Pith. sign in

REVIEW 4 major objections 6 minor 83 references

Fuzzy Linkography: Automatic Graphical Summarization of Creative Activity Traces

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

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

Pith's one-line read This paper claims that linkographs—diagrams of how design moves build on one another—can be constructed automatically by scoring semantic similarity between move texts with sentence embeddings, and that the resulting fuzzy graphs preserve…

desk verdict Fuzzy linkography is a promising, clearly-presented method for cheap linkographs, but its load-bearing premise—embedding similarity as a stand-in for human link annotation—is untested, and the paper's own limitations section admits as much. read the letter →

arxiv 2502.04599 v2 pith:IUQIS4B4 submitted 2025-02-07 cs.HC

Amy Smith , Barrett R. Anderson , Jasmine Tan Otto , Isaac Karth , Yuqian Sun , John Joon Young Chung , Melissa Roemmele , Max Kreminski This is my paper · ORCID
classification cs.HC
keywords fuzzylinkographylinkographcreativitysupporttoolsdesignmovessemanticsimilaritysentenceembeddingscreativeactivitytracesvisualization
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

This paper proposes fuzzy linkography, a way to build a linkograph automatically from a sequence of recorded design moves. Linkography is a method, dating to the early 1990s, that visualizes a creative episode as a row of design moves with lines drawn between moves that build on one another; until now it has required a human coder to annotate both moves and links, which has kept it from being applied at scale. The authors' method replaces the human coder with a sentence-embedding model: each move's text is turned into a vector, cosine similarity between pairs of moves sets the strength of a link from 0 to 1, and links below a threshold are discarded. They apply the method to text-to-image prompting logs, LLM-supported ideation sessions, and researcher publication histories, and argue that the resulting imperfect graphs still function as graphical summaries that reveal recurring structural patterns and support quantitative analysis. The payoff, if the method holds up, is that linkographic analysis becomes cheap enough to run over thousands of traces instead of a handful.

What carries the argument

The central object is the fuzzy linkograph, a graph in which design moves are nodes laid out left to right and links carry weights between 0 and 1 instead of binary on/off values. The inference machinery is cosine similarity between sentence embeddings of move texts: the all-MiniLM-L6-v2 model converts each move's text into a vector, and for every pair of moves the cosine similarity is computed; values above the threshold $t = 0.35$ are linearly rescaled onto $[0,1]$ and rendered as varying link darkness. The traditional linkographic statistics are then redefined on the weighted graph—forelink and backlink weights as sums of link strengths, link density index as total link strength divided by move count, and link entropy by treating each link strength as the probability that a binary link exists, following the standard derivation in [38]. This object carries the argument because every claimed finding, from refinement webs to critical moves to cluster archetypes, is read off this automatically constructed weighted graph.

What would settle it

Take a sample of design episodes, have several human coders independently annotate links between moves, and compute the correlation between the human link judgments and the fuzzy link strengths produced by all-MiniLM-L6-v2 with $t = 0.35$. If the correlation is at or near chance, or if the paper's own failure example—no link between "not much money in the kitty" and "piggy bank"—turns out to generalize to many obvious semantic relations, the central claim fails; a positive correlation would support it. Running the same analysis across a range of thresholds would also show whether the reported motifs are an artifact of the chosen threshold value.

Watch

Extended reading notes

Core claim

The paper claims that a linkograph—the network of design moves and the semantic links between them—can be produced by a purely computational pipeline and still carry the analytical value linkography has when built by hand. Formally, each recorded move is represented by a sentence embedding, every pair of moves receives a cosine-similarity score, scores above a fixed threshold $t = 0.35$ are linearly rescaled to link strengths in $[0,1]$, and the resulting graph is drawn with link darkness encoding strength. On this basis the authors report that familiar manual-linkography phenomena (webs, chunks, sawtooths, critical moves, link entropy) survive the automation, and they identify new motifs in image prompting traces (refinement webs, curiosity zigzags, converging zigzags), asymmetric human-to-machine backlink densities in LLM-supported ideation, and recognizable career shapes in publication histories. The intended contribution is not a perfect replica of human annotation but a cheap, approximate, continuously valued summary that scales to thousands of traces.

Load-bearing premise

The method rests on one premise: cosine similarity between sentence embeddings of move texts is a valid stand-in for a human coder's judgment that two design moves are related. The authors do not test this directly—the threshold $0.35$ is chosen because it "seems to work well," and Section 8 asks for a formal comparison between human-constructed and machine-constructed linkographs—so if embedding similarity diverges from human perception of relatedness, the inferred links and all patterns built on them lose their grounding.

