Pith. sign in

REVIEW 4 major objections 4 minor 51 references

Cognitively-Inspired Emergent Communication via Knowledge Graphs for Assisting the Visually Impaired

T0 review · 4 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Graph-structured inputs make emergent communication symbols more compositional and context-stable, the paper reports.

desk verdict Plausible idea, unsupported evaluation: the graph-based EC pipeline is clearly described, but the baseline, metric provenance, and figure labels do not back the headline numbers. read the letter →

arxiv 2505.22087 v1 pith:JBMV5K4U submitted 2025-05-28 cs.AI

Ruxiao Chen , Dezheng Han , Wenjie Han , Shuaishuai Guo This is my paper
classification cs.AI
keywords emergentcommunicationknowledgegraphsvisualattentioncognitivemapsassistivetechnologyvisuallyimpairedreferentialgameGumbel-Softmax
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

VAG-EC (Visual Attention Graph-based Emergent Communication) claims that agents communicating about a visual scene develop better symbolic languages when they reason over a knowledge graph—objects as nodes, spatial relations as edges—than over raw image features. The graph is built by segmenting the image, embedding the top objects, and connecting each object to its nearest spatial neighbours, with attention weights marking task-relevant nodes. In a referential game, the speaker sends a discrete message from this graph embedding and the listener picks the matching scene, trained end-to-end with a Gumbel-Softmax relaxation. The paper reports that VAG-EC beats the original emergent-communication baseline in Topographic Similarity—similar scenes yielding similar messages—and Context Independence—tokens keeping stable meanings—at vocabulary sizes 10, 20, and 80, with Context Independence up 30–60%, while using the vocabulary more evenly. The practical claim is that this gives a fast, compact, interpretable symbolic channel for real-time assistive feedback to visually impaired users, where natural language is too slow and prior symbolic protocols are too shallow.

What carries the argument

The load-bearing mechanism is the attention-augmented knowledge-graph encoder shared by speaker and listener. An input image is segmented with the Segment Anything Model, the top-N objects become graph nodes carrying CNN-derived embeddings, edges connect each node to its nearest spatial neighbours, and a Graph Convolutional Network computes node features that are pooled into a single scene embedding $J = \sum_j w_j h_j$, where $w_j$ is an attention weight reflecting task-relevant salience. This object-relation structure is what, in the paper's account, pre-factorises the semantic space so that discrete tokens align with entities and relations instead of raw visual patterns. The speaker and listener are separate GRUs trained end-to-end through a Gumbel-Softmax relaxation of the discrete message in the Lewis signaling game, which is what lets the whole graph-to-symbol pipeline be optimized by gradient descent.

What would settle it

Recompute TopSim and Context Independence with concept embeddings from a frozen, pretrained vision encoder that was not trained with the speaker, and run the comparison on real-world dining images; if VAG-EC's advantage over the baseline shrinks or disappears, the reported structure and stability are artifacts of aligning the message space to the model's own graph embeddings.

Watch

Extended reading notes

Core claim

The central claim is that grounding emergent communication in graph-structured cognitive maps changes the character of the learned language, not just its accuracy: messages become compositional over object–relation elements and stable across contexts. VAG-EC encodes each scene as a knowledge graph—nodes from Segment Anything Model (SAM) segmentation and CNN embeddings, edges from spatial proximity—applies attention-weighted GCN encoding, and has a GRU speaker emit discrete tokens that a GRU listener matches to candidate graphs in a Lewis signaling game; Gumbel-Softmax makes the discrete channel trainable. On synthetic dining-scene data generated by a diffusion model, with vocabulary sizes 10, 20, and 80, the paper reports that VAG-EC exceeds the original EC baseline in Topographic Similarity and Context Independence at every vocabulary size, with a 30–60% Context Independence improvement, and matches or improves accuracy at small vocabularies. The token statistics back the qualitative story: a flatter Zipf curve, slower 90% coverage, and more even frequency distribution indicate less token collapse and fuller use of the symbol inventory. The authors present this as evidence that attention-guided reasoning over structured object relations yields emergent languages that are compact, interpretable, and suitable for real-time assistive feedback to visually impaired users.

Load-bearing premise

The gains assume that TopSim and Context Independence, computed from concept embeddings, measure how well a person would understand the symbols; if those embeddings come from the same graph encoder trained jointly with the speaker, the metrics can reward internal self-consistency rather than human interpretability.

