antti Oulasvirta, Jussi P. P. Jokinen, and Andrew Howes. 2022. Computational Rationality as a Theory of Interaction. InProceedings of the 2022 CHI Conference on Human Factors in Computing Systems (CHI ’22). Article 359, 14 pages. doi:10.1145/3491102.3517739

    work page doi:10.1145/3491102.3517739 2022

  61. [61]

    Antti Oulasvirta, Anna Reichel, Wenbin Li, Yan Zhang, Myroslav Bachynskyi, Keith Vertanen, and Per Ola Kristensson. 2013. Improving two-thumb text entry on touchscreen devices. InProceedings of the SIGCHI Conference on Human Factors in Computing Systems. 2765–2774

    work page 2013

  62. [62]

    Kseniia Palin, Anna Maria Feit, Sunjun Kim, Per Ola Kristensson, and Antti Oulasvirta. 2019. How do people type on mobile devices? Observations from a study with 37,000 volunteers. InProceedings of the 21st International Conference on Human-Computer Interaction with Mobile Devices and Services. 1–12. https://doi.org/10.1145/3338286.3340120

    work page doi:10.1145/3338286.3340120 2019

  63. [63]

    Parasuraman

    A. Parasuraman. 2000. Technology Readiness Index (Tri): A Multiple-Item Scale to Measure Readiness to Embrace New Technologies.Journal of Service Research2, 4 (2000), 307–320. arXiv:https://doi.org/10.1177/109467050024001 doi:10.1177/109467050024001

    work page doi:10.1177/109467050024001 2000

  64. [64]

    Shubham Pateria, Budhitama Subagdja, Ah-hwee Tan, and Chai Quek. 2021. Hierarchical reinforcement learning: A comprehensive survey.ACM Computing Surveys (CSUR)54, 5 (2021), 1–35

    work page 2021

  65. [65]

    Svetlana Pinet, Johannes C Ziegler, and F-Xavier Alario. 2016. Typing is writing: Linguistic properties modulate typing execution.Psychonomic bulletin & review23, 6 (2016), 1898–1906

    work page 2016

  66. [66]

    Philip Quinn and Shumin Zhai. 2016. A cost-benefit study of text entry suggestion interaction. InProceedings of the 2016 CHI conference on human factors in computing systems. 83–88. https://doi.org/10.1145/2858036.2858305

    work page doi:10.1145/2858036.2858305 2016

  67. [67]

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. OpenAI blog1, 8 (2019), 9

    work page 2019

  68. [68]

    Evan F Risko and Sam J Gilbert. 2016. Cognitive offloading.Trends in cognitive sciences20, 9 (2016), 676–688

    work page 2016

  69. [69]

    Quentin Roy, Sébastien Berlioux, Géry Casiez, and Daniel Vogel. 2021. Typing efficiency and suggestion accuracy influence the benefits and adoption of word suggestions. InProceedings of the 2021 CHI Conference on Human Factors in Computing Systems. 1–13. https://doi.org/10.1145/3411764.3445725

    work page doi:10.1145/3411764.3445725 2021

  70. [70]

    Quentin Roy, Géry Casiez, and Daniel Vogel. 2025. Are word suggestions beneficial? The effect of typing efficiency and suggestion accuracy.ACM Transactions on Computer-Human Interaction(2025)

    work page 2025

  71. [71]

    Dario D Salvucci. 2001. An integrated model of eye movements and visual encoding.Cognitive Systems Research1, 4 (2001), 201–220

    work page 2001

  72. [72]

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. 2017. Proximal policy optimization algorithms.arXiv preprint arXiv:1707.06347(2017)

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  73. [73]

    Danqing Shi, Yao Wang, Yunpeng Bai, Andreas Bulling, and Antti Oulasvirta. 2025. Chartist: Task-driven Eye Movement Control for Chart Reading. InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems. 1–14

    work page 2025

  74. [74]

    Danqing Shi, Yujun Zhu, Francisco Erivaldo Fernandes Junior, Shumin Zhai, and Antti Oulasvirta. 2025. Simulating Errors in Touchscreen Typing. InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems. 1–13

    work page 2025

  75. [75]

    Danqing Shi, Yujun Zhu, Jussi PP Jokinen, Aditya Acharya, Aini Putkonen, Shumin Zhai, and Antti Oulasvirta. 2024. CRTypist: Simulating Touchscreen Typing Behavior via Computational Rationality. InProceedings of the CHI Conference on Human Factors in Computing Systems. 1–17. https://doi.org/10.1145/3613904.3642918

    work page doi:10.1145/3613904.3642918 2024

  76. [76]

    1998.Reinforcement learning: An introduction

    Richard S Sutton, Andrew G Barto, et al. 1998.Reinforcement learning: An introduction. Vol. 1. MIT press Cambridge

    work page 1998

  77. [77]

    Diana-Cezara Toader, Grat,iela Boca, Rita Toader, Mara Măcelaru, Cezar Toader, Diana Ighian, and Adrian T Rădulescu. 2019. The effect of social presence and chatbot errors on trust.Sustainability12, 1 (2019), 256

    work page 2019

  78. [78]

    2022.Bayesian Methods for Interaction and Design

    John H Williamson, Antti Oulasvirta, Per Ola Kristensson, and Nikola Banovic. 2022.Bayesian Methods for Interaction and Design. Cambridge University Press

    work page 2022

  79. [79]

    Seunghak Yu, Nilesh Kulkarni, Haejun Lee, and Jihie Kim. 2018. On-device neural language model based word prediction. InProceedings of the 27th international conference on computational linguistics: system demonstrations. 128–131. https://aclanthology.org/C18-2028

    work page 2018

  80. [80]

    Mingrui Ray Zhang and Shumin Zhai. 2021. PhraseFlow: Designs and empirical studies of phrase-level input. InProceedings of the 2021 CHI Conference on Human Factors in Computing Systems. 1–13

    work page 2021

Showing first 80 references.