Editorial extensions

If this is right

  • Linkographic analysis moves from a scarce, labor-intensive research method to a routine computation: any text-logged creative activity trace can be turned into a linkograph at near-zero marginal cost.
  • At scale, linkographic hypotheses—such as which structural patterns or entropy values accompany successful ideation—can be tested by comparing automatically built linkographs against subjective creativity ratings, product quality, or other outcome measures.
  • Real-time construction of fuzzy linkographs becomes possible, allowing reflective visualization in which participants watch their own creative process take shape while it is still happening.
  • Creativity support tools could use linkographic metrics to guide their own behavior, for example gauging creative momentum, deciding when to intervene, or finding previously unexplored moves to build on.
  • Domain-specific results become visible for the first time at scale, such as refinement webs and curiosity zigzags in text-to-image prompting and asymmetric human-machine influence patterns in LLM-supported ideation.

Reading between the lines

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

  • The paper stops short of claiming that fuzzy linkographs can replace human annotation; a natural extension it does not defend is that these graphs are most reliable at the level of whole-episode motifs and statistics rather than individual links, so validity should be sought in how well motif frequencies track outcomes, not in per-link accuracy.
  • If the threshold $t = 0.35$ and the choice of embedding model were varied, the resulting linkographs would change; nothing in the paper establishes that the reported motifs are robust to these choices, so a direct sensitivity analysis is a testable next step.
  • The same pipeline applies to any artifact that can be embedded—images, music, user interfaces, game states—so the method's real reach may be as a domain-agnostic telemetry layer for creative tools, with linkographic abundance becoming a routine dashboard rather than an analytical endpoint.
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 introduces 'fuzzy linkography,' a pipeline that automatically converts sequences of textual design moves into linkographs by computing cosine similarity between sentence embeddings, thresholding at t=0.35, and rescaling the remaining similarities into link strengths. It defines fuzzy analogues of forelink/backlink weights, link density, and link entropy, and applies the method to three domains: text-to-image prompting traces, LLM-supported ideation sessions, and researcher publication histories. The case studies yield recurring motifs, interaction-pattern claims, and career-shape observations, and the authors release their code open-source. The paper's central methodological premise is that embedding similarity can stand in for human link annotation; the manuscript does not currently validate that premise, and the authors explicitly defer such validation to future work.

Significance. If validated, the contribution is genuinely useful: it would lower the cost of linkographic analysis, enable large-scale exploratory studies, and support real-time reflective visualization. The open-source implementation, the diversity of the three case studies, the clear operationalization of fuzzy link metrics, and the candid limitations section are all strengths. However, the central premise that cosine similarity between MiniLM embeddings is a valid proxy for human judgments of move relatedness is unvalidated, and several specific downstream claims rest on ad hoc parameter choices and visual inspection. As it stands, the paper is better characterized as an exploratory visualization method for embedding-similarity structure than as a validated automatic replacement for manual linkography.

major comments (4)
  1. [Section 3 and Section 8] The load-bearing premise of the paper is that cosine similarity between design-move embeddings can stand in for human link annotation. This premise is never tested: the threshold t=0.35 is chosen because it 'seems to work well' (Section 3), and Section 8 explicitly defers 'formal comparison between human-constructed and machine-constructed linkographs' to future work. Since every quantitative measure, motif, and interpretation in the paper is a function of these link strengths, the absence of validation is a major gap. Please add a validation experiment in which human annotators create linkographs for a sample of the same move sequences (ideally with multiple annotators to measure inter-rater reliability), and compare the machine linkographs with human linkographs using threshold-independent metrics such as rank correlation or AUC for link existence. Also report sensitivity of the main conclusions to the choice of t. If such validation is considered out of scope, the paper should be explicitly reframed as an exploration of embedding-similarity structure rather than as automatic linkography.
  2. [Section 4.1] The recurring-motif taxonomy (refinement webs, curiosity zigzags, converging zigzags, temporal structures) is derived by 'visually inspecting the linkographs of each trace' with no inter-rater reliability, no operational definitions of the motifs, and no test that the motifs occur more often than would be expected by chance. These motifs are the main empirical contribution of the first case study, but because they are extracted from the very linkographs produced by the unvalidated model, they cannot independently confirm the method's usefulness. The motifs should either be coded by multiple independent raters with reliability statistics, or explicitly presented as hypotheses for future confirmation.
  3. [Section 4.2] The trace-clustering analysis uses k=5 and a z-score outlier cutoff of 3 with no reported justification, and the cluster archetype descriptions (e.g., 'members of cluster 4 ... single major fascination') appear to be assigned by inspecting example linkographs. No cluster-validation metrics, stability analysis, or sensitivity to k, threshold t, or the embedding model are reported. Consequently, the design implications drawn from the clusters (e.g., scaffolding for goal formation, per-archetype UI adaptation) are not supported. Please add quantitative cluster validation and sensitivity checks, or restrict the claims to descriptive observations.
  4. [Section 6] The publication-history case study reports that automatically identified critical moves are 'roughly half' the time sensible to the authors, which is a high failure rate for a core linkographic quantity and is not evaluated formally. The top-three-by-weight selection rule is also arbitrary. The claim that fuzzy linkography can identify critical moves in careers needs a systematic evaluation against human judgments (with precision and recall, or a ranked-ground-truth comparison); otherwise this part of the analysis should be labeled a negative result and not listed among the successful applications of the method.
minor comments (6)
  1. [Section 5.1, Table 1] The sentence '51 of 72 total traces (70.83%) feature a greater total backlink density from machine moves to human moves than vice versa' contradicts Table 1 and the following sentence, which states that the average density of backlinks from human to machine moves is nearly 2x that of machine to human moves. Please correct the direction or the table.
  2. [Table 1] The 'Machine to...' row appears to have only two numerical entries for three columns; please clarify whether the Machine (YC) entry is identical to the Machine (NC) entry or was omitted.
  3. [Section 4.1] The definitions of a 'substantial' trace (at least seven prompts) and a 'session break' (at least 30 minutes) are reasonable, but the paper does not report whether the identified motifs are robust to these choices; a brief sensitivity note would strengthen the results.
  4. [Section 6] The publication-history sample consists of 10 researchers well-known to the authors, which is a convenience sample with possible selection bias; this should be stated explicitly when interpreting the generality of the career shapes.
  5. [Section 3] The claim that all-MiniLM-L6-v2 'has previously been validated against a human baseline for assessment of semantic similarity' cites [3], but that prior validation concerned ideation-output similarity, not pairwise link judgments; this should not be presented as validation of the link-inference threshold.
  6. [Section 8] The missed-link example of 'not much money in the kitty' and 'piggy bank' is useful; connecting it explicitly to the cited part-of-speech bias in sentence transformers [59] would make the limitation paragraph more precise.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: fuzzy linkographs are explicitly constructed from embedding similarity, and the paper's interpretive claims are exploratory with stated limitations; validity concerns are not circularity.

full rationale

The paper's derivation chain is self-contained. Link strengths are explicitly defined as rescaled cosine similarities between sentence embeddings (Section 3, "we use cosine similarity between the embedding vectors representing each move to determine the strength of the link between these moves"), so every linkograph, downstream statistic, and motif is a deterministic function of the input text. That is the proposed method, not a hidden circular reduction. The interpretive findings—refinement webs, curiosity zigzags, inter-actor influence asymmetries, and career shapes—are presented as exploratory observations from the constructed graphs, with explicit caveats that the graphs are imperfect proxies and that "formal comparison between human-constructed and machine-constructed linkographs" remains future work (Section 8). The threshold t=0.35 is an openly described heuristic ("this value seems to work well with our chosen embedding model"), not a parameter fitted to a target and then renamed as a prediction. The one notable self-citation, [3] (Anderson, Shah, and Kreminski, with two overlapping authors), is used to support the embedding model's prior validation against a human baseline for semantic similarity in ideation; this is external, falsifiable evidence and is not the exclusive load-bearing basis of the central claim. The Section 6 result that automatic critical-move identification is sensible only "roughly half the time" is an honest negative result, not a circular confirmation. Concerns about whether cosine similarity adequately captures human link perception are construct-validity and correctness risks, not circularity.

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

The approach rests on several unvalidated domain assumptions about embedding similarity as a proxy for human link judgments, plus multiple hand-picked parameters (t, k, thresholds). No new physical entities are introduced.

free parameters (5)
  • similarity_threshold_t = 0.35
    Chosen ad hoc because it 'seems to work well' with the embedding model; no principled justification or sensitivity analysis. Section 3.
  • cluster_count_k = 5
    Arbitrary choice for k-means clustering of trace signature vectors; no validation of cluster number. Section 4.2.
  • minimum_trace_length = 7 moves
    Filtering criterion for 'substantial' image prompting traces; affects dataset composition. Section 4.
  • outlier_z_threshold = 3
    Used to exclude 59 traces before clustering; arbitrary selection without justification. Section 4.2.
  • critical_move_selection = top 3 by forelink/backlink weight
    Arbitrary rule for identifying critical moves in publication histories; no principled basis. Section 6.
assumptions (5)
  • domain assumption Textual descriptions of design moves are an adequate representation for linkographic analysis.
    The method embeds move text, so its validity depends on the assumption that move text carries the semantic content needed to determine relatedness. Invoked throughout, e.g., Section 3.
  • domain assumption Cosine similarity in the embedding space is a valid proxy for human judgments of move relatedness.
    The entire link inference relies on this; the paper does not validate it against human linkographs. Sections 3 and 8.
  • domain assumption The all-MiniLM-L6-v2 embedding model is appropriate for all three domains.
    The model was validated for ideation in prior work [3], but not for image prompts or publication titles. Section 3.
  • domain assumption Link strength can be treated as the probability of a binary link for entropy calculation.
    The paper adapts Kan and Gero's entropy formulation by substituting summed strengths for binary counts in Eq. (1). The authors note this conflates association strength with probability. Section 3.
  • domain assumption The standard linkographic pattern taxonomy (webs, chunks, sawtooths) applies to automatically generated linkographs.
    The paper interprets patterns in fuzzy linkographs using categories from manual linkography, implicitly assuming the same structures are meaningful when produced by embeddings. Section 4.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fuzzy Linkography: Automatic Graphical Summarization of Creative Activity Traces." pith.science (2026). https://pith.science/paper/IUQIS4B4

@misc{pith2026250204599,
  author       = {Pith},
  title        = {Pith review of: Fuzzy Linkography: Automatic Graphical Summarization of Creative Activity Traces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IUQIS4B4}},
  note         = {Machine review of arXiv:2502.04599}
}
read the original abstract

Linkography -- the analysis of links between the design moves that make up an episode of creative ideation or design -- can be used for both visual and quantitative assessment of creative activity traces. Traditional linkography, however, is time-consuming, requiring a human coder to manually annotate both the design moves within an episode and the connections between them. As a result, linkography has not yet been much applied at scale. To address this limitation, we introduce fuzzy linkography: a means of automatically constructing a linkograph from a sequence of recorded design moves via a "fuzzy" computational model of semantic similarity, enabling wider deployment and new applications of linkographic techniques. We apply fuzzy linkography to three markedly different kinds of creative activity traces (text-to-image prompting journeys, LLM-supported ideation sessions, and researcher publication histories) and discuss our findings, as well as strengths, limitations, and potential future applications of our approach.

Figures

Figures reproduced from arXiv: 2502.04599 by the authors.

Figure 1
Figure 1. Fuzzy linkography allows for the rapid translation of user activity logs from digital creativity support tools (and [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Standard linkographic patterns. Moves 0–3 form a [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. An example fuzzy linkograph of an open-ended stream-of-consciousness ideation activity. Links between moves are [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (9 more)
Figure 5
Figure 5. Figure 5: This image prompting linkograph demonstrates a [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: This image prompting linkograph’s “converging zigzag” shape combines the [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: This image prompting linkograph shows clear se [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 9
Figure 9. Figure 9: Five clusters of image prompting linkographs, five random examples per cluster. [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 10
Figure 10. Figure 10: Three LLM-supported ideation linkographs demonstrating many of the archetypal patterns we observed in ideation [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]
Figure 11
Figure 11. Figure 11: Even in LLM-supported ideation linkographs [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]
Figure 12
Figure 12. Figure 12: An LLM-supported ideation linkograph demon [PITH_FULL_IMAGE:figures/full_fig_p009_12.png]
Figure 13
Figure 13. Figure 13: Three publication history linkographs demonstrating three markedly different career shapes. The leftmost linkograph [PITH_FULL_IMAGE:figures/full_fig_p010_13.png]
Figure 14
Figure 14. Figure 14: Linkographic abundance: fuzzy linkographs of 100 randomly selected image prompting traces, laid out as thumbnails. As new activity traces are logged, CST researchers might automatically generate linkographs for these traces and monitor the overall distribution, or use…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

83 extracted references · 73 canonical work pages

  1. [1]

    Leyla Alipour, Mohsen Faizi, Asghar Mohammad Moradi, and Gholamreza Akrami. 2018. A review of design fixation: Research directions and key fac- tors. International Journal of Design Creativity and Innovation 6, 1-2 (2018), 22–35

    2018

  2. [2]

    Alberto Alvarez, Jose Font, and Julian Togelius. 2022. Toward Designer Modeling Through Design Style Clustering. IEEE Transactions on Games 14, 4 (2022), 676–686

    2022

  3. [3]

    Barrett R Anderson, Jash Hemant Shah, and Max Kreminski. 2024. Homogeniza- tion effects of large language models on human creative ideation. In Proceedings of the 16th Conference on Creativity & Cognition . 413–425

    2024

  4. [4]

    Jack Armitage, Thor Magnusson, and Andrew McPherson. 2023. Design Process in Visual Programming: Methods for Visual and Temporal Analysis. In Proc. Sound and Music Computing Conference

    work page 2023

  5. [5]

    Jesse Josua Benjamin, Arne Berger, Nick Merrill, and James Pierce. 2021. Machine learning uncertainty as a design material: A post-phenomenological inquiry. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems

    work page 2021

  6. [6]

    Umang Bhatt, Javier Antorán, Yunfeng Zhang, Q Vera Liao, Prasanna Sattigeri, Riccardo Fogliato, Gabrielle Melançon, Ranganath Krishnan, Jason Stanley, Omesh Tickoo, Lama Nachman, Rumi Chunara, Madhulika Srikumar, Adrian Weller, and Alice Xiang. 2021. Uncertainty as a form of transparency: Measuring, communicating, and using uncertainty. In Proceedings of ...