Editorial extensions

If this is right

  • If the gains hold, assistive devices could encode a scene into a short string of tokens for haptic or audio presentation, because VAG-EC keeps messages compact while preserving object-relation meaning.
  • Small-vocabulary operation becomes practical: the reported Context Independence gain means tokens do not become ambiguous when the symbol inventory is tight, which is exactly the regime of constrained tactile channels.
  • The emergent code is more evenly distributed, so downstream interfaces can map a stable set of symbols to fixed meanings rather than coping with a few overloaded tokens.
  • Because the method trains on synthetic dining scenes and tests on real-world images, the framework offers a route to data-efficient assistive communication systems for domains where labeled real data are scarce.

Reading between the lines

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

  • A decisive next test the paper does not run is whether the symbols are decodable by humans: TopSim and Context Independence are distributional measures, so an outside probe using a frozen pretrained encoder or human listeners would separate genuine interpretability from the model's internal self-consistency.
  • Replacing generic CNN node features with semantic labels or functional attributes would likely make the emergent tokens map onto human words and could enable zero-shot transfer to scenes whose object categories were never seen in training.
  • Pooling multiple domains into one vocabulary is the natural stress test: the paper's own limitations note that broader scenes may blur relational cues, so a mixed-domain referential game would show whether object-relation symbols remain stable across bathrooms, kitchens, and corridors.
  • The same attention-weighted graph encoding could serve other time-critical symbolic interfaces, such as navigation alerts or emergency signals, wherever a short unambiguous message must beat natural-language latency.
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 / 4 minor

Summary. The paper proposes VAG-EC, an emergent communication framework for assistive navigation for visually impaired users. Images are converted into knowledge graphs via SAM segmentation and CNN features, encoded by an attention-augmented GCN, and then used by GRU-based speaker and listener agents in a referential game trained with Gumbel-Softmax relaxation. The paper claims that this graph-structured, attention-based pipeline outperforms an 'Original EC' baseline in Topographic Similarity (TopSim), Context Independence (CI), and Accuracy across vocabulary sizes 10, 20, and 80, and that it yields more Zipf-like and less collapsed token distributions. The evaluation is based on a synthetic dining dataset generated by diffusion models and real-world dining images used for testing.

Significance. The cognitive motivation is appealing: representing scenes as knowledge graphs with attention is a reasonable way to inject structure into emergent communication, and the paper is honest about its limitations, including the absence of human-in-the-loop evaluation. If the empirical claims were substantiated, the framework would be a useful step toward low-latency symbolic communication for assistive systems. However, the current evidence is not verifiable: the baseline is unspecified, the TopSim/CI metrics are defined relative to an unspecified concept space, and the only quantitative figure lacks error bars and contains numeric labels that do not clearly match the stated 30-60% improvement. The promise of code availability is a positive, but the code URL is inconsistent between the abstract and the full text.

major comments (4)
  1. [Section 6.1 and Figure 3(d)] The central quantitative claim is not supported by the figure as printed. The text states that VAG-EC gives a '30-60% CI improvement across all setups,' but with the most natural row grouping of the printed values, the CI improvements are approximately 61% for V=10, 23% for V=20, and 29% for V=80; only the V=10 case falls in the stated range. The figure also has no error bars, confidence intervals, or seed counts, so the observed differences cannot be distinguished from run-to-run noise. Please report per-cell means with variance or per-seed values, make the label-to-cell mapping explicit, and reconcile the text with the figure.
  2. [Section 6.1, 'Original EC' baseline] The baseline is never defined. The reader is not told whether it uses raw pixels, CNN features, or learned embeddings, nor its architecture, capacity, or training procedure. Without this information, the comparison cannot be reproduced and the reported gains cannot be attributed to knowledge-graph construction or attention rather than to model size or optimization differences. Please specify the baseline fully and, ideally, include ablations that isolate graph structure and attention.
  3. [Appendix B, Eqs. (7) and (8)] TopSim and CI are defined in terms of 'concepts', 'concept embeddings', and conditional probabilities p_m and p_c, but the manuscript never states where these concepts and embeddings come from or how the probabilities are estimated. If the concept embeddings are produced by the same GCN/attention encoder that is jointly trained with the speaker through the referential-game loss in Eq. (5), then the TopSim distance matrix and the CI mapping are partly aligned by the training objective itself, so the reported gains may reflect self-consistency rather than external interpretability. Please specify the provenance of the concept embeddings (e.g., fixed pretrained features or human-labeled concepts) and the estimation procedure for Eq. (7), and recompute the metrics with an external grounding if necessary.
  4. [Sections 5 and 8, experimental protocol] The experimental protocol is under-specified. The synthetic dataset is described only by a prompt template, with no statement of the number of images, number of prompts, train/test split, or diversity statistics, and the real-world test set is only referenced via Figure 5. Section 8 states that all evaluation uses simulated agents, so none of the reported Accuracy, TopSim, or CI numbers have been validated against human interpretability. Please provide dataset statistics, seed information, and hyperparameter values, including the graph construction choices top-N and n-nearest-neighbors, or clearly state which numbers are preliminary.
minor comments (4)
  1. [Section 8 and Appendix B, naming consistency] Section 8 refers to 'KG-AEC' and Appendix B to 'GESC', while the rest of the paper uses 'VAG-EC'; the abstract and full text also give different GitHub URLs. Please harmonize the framework name and repository link.
  2. [Abstract and Section 6] The abstract says experiments vary 'message lengths', but Section 6 fixes L=10 and varies only vocabulary size; either add an L-sweep or correct the claim.
  3. [Figure 3(a)-(c)] The panels would benefit from explicit axis labels and legends; the x-axis of panel (c) is 'Token ID' while panels (a) and (b) use rank and cumulative counts, making direct comparison harder than necessary.
  4. [Throughout] There are minor typos, including 'menioned' in Section 2.2 and the garbled inline text 'cup, next to ← − − −forks' in Section 6.1, and some notation is defined only loosely, such as the term 'concepts' in Eq. (7). Please proofread and clarify.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains are empirical comparisons, and the unspecified origin of concept embeddings is a reproducibility concern, not a circular reduction.

full rationale

The central claim of the paper is empirical: VAG-EC is compared with an "Original EC" baseline on Accuracy, TopSim, and CI (Section 6.1, Figure 3(d)). No equation in the paper defines a claimed result in terms of its own input. TopSim (Eq. 8) and CI (Eq. 7) are standard post-hoc metrics computed over the trained models' outputs; they are not loss terms, and no parameter is fitted to them and then re-reported as a prediction. The paper contains no load-bearing self-citations: the cited graph, vision, and EC works (e.g., Kirillov et al. for SAM, Pope et al. for GCN, Lazaridou et al. for TopSim) are external, and no "uniqueness theorem" is invoked. Two missing pieces—the definition of the "concept embeddings" used in Eq. 8 and the architecture of the baseline—are genuine reproducibility gaps and could make the TopSim/CI comparison unfair if the embeddings are the same jointly trained GCN outputs used by the speaker/listener. However, the paper does not state that identity, so claiming circularity would require speculation. Section 8 explicitly concedes that no human-in-the-loop evaluation was performed, which limits the assistive framing but does not make the comparative EC experiment circular. The paper's asserted 30-60% CI improvement is also difficult to verify from the figure's labels, and the baseline is underspecified; these are correctness and reproducibility risks, not circular reductions. Under the hard rule that circularity must be exhibited by a specific reduction, the honest finding is no significant circularity (score 0).

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

The central claim rests on a chain of modeling choices: the graph construction, the synthetic training distribution, and the evaluation metrics. These are not independently validated, and the evaluation metrics are computed on the model's own representations, which is why the circularity burden is high.

free parameters (4)
  • Gumbel-Softmax temperature tau = 1.0
    Temperature chosen by hand to balance training stability and approximation quality; Section 4.4.
  • Message length L = 10
    Fixed in all experiments, Section 6.
  • Graph construction hyperparameters (top-N, n-nearest neighbors) = unspecified
    Number of object nodes and edge neighbors are not reported; Section 3.
  • Diffusion prompt categories = food, drink, tableware, directions
    Hand-designed prompt vocabulary defines the synthetic training distribution; Section 5 and Appendix A.
assumptions (4)
  • domain assumption SAM segmentation and CNN features yield semantically meaningful object nodes
    Section 3 assumes segments correspond to objects and that fixed CNN embeddings capture task-relevant visