Pith. sign in

REVIEW 2 major objections 6 minor 30 references

Semantic Bundling: Interactive Node and Edge Bundling to Simplify Knowledge Graphs using Large Language Models

T0 review · 2 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Semantic Bundling lets large language models transform tangled knowledge graphs into labeled, evidence-linked super nodes and super edges.

desk verdict Useful technique paper with a fixable traceability gap; worth a serious referee. read the letter →

arxiv 2608.04002 v1 pith:IEYBNQDC submitted 2026-08-04 cs.HC

Adam Coscia , Zeyu Hua , Eric Krokos , Timothy Lin , Alex Endert This is my paper · ORCID
classification cs.HC
keywords knowledgegraphsvisualanalyticslargelanguagemodelsgraphbundlingsupernodesedgessensemakingtraceability
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper is trying to establish that large language models can serve as a semantic bundling layer for knowledge graphs: instead of merely decluttering a graph, a user can collapse a dense region into a labeled super node or summarize the connection between two entities as a super edge, with every summarized claim linked to the underlying triples and source documents. If true, document analysis could proceed on the graph: analysts read generated summaries, verify them sentence by sentence, and only open the source text to confirm or explore. The paper contributes the technique, an open-source system called AgentK that implements it, and two use cases suggesting the kind of insight this enables.

What carries the argument

The central mechanism is the pair of bundling operations. A super node is created when the user selects a node or lassoes a subgraph: the triples in that region are grouped by a chosen granularity, an LLM labels and summarizes each group, a cross-encoder links each summary sentence to the triple it best supports, and motif detection labels fans, cliques, and connectors so the collapsed glyph preserves the structural shape of the region. A super edge is created when the user selects two entities: parallel edges or bounded simple paths between them are gathered and grouped by strategy, an LLM produces a summary triple as the label plus a natural-language summary, and the same sentence-to-triple linkage makes the result traceable. That linkage is what turns an LLM summary from a plausible paraphrase into an auditable claim.

What would settle it

Generate a hundred bundles from a corpus with known ground-truth relationships, then have annotators mark every summary sentence that asserts a relationship absent from the triples it is linked to; any such sentence is a direct counterexample to the traceability claim. A cheaper automated version would inject a fabricated relationship into a summary and check whether the cross-encoder still links it to a supported triple.

Watch

Extended reading notes

Core claim

The central claim is that a knowledge graph built from a closed document set can be reorganized, on demand and by direct manipulation, into higher-level structures: super nodes collapse and summarize a region of the graph, and super edges summarize how two entities are connected, whether through parallel edges or multi-hop paths. Because each generated label and summary sentence is linked to the triples that best support it, and each triple links to its source documents, the bundles remain grounded in evidence rather than floating free as LLM text. The paper realizes this claim in AgentK, which maps graph interactions to bundling operations, and demonstrates through use cases on movie reviews and an intelligence analysis scenario that a user can go from a hairball to a hypothesis, then from any summary sentence down to source text.

Load-bearing premise

The technique's traceability rests on the LLM sticking to the triples it is shown and on the cross-encoder correctly matching each summary sentence to the triple that best supports it; if either step fails, a claim can appear grounded when the underlying data does not support it.

Editorial extensions

If this is right

  • A user does not need to read a dense corpus to learn what a region of its knowledge graph means; the super node labels and summaries supply the meaning in place.
  • A connection between two entities with no direct edge can still be read as a single labeled relationship, because super edges summarize bounded multi-hop paths.
  • Every sentence in a bundle is traceable: hovering a sentence highlights the triples that support it, and clicking a triple opens the source document, so a generated claim can be checked rather than trusted.
  • Different path-selection and grouping strategies produce genuinely different super edges over the same graph, letting a user contrast alternative interpretations of how two entities are linked.
  • The technique transfers across domains without schema engineering, since the LLM defines the ontology at extraction time.

Reading between the lines

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

  • The paper does not develop this, but the sentence-to-triple linkage could serve as a general provenance layer for any LLM summarization of graph-structured data, not just node-link displays.
  • Because a bundle can be deleted to restore the original graph, the technique implies a lightweight exploration history, yet the paper notes only single-step undo.
  • A controlled experiment comparing analysts with and without Semantic Bundling on hypothesis generation or evidence verification would test whether the qualitative benefits shown in the use cases hold up; the paper explicitly leaves that to future work.
  • The same operations could plausibly extend to temporal knowledge graphs, comparing bundles across time, or to reconciling entities across separate graphs, both of which the paper lists as open directions.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. Semantic Bundling proposes two LLM-driven operations—super nodes, which collapse and summarize a region of a knowledge graph, and super edges, which summarize direct or multi-hop connections between two entities—while linking the generated text back to underlying triples and source documents. The authors instantiate the technique in AgentK, an open-source visual analytics system that builds a KG from documents, maps direct-manipulation selections to bundling operations, and renders bundles as glyphs with details-on-demand. The paper reports two usage scenarios (IMDb movie reviews and the Kronos intelligence corpus) and discusses trade-offs of LLM-defined ontologies and limitations.

Significance. The technical design is clearly described, and the open-source implementation is a concrete contribution that the community can build on. The idea of using LLMs to relabel collapsed KG structures is timely, and the evidence-linkage mechanism is a reasonable first response to the hallucination concern. If the traceability guarantee were verified, the work would be a solid step toward KG sensemaking. However, the current manuscript does not verify the central grounding claim, and the usage scenarios are self-authored and favorable, so the significance is conditional on additional evidence.

major comments (2)
  1. [III-E, III-C step 3, III-D step 4] The grounding mechanism is one-directional: the implementation links each triple to the summary sentence it best supports, but nothing checks that every summary sentence is supported by some triple. An ungrounded sentence produced by the LLM simply has no linked triples and is not flagged in the UI. Because traceability is the property that distinguishes Semantic Bundling from structure-only bundling (challenge C4), this missing verification is load-bearing. Moreover, stsb-distilroberta-base is a semantic-textual-similarity model, not an entailment model, so "best supports" is a similarity ranking and cannot detect contradictions or ungrounded synthesis. Please add a sentence-level faithfulness check (for example, NLI-based entailment of each sentence against the selected triples) and report its accuracy, or revise the claim to describe triples being linked to sentences rather than every summary being traceable to evidence.
  2. [V, VI-C] The two usage scenarios are authored by the system's designers, rely on their own LLM prompts, and select favorable examples; they do not independently establish that the summaries are faithful or that the reported insights generalise. The limitation paragraph in §VI-C acknowledges the absence of a controlled study, but the abstract and conclusion still state that the use cases show how Semantic Bundling reveals new insights. At minimum, the paper should report a systematic audit of summary-sentence support (e.g., the fraction of sentences with at least one linked triple, and human judgments of whether the linked triples entail the sentence) on the two datasets, or explicitly downgrade the "reveals new insights" claim to "illustrates potential."
minor comments (6)
  1. [VI] There is a typo in the opening of Section VI: "ontolgies" should be "ontologies."
  2. [II-B] The characterization of Biset [21] as prior work that "bundles such edges by structure" is inaccurate: Biset is explicitly a semantic edge bundling technique. The actual novelty of Semantic Bundling lies in LLM-generated natural-language labels and summaries with evidence linkage, not in semantic bundling per se. Please correct the related-work framing so the contribution is stated against the correct baseline.
  3. [V-A] The statement that there was "no direct edge initially linked director Christopher Nolan to actor Matthew McConaughey" should report the actual path length and path enumeration settings used, so the reader can judge the claim.
  4. [IV-D] The paper does not report basic graph statistics (number of entities, triples, edges, and paths enumerated) for the IMDb and Kronos graphs; these numbers would help assess the scalability limitations discussed in §VI-C.
  5. [III-C, III-D] The full LLM prompts are not included in the paper; since the technique's behavior depends on the exact prompts, including them in an appendix (as promised in the supplemental materials) is important for reproducibility.
  6. [Fig. 1] The illustrative summary in Figure 1 includes statements such as "A is more talkative" and "B likes to order pizza from time to time," which appear to go beyond the example triples about phone calls and cafe meetings; this underlines the need for the faithfulness audit requested in the major comments.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper presents a new LLM-driven visual analytics technique with no equation-level derivation, no fitted parameters, and no load-bearing self-citation chain.

full rationale

Semantic Bundling is a design and system paper rather than a quantitative derivation. The central claims are that LLM-generated super nodes and super edges summarize knowledge-graph regions while remaining linked to underlying triples and source documents. These claims are supported by a described algorithm (Section III-C/D), an implementation (Section IV), and illustrative use cases (Section V), not by a formal predictive model that could reduce to its own inputs by construction. The traceability mechanism links each triple to the summary sentence it best supports via a cross-encoder, and the paper explicitly acknowledges that the LLM is only prompted, not verified, to avoid speculation. That is a potential correctness or evaluation gap, but it is not circularity: the technique does not define its output in terms of its input, fit a parameter and then rename it a prediction, or import a uniqueness result from the authors' prior work. The only self-citation (Ref. [11], Vispile) appears in related work and is not load-bearing for the paper's contributions. Use cases are authored demonstrations rather than controlled benchmarks, but absence of external validation is a generalizability concern, not a circular-reasoning concern. Overall, no circular step is exhibited.

Assumptions & free parameters 3 free parameters · 4 assumptions · 2 invented entities

This is a technique-effectiveness claim, not a numerical derivation, so no constants are fitted to data. The core burden is a set of behavioral assumptions about LLM faithfulness, evidence linkage, and interaction mapping; these are plausible but unmeasured. The introduced entities, super nodes and super edges, are the contribution itself, so their benefit is the claim under test rather than an independent axiom.

free parameters (3)
  • Maximum path length for multi-hop super edges
    User-configurable bound on simple-path enumeration; the value determines which distant connections can be discovered, and no default or sensitivity analysis is reported.
  • Granularity and clustering settings for super nodes and super edges
    Number of groups, topic ranking, and HDBSCAN or silhouette settings are chosen per user or by automatic heuristics; the paper does not report how these choices affect summary quality.
  • LLM prompt and generation settings
    Exact prompts are deferred to supplemental materials and generation parameters such as temperature are not reported, so the quality of labels and summaries cannot be reproduced or benchmarked.
assumptions (4)
  • domain assumption LLM extraction produces a knowledge graph whose triples are checkable and faithful to source documents.
    AgentK's extraction pipeline in Section IV-A uses GPT-5.2 to build the graph; the paper later acknowledges that the quality of the extracted graph bounds everything built on top of it in Section VI-C.
  • domain assumption The LLM summarization, when constrained to selected triples, does not speculate beyond the given triples.
    Section III-E claims the LLM labels and condenses the graph but does not speculate; this is an assumption about model behavior, not a verified property.
  • domain assumption Cross-encoder sentence-to-triple linking accurately identifies the evidence supporting each summary sentence.
    Sections III-C and III-D rely on this linkage for traceability; no accuracy evaluation is provided.
  • domain assumption The mapping from selection type to operation, such as single selection or lasso to super node and two nodes to super edge, matches user intent.
    Section IV-B describes this mapping as a design choice; no user study validates it.
invented entities (2)
  • Super node
    purpose: Collapses and summarizes a selected region of a knowledge graph into a single labeled glyph while preserving motif structure.
    Introduced by this paper; its demonstrated benefit comes only from the authors' own use cases, with no external benchmark.
  • Super edge
    purpose: Summarizes parallel edges or multi-hop paths between two entities into a labeled relationship triple.
    Introduced by this paper; the semantic-edge version overlaps with Biset [21] and the effectiveness evidence is limited to the paper's own scenarios.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semantic Bundling: Interactive Node and Edge Bundling to Simplify Knowledge Graphs using Large Language Models." pith.science (2026). https://pith.science/paper/IEYBNQDC

@misc{pith2026260804002,
  author       = {Pith},
  title        = {Pith review of: Semantic Bundling: Interactive Node and Edge Bundling to Simplify Knowledge Graphs using Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IEYBNQDC}},
  note         = {Machine review of arXiv:2608.04002}
}
read the original abstract

We present Semantic Bundling, a visual analytics technique for making sense of text documents represented as knowledge graphs (KGs). Representing a document corpus as a KG makes relationships between entities explicit, making KGs useful both to analyze directly and in computational workflows including ML pipelines and generative AI backends. However, as KGs grow they become difficult to interpret and visualize for specific tasks (e.g., the ``hairball problem''), with the meaning of each relationship often buried in dense source text. Semantic Bundling uses large language models (LLMs) to support user-driven bundling of nodes and edges in a KG into higher-level graph structures: super nodes, which collapse and summarize a region of the graph, and super edges, which summarize the connection between two entities. Results are linked to underlying triples and source documents, grounding summaries in evidence. We implement Semantic Bundling in AgentK, an open-source system that builds a KG from text documents and maps graph interactions to bundling operations. Through use cases on movie reviews and an intelligence analysis scenario, we show how Semantic Bundling reveals new insights in document collections, and synthesize our findings into a discussion of emerging challenges and opportunities in knowledge graph sensemaking.

Figures

Figures reproduced from arXiv: 2608.04002 by the authors.

Figure 1
Figure 1. Semantic Bundling maps a user’s graph selection to an operation that collapses the selected structure into a labeled, summarized glyph. Selecting [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Semantic Bundling uses several glyphs to convey the structure of the consolidated subgraph. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. The AgentK main view loaded with the KRONOS dataset [15]. Users (a) explore the knowledge graph as a force-directed node-link diagram, (b) [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The AgentK results view loaded with the KRONOS dataset [15]. Selecting nodes/edges in the main view opens the Bundles tab (a), which exposes [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: The dense hairball surrounding the node “Interstellar” is consolidated [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: User-selected subgraph is abstracted into a super node, preserving [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Nodes related to the focal topic “science” are selectively clustered and [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: A logical multi-hop relationship between Nolan and McConaughey [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 10
Figure 10. Figure 10: The initial overview of the KRONOS dataset in AgentK. The [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]
Figure 11
Figure 11. Figure 11: Betty’s super node workflow: (a) Inspecting the semantic summary of [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]
Figure 13
Figure 13. Figure 13: Betty’s intent discovery workflow: (a) Querying keywords in the [PITH_FULL_IMAGE:figures/full_fig_p011_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 26 canonical work pages

  1. [21]

    Biset: Semantic edge bundling with biclusters for sensemaking,

    M. Sun, P. Mi, C. North, and N. Ramakrishnan, “Biset: Semantic edge bundling with biclusters for sensemaking,”IEEE Transactions on Visualization and Computer Graphics, vol. 22, no. 1, pp. 310–319, 2016

    work page 2016

  2. [1]

    Lex2kg: automatic conversion of legal documents to knowledge graph,

    M. Abdurahman, F. Darari, H. Lesmana, M. Hartopo, I. Rhesa, and B. C. L. Tobing, “Lex2kg: automatic conversion of legal documents to knowledge graph,” in2021 International Conference on Advanced Computer Science and Information Systems (ICACSIS). IEEE, 2021, pp. 1–5

    work page 2021

  3. [2]

    Doc2kg: Transforming document repositories to knowledge graphs,

    N. Stylianou, D. Vlachava, I. Konstantinidis, N. Bassiliades, and V . Peri- steras, “Doc2kg: Transforming document repositories to knowledge graphs,”International Journal on Semantic Web and Information Sys- tems (IJSWIS), vol. 18, no. 1, pp. 1–20, 2022

    work page 2022

  4. [3]

    Building and using a knowledge graph to combat human trafficking,

    P. Szekely, C. A. Knoblock, J. Slepicka, A. Philpot, A. Singh, C. Yin, D. Kapoor, P. Natarajan, D. Marcu, K. Knightet al., “Building and using a knowledge graph to combat human trafficking,” inInternational Semantic Web Conference. Springer, 2015, pp. 205–221

    work page 2015

  5. [4]

    Obsidian,

    S. Li and E. Xu, “Obsidian,” 2026. [Online]. Available: https: //obsidian.md/

    work page 2026

  6. [5]

    New, “Atlas,” 2026

    J. New, “Atlas,” 2026. [Online]. Available: https://www.atlasworkspace. ai/

    work page 2026

  7. [6]

    Knowledge graph builder,

    N. Inc., “Knowledge graph builder,” Jun. 2026. [Online]. Available: https://github.com/neo4j-labs/llm-graph-builder

    work page 2026

  8. [7]

    Themeriver: Visualiz- ing thematic changes in large document collections,

    S. Havre, E. Hetzler, P. Whitney, and L. Nowell, “Themeriver: Visualiz- ing thematic changes in large document collections,”IEEE transactions on visualization and computer graphics, vol. 8, no. 1, pp. 9–20, 2002

    work page 2002

Show all 30 references
  1. [8]

    Visualizing the non-visual: Spatial analysis and interaction with information from text documents,

    J. A. Wise, J. J. Thomas, K. Pennock, D. Lantrip, M. Pottier, A. Schur, and V . Crow, “Visualizing the non-visual: Spatial analysis and interaction with information from text documents,” inProceedings of Visualization 1995 Conference. IEEE, 1995, pp. 51–58

    1995

  2. [9]

    Jigsaw: Supporting investigative analysis through interactive visualization,

    J. Stasko, C. Gorg, Z. Liu, and K. Singhal, “Jigsaw: Supporting investigative analysis through interactive visualization,” in2007 IEEE Symposium on Visual Analytics Science and Technology, 2007, pp. 131– 138

    2007

  3. [10]

    Semantic interaction for visual text analytics,

    A. Endert, P. Fiaux, and C. North, “Semantic interaction for visual text analytics,” inProceedings of the SIGCHI conference on Human factors in computing systems, 2012, pp. 473–482

    2012

  4. [11]

    Vispile: A visual analytics system for analyzing multiple text documents with large language models and knowledge graphs,

    A. Coscia and A. Endert, “Vispile: A visual analytics system for analyzing multiple text documents with large language models and knowledge graphs,” inProceedings of the 59th Hawaii International Conference on System Sciences, ser. HICSS-59, 2026

    2026

  5. [12]

    Apolo: making sense of large network data by combining rich user interaction and machine learning,

    D. H. Chau, A. Kittur, J. I. Hong, and C. Faloutsos, “Apolo: making sense of large network data by combining rich user interaction and machine learning,” inProceedings of the SIGCHI conference on human factors in computing systems, 2011, pp. 167–176

    2011

  6. [13]

    Strat- isfimal layout: A modular optimization model for laying out layered node-link network visualizations,

    S. Di Bartolomeo, M. Riedewald, W. Gatterbauer, and C. Dunne, “Strat- isfimal layout: A modular optimization model for laying out layered node-link network visualizations,”IEEE Transactions on Visualization and Computer Graphics, vol. 28, no. 1, pp. 324–334, 2021

    2021

  7. [14]

    Imdb movie reviews dataset,

    A. Pal, A. Barigidad, and A. Mustafi, “Imdb movie reviews dataset,”

  8. [15]

    Vast challenge 2014: The kronos incident,

    M. Whiting, K. Cook, G. Grinstein, K. Liggett, M. Cooper, J. Fallon, and M. Morin, “Vast challenge 2014: The kronos incident,” in2014 IEEE Conference on Visual Analytics Science and Technology (VAST), 2014, pp. 295–300

    2014

  9. [16]

    The sensemaking process and leverage points for analyst technology as identified through cognitive task analysis,

    P. Pirolli and S. Card, “The sensemaking process and leverage points for analyst technology as identified through cognitive task analysis,” in Proceedings of international conference on intelligence analysis, vol. 5. McLean, V A, USA, 2005, pp. 2–4

    2005

  10. [17]

    Visual exploration of neural document embedding in information retrieval: Semantics and feature selection,

    X. Ji, H.-W. Shen, A. Ritter, R. Machiraju, and P.-Y . Yen, “Visual exploration of neural document embedding in information retrieval: Semantics and feature selection,”IEEE transactions on visualization and computer graphics, vol. 25, no. 6, pp. 2181–2192, 2019

    2019

  11. [18]

    Topic-based exploration and embedded visualizations for research idea generation,

    H. Guo and D. H. Laidlaw, “Topic-based exploration and embedded visualizations for research idea generation,”IEEE transactions on visu- alization and computer graphics, vol. 26, no. 3, pp. 1592–1607, 2018

    2018

  12. [19]

    Kg4vis: A knowledge graph-based approach for visualization recommendation,

    H. Li, Y . Wang, S. Zhang, Y . Song, and H. Qu, “Kg4vis: A knowledge graph-based approach for visualization recommendation,”IEEE Trans- actions on Visualization and Computer Graphics, vol. 28, no. 1, pp. 195–205, 2021

    2021

  13. [20]

    Visualization and exploration of linked data using virtual reality,

    A. J. Kellmann, M. Postema, J. de Keijser, P. Svetachov, R. C. Wilson, E. J. van Enckevort, and M. A. Swertz, “Visualization and exploration of linked data using virtual reality,”Database, vol. 2024, p. baae008, 2024

    2024

  14. [22]

    Motif simplification: improving net- work visualization readability with fan, connector, and clique glyphs,

    C. Dunne and B. Shneiderman, “Motif simplification: improving net- work visualization readability with fan, connector, and clique glyphs,” inProceedings of the SIGCHI conference on human factors in computing systems, 2013, pp. 3247–3256

    2013

  15. [23]

    Hierarchical aggregation for informa- tion visualization: Overview, techniques, and design guidelines,

    N. Elmqvist and J.-D. Fekete, “Hierarchical aggregation for informa- tion visualization: Overview, techniques, and design guidelines,”IEEE transactions on visualization and computer graphics, vol. 16, no. 3, pp. 439–454, 2009

    2009

  16. [24]

    Unifying large language models and knowledge graphs: A roadmap,

    S. Pan, L. Luo, Y . Wang, C. Chen, J. Wang, and X. Wu, “Unifying large language models and knowledge graphs: A roadmap,”IEEE Transactions on Knowledge and Data Engineering, vol. 36, no. 7, pp. 3580–3599, 2024

    2024

  17. [25]

    Think-on-graph: Deep and responsible reasoning of large language model on knowledge graph,

    J. Sun, C. Xu, L. Tang, S. Wang, C. Lin, Y . Gong, L. Ni, H.-Y . Shum, and J. Guo, “Think-on-graph: Deep and responsible reasoning of large language model on knowledge graph,” inInternational Conference on Learning Representations, vol. 2024, 2024, pp. 3868–3898

    2024

  18. [26]

    Ava: An automated and ai-driven intelligent visual analytics framework,

    J. Wang, X. Li, C. Li, D. Peng, A. Z. Wang, Y . Gu, X. Lai, H. Zhang, X. Xu, X. Dong, Z. Lin, J. Zhou, X. Liu, and W. Chen, “Ava: An automated and ai-driven intelligent visual analytics framework,”Visual Informatics, vol. 8, no. 2, pp. 106–114, 2024. [Online]. Available: https...

    2024

  19. [27]

    Llm- assisted visual analytics: Opportunities and challenges,

    M. Hutchinson, R. Jianu, A. Slingsby, and P. Madhyastha, “Llm- assisted visual analytics: Opportunities and challenges,”arXiv preprint arXiv:2409.02691, 2024

    2024 arXiv

  20. [28]

    Leva: Using large language models to enhance visual analytics,

    Y . Zhao, Y . Zhang, Y . Zhang, X. Zhao, J. Wang, Z. Shao, C. Turkay, and S. Chen, “Leva: Using large language models to enhance visual analytics,”IEEE transactions on visualization and computer graphics, vol. 31, no. 3, pp. 1830–1847, 2024

    2024

  21. [29]

    Semdedup: Data-efficient learning at web-scale through semantic deduplication,

    A. K. M. Abbas, K. Tirumala, D. Simig, S. Ganguli, and A. S. Morcos, “Semdedup: Data-efficient learning at web-scale through semantic deduplication,” inICLR 2023 Workshop on Mathematical and Empirical Understanding of Foundation Models, 2023. [Online]. Available: https://openr...

    2023

  22. [2020]

    Available: https://dx.doi.org/10.21227/zm1y-b270

    [Online]. Available: https://dx.doi.org/10.21227/zm1y-b270

    doi

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.