    work page 2021

  7. [7]

    Alexandra Bremers and Wendy Ju. 2024. Can Machines Tell What People Want? Bringing Situated Intelligence to Generative AI. In Proceedings of the Halfway to the Future Symposium

    work page 2024

  8. [8]

    Hui Cai, Ellen Yi-Luen Do, and Craig M Zimring. 2010. Extended linkography and distance graph in design evaluation: an empirical study of the dual effects of inspiration sources in creative design. Design Studies 31, 2 (2010), 146–168

    work page 2010

Show all 83 references
  1. [9]

    Joel Chan, Steven P Dow, and Christian D Schunn. 2018. Do the best design ideas (really) come from conceptually distant sources of inspiration? Engineering a Better Future: Interplay between Engineering, Social Sciences, and Innovation (2018), 111–139. Fuzzy Linkography: Autom...

    2018

  2. [10]

    Minsuk Chang, Stefania Druga, Alexander J Fiannaca, Pedro Vergani, Chinmay Kulkarni, Carrie J Cai, and Michael Terry. 2023. The prompt artists. InProceedings of the 15th Conference on Creativity and Cognition . 75–87

    2023

  3. [11]

    Erin Cherry and Celine Latulipe. 2014. Quantifying the creativity support of dig- ital tools through the Creativity Support Index. ACM Transactions on Computer- Human Interaction (TOCHI) 21, 4 (2014)

    2014

  4. [12]

    Shyan-Bin Chou. 2007. A method for evaluating creativity in Linkography. In 10th QMOD Conference of Quality Management and Organizational Development. Our Dreams of Excellence , Vol. 26. 18–20

    2007

  5. [13]

    Shyan-Bin Chou, Huey-Wen Chou, and Yui-Liang Chen. 2013. Entropy of Linkog- raphy: Evaluating the creativity of short animation. Creativity Research Journal 25, 1 (2013), 33–37

    2013

  6. [14]

    John Joon Young Chung, Shiqing He, and Eytan Adar. 2021. The intersection of users, roles, interactions, and technologies in creativity support tools. InDesigning Interactive Systems Conference 2021 . 1817–1833

    2021

  7. [15]

    John Joon Young Chung and Max Kreminski. 2024. Patchview: LLM-Powered Worldbuilding with Generative Dust and Magnet Visualization. In Proceedings of the 37th Annual ACM Symposium on User Interface Software and Technology

    2024

  8. [16]

    Kate Compton and Michael Mateas. 2015. Casual Creators. In International Conference on Computational Creativity . 228–235

    2015

  9. [17]

    Nathan Crilly. 2019. Creativity and fixation in the real world: A literature review of case study research. Design Studies 64 (2019), 154–168

    2019

  10. [18]

    Nicholas Davis, Chih-Pin Hsiao, Kunwar Yashraj Singh, Brenda Lin, and Brian Magerko. 2017. Creative sense-making: Quantifying interaction dynamics in co-creation. In Proceedings of the 2017 ACM SIGCHI Conference on Creativity and Cognition. 356–366

    2017

  11. [19]

    Stefano Delle Monache and Davide Rocchesso. 2016. Cooperative sound design: A protocol analysis. In Proceedings of the Audio Mostly 2016 . 154–161

    2016

  12. [20]

    Manoj Deshpande, Jisu Park, Supratim Pait, and Brian Magerko. 2024. Perceptions of Interaction Dynamics in Co-Creative AI: A Comparative Study of Interaction Modalities in Drawcto. In Proceedings of the 16th Conference on Creativity & Cognition. 102–116

    2024

  13. [21]

    Shachar Don-Yehiya, Leshem Choshen, and Omri Abend. 2023. Human Learning by Model Feedback: The Dynamics of Iterative Prompting with Midjourney. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, Houda Bouamor, Juan Pino, and Kalika Bali...

    2023 doi

  14. [22]

    Gerhard Fischer. 2005. Distances and diversity: sources for social creativity. In Proceedings of the 5th Conference on Creativity & Cognition . 128–136

    2005

  15. [23]

    Jonas Frich, Lindsay MacDonald Vermeulen, Christian Remy, Michael Mose Biskjaer, and Peter Dalsgaard. 2019. Mapping the landscape of creativity support tools in HCI. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems

    2019

  16. [24]

    Gabriela Goldschmidt. 1990. Linkography: assessing design productivity. In Cy- berbetics and System ’90, Proceedings of the Tenth European Meeting on Cybernetics and Systems Research, Singapore . World Scientific, 291–298

    1990

  17. [25]

    Gabriela Goldschmidt. 2014. Linkography: Unfolding the Design Process . MIT Press

    2014

  18. [26]

    Gabriela Goldschmidt. 2016. Linkographic evidence for concurrent divergent and convergent thinking in creative design. Creativity Research Journal 28, 2 (2016), 115–122

    2016

  19. [27]

    Adam E Green, Roger E Beaty, Yoed N Kenett, and James C Kaufman. 2024. The process definition of creativity. Creativity Research Journal 36, 3 (2024), 544–572

    2024

  20. [28]

    Zixun Guo, Jaeyong Kang, and Dorien Herremans. 2023. A domain-knowledge- inspired music embedding space and a novel attention mechanism for sym- bolic music modeling. In Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-Fifth Conference on...

    2023 doi

  21. [29]

    Matthew Guzdial, Nicholas Liao, Jonathan Chen, Shao-Yu Chen, Shukan Shah, Vishwa Shah, Joshua Reno, Gillian Smith, and Mark O Riedl. 2019. Friend, collaborator, student, manager: How design of an AI-driven game level editor affects creators. In Proceedings of the 2019 CHI Conf...

    2019

  22. [30]

    Tellervo Härkki. 2023. Mobile gaze tracking and an extended linkography for collaborative sketching and designing. International Journal of Technology and Design Education 33, 2 (2023), 379–413

    2023

  23. [31]

    Gillian Hatcher, William Ion, Ross Maclachlan, Marion Marlow, Barbara Simpson, Nicky Wilson, and Andrew Wodehouse. 2018. Using linkography to compare creative methods for group ideation. Design Studies 58 (2018), 127–152

    2018

  24. [32]

    Tom Hewett, Mary Czerwinski, Michael Terry, Jay Nunamaker, Linda Candy, Bill Kules, and Elisabeth Sylvan. 2005. Creativity support tool evaluation methods and metrics. In Creativity Support Tools: A workshop sponsored by the National Science Foundation. 10–24

    2005

  25. [33]

    Qingqing Huang, Aren Jansen, Joonseok Lee, Ravi Ganti, Judith Yue Li, and Daniel P. W. Ellis. 2022. MuLan: A Joint Embedding of Music Audio and Natural Language. In Proceedings of the 23rd International Society for Music Information Retrieval Conference, ISMIR 2022, Bengaluru,...

    2022

  26. [34]

    David G Jansson and Steven M Smith. 1991. Design fixation. Design Studies 12, 1 (1991), 3–11

    1991

  27. [35]

    Youngseung Jeon, Seungwan Jin, Patrick C Shih, and Kyungsik Han. 2021. Fash- ionQ: an AI-driven creativity support tool for facilitating ideation in fashion design. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems

    2021

  28. [36]

    Jeff WT Kan, Zafer Bilda, and John S Gero. 2007. Comparing entropy measures of idea links in design protocols: Linkography entropy measurement and analysis of differently conditioned design sessions. AI EDAM 21, 4 (2007), 367–377

    2007

  29. [37]

    Jeff WT Kan and John S Gero. 2005. Can entropy indicate the richness of idea generation in team designing? In CAADRIA05. 451–457

    2005

  30. [38]

    Jeff WT Kan and John S Gero. 2008. Acquiring information from linkography in protocol studies of designing. Design Studies 29, 4 (2008), 315–337

    2008

  31. [39]

    Max Kreminski and John Joon Young Chung. 2024. Intent Elicitation in Mixed- Initiative Co-Creativity. In Joint Proceedings of the ACM IUI Workshops

    2024

  32. [40]

    Max Kreminski, Melanie Dickinson, Joseph Osborn, Adam Summerville, Michael Mateas, and Noah Wardrip-Fruin. 2020. Germinate: A mixed-initiative casual creator for rhetorical games. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertai...

    2020

  33. [41]

    Max Kreminski, Isaac Karth, Michael Mateas, and Noah Wardrip-Fruin. 2022. Evaluating mixed-initiative creative interfaces via expressive range coverage analysis. In IUI Workshops. 34–45

    2022

  34. [42]

    Max Kreminski and Michael Mateas. 2021. Reflective creators. In International Conference on Computational Creativity . 309–318

    2021

  35. [43]

    Arniel Labrada, Benjamin Bustos, and Ivan Sipiran. 2024. A convolutional archi- tecture for 3D model embedding using image views. The Visual Computer 40, 3 (2024), 1601–1615

    2024

  36. [44]

    Tomas Lawton, Francisco J Ibarrola, Dan Ventura, and Kazjon Grace. 2023. Draw- ing with Reframer: Emergence and control in co-creative AI. In Proceedings of the 28th International Conference on Intelligent User Interfaces . 264–277

    2023

  37. [45]

    Alghamdi, Tal August, Avinash Bhat, Madiha Zahrah Choksi, Senjuti Dutta, Jin L.C

    Mina Lee, Katy Ilonka Gero, John Joon Young Chung, Simon Buckingham Shum, Vipul Raheja, Hua Shen, Subhashini Venugopalan, Thiemo Wambsganss, David Zhou, Emad A. Alghamdi, Tal August, Avinash Bhat, Madiha Zahrah Choksi, Senjuti Dutta, Jin L.C. Guo, Md Naimul Hoque, Yewon Kim, S...

    2024

  38. [46]

    Seung Won Lee, Tae Hee Jo, Semin Jin, Jiin Choi, Kyungwon Yun, Sergio Bromberg, Seonghoon Ban, and Kyung Hoon Hyun. 2024. The Impact of Sketch- guided vs. Prompt-guided 3D Generative AIs on the Design Exploration Process. In Proceedings of the CHI Conference on Human Factors i...

    2024

  39. [47]

    Jialang Victor Li, Max Kreminski, Sean M Fernandes, Anya Osborne, Joshua McVeigh-Schultz, and Katherine Isbister. 2022. Conversation balance: A shared VR visualization to support turn-taking in meetings. InCHI Conference on Human Factors in Computing Systems Extended Abstracts

    2022

  40. [48]

    Antonios Liapis, Georgios Yannakakis, and Julian Togelius. 2013. Designer modeling for personalized game content creation tools. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment , Vol. 9. 11–16

    2013

  41. [49]

    Ryan Louie, Andy Coenen, Cheng Zhi Huang, Michael Terry, and Carrie J Cai

  42. [50]

    Ryan Louie, Jesse Engel, and Cheng-Zhi Anna Huang. 2022. Expressive communi- cation: Evaluating developments in generative models and steering interfaces for music creation. In Proceedings of the 27th International Conference on Intelligent User Interfaces. 405–417

    2022

  43. [51]

    Todd I Lubart. 2001. Models of the creative process: Past, present and future. Creativity Research Journal 13, 3-4 (2001), 295–308

    2001

  44. [52]

    Shruti Mahajan, Leo Bunyea, Nathan Partlan, Dylan Schout, Casper Harteveld, Camillia Matuk, Will Althoff, Tyler Duke, Steven Sutherland, and Gillian Smith

  45. [53]

    Atefeh Mahdavi Goloujeh, Anne Sullivan, and Brian Magerko. 2024. Is It AI or Is It Me? Understanding Users’ Prompt Journey with Text-to-Image Generative AI Tools. In Proceedings of the CHI Conference on Human Factors in Computing Systems. C&C ’25, June 23–25, 2025, Virtual, Un...

    2024

  46. [54]

    Jon McCormack, Camilo Cruz Gambardella, Nina Rajcic, Stephen James Krol, Maria Teresa Llano, and Meng Yang. 2023. Is writing prompts really making art? In International Conference on Computational Intelligence in Music, Sound, Art and Design (Part of EvoStar) . Springer, 196–211

    2023

  47. [55]

    Jon McCormack, Maria Teresa Llano, Stephen James Krol, and Nina Rajcic. 2024. No Longer Trending on Artstation: Prompt Analysis of Generative AI Art. In International Conference on Computational Intelligence in Music, Sound, Art and Design (Part of EvoStar) . Springer, 279–295

    2024

  48. [56]

    Mark J Nelson, Swen E Gaudl, Simon Colton, and Sebastian Deterding. 2018. Curious users of casual creators. InProceedings of the 13th International Conference on the Foundations of Digital Games

    2018

  49. [57]

    Tricia J Ngoon, C Ailie Fraser, Ariel S Weingarten, Mira Dontcheva, and Scott Klemmer. 2018. Interactive guidance techniques for improving creative feedback. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems

    2018

  50. [58]

    Tricia J Ngoon, Joy O Kim, and Scott Klemmer. 2021. Shöwn: Adaptive Conceptual Guidance Aids Example Use in Creative Tasks. In Proceedings of the 2021 ACM Designing Interactive Systems Conference . 1834–1845

    2021

  51. [59]

    Dmitry Nikolaev and Sebastian Padó. 2023. Representation biases in sentence transformers. arXiv preprint arXiv:2301.13039 (2023)

    2023 arXiv

  52. [60]

    Peter Organisciak, Selcuk Acar, Denis Dumas, and Kelly Berthiaume. 2023. Be- yond semantic distance: Automated scoring of divergent thinking greatly im- proves with large language models. Thinking Skills and Creativity 49 (2023), 101356

    2023

  53. [61]

    Lace Padilla, Matthew Kay, and Jessica Hullman. 2022. Uncertainty visualization. In Computational Statistics in Data Science . Wiley, 405–421

    2022

  54. [62]

    Morteza Pourmohamadi and John S Gero. 2011. LINKOgrapher: An analysis tool to study design protocols based on FBS coding scheme. In DS 68-2: Proceedings of the 18th International Conference on Engineering Design (ICED 11), Impacting So- ciety through Engineering Design, Vol. 2...

    2011

  55. [63]

    A Terry Purcell and John S Gero. 1996. Design and other types of fixation.Design Studies 17, 4 (1996), 363–383

    1996

  56. [64]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. 2021. Learning transferable visual mod- els from natural language supervision. In Internati...

    2021

  57. [65]

    Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. In Proceedings of the 2019 Conference on Em- pirical Methods in Natural Language Processing . Association for Computational Linguistics. https://arxiv.org/abs/1908.10084

    2019 arXiv

  58. [66]

    Christian Remy, Lindsay MacDonald Vermeulen, Jonas Frich, Michael Mose Biskjaer, and Peter Dalsgaard. 2020. Evaluating creativity support tools in HCI research. In Proceedings of the 2020 ACM Designing Interactive Systems Conference . 457–476

    2020

  59. [67]

    Huda Salman. 2014. Linkography for evaluating ideas connectivity of Computer Aided Design-based protocols. In eCAADe, Vol. 1. 573–581

    2014

  60. [68]

    Téo Sanchez. 2023. Examining the Text-to-Image Community of Practice: Why and How do People Prompt Generative AIs? In Proceedings of the 15th Conference on Creativity and Cognition . 43–61

    2023

  61. [69]

    Ben Shneiderman. 2007. Creativity support tools: accelerating discovery and innovation. Commun. ACM 50, 12 (2007), 20–32

    2007

  62. [70]

    Pao Siangliulue, Kenneth C Arnold, Krzysztof Z Gajos, and Steven P Dow. 2015. Toward collaborative ideation at scale: Leveraging ideas from others to generate more creative and diverse ideas. In Proceedings of the 18th ACM Conference on Computer Supported Cooperative Work & So...

    2015

  63. [71]

    Amy Smith and Michael Cook. 2023. AI-Generated Imagery: A New Era for the ‘Readymade’. In SIGGRAPH Asia 2023 Art Papers

    2023

  64. [72]

    Gillian Smith and Jim Whitehead. 2010. Analyzing the expressive range of a level generator. In Proceedings of the 2010 Workshop on Procedural Content Generation in Games

    2010

  65. [73]

    Saba Sturua, Isabelle Mohr, Mohammad Kalim Akram, Michael Günther, Bo Wang, Markus Krimmel, Feng Wang, Georgios Mastrapas, Andreas Koukounas, Andreas Koukounas, Nan Wang, and Han Xiao. 2024. jina-embeddings-v3: Multilingual Embeddings With Task LoRA. arXiv:2409.10173 [cs.CL] h...

    2024 arXiv

  66. [74]

    Mikaela Angelina Uy, Jingwei Huang, Minhyuk Sung, Tolga Birdal, and Leonidas Guibas. 2020. Deformation-aware 3D model embedding and retrieval. InComputer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part VII 16 . Springer, 397–413

    2020

  67. [75]

    Rosario Vidal, Elena Mulet, and Eliseo Gómez-Senent. 2004. Effectiveness of the means of expression in creative problem-solving in design groups. Journal of Engineering Design 15, 3 (2004), 285–298

    2004

  68. [76]

    Benjamin Warner, Antoine Chaffin, Benjamin Clavié, Orion Weller, Oskar Hall- ström, Said Taghadouini, Alexis Gallagher, Raja Biswas, Faisal Ladhak, Tom Aarsen, Nathan Cooper, Griffin Adams, Jeremy Howard, and Iacopo Poli. 2024. Smarter, Better, Faster, Longer: A Modern Bidirec...

    2024 arXiv

  69. [77]

    Andy Warr and Eamonn O’Neill. 2005. Understanding design as a social creative process. In Proceedings of the 5th Conference on Creativity & Cognition . 118–127

    2005

  70. [78]

    Jason Wu, Yi-Hao Peng, Xin Yue Amanda Li, Amanda Swearngin, Jeffrey P Bigham, and Jeffrey Nichols. 2024. UIClip: a data-driven model for assessing user interface design. In Proceedings of the 37th Annual ACM Symposium on User Interface Software and Technology

    2024

  71. [79]

    Yutong Xie, Zhaoying Pan, Jinge Ma, Luo Jie, and Qiaozhu Mei. 2023. A prompt log analysis of text-to-image generation systems. In Proceedings of the ACM Web Conference 2023. 3892–3902

    2023

  72. [80]

    Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. 2023. Sig- moid loss for language image pre-training. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision . 11975–11986

    2023

  73. [81]

    Xiaoxuan Zhang, Zeping Zhan, Misha Holtz, and Adam M Smith. 2018. Crawling, indexing, and retrieving moments in videogames. In Proceedings of the 13th International Conference on the Foundations of Digital Games

    2018

  74. [2019]

    In Proceedings of the 15th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE-19)

    Toward automated critique for student-created interactive narrative projects. In Proceedings of the 15th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE-19)

  75. [2020]

    In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems

    Novice-AI music co-creation via AI-steering tools for deep generative models. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems

    2020

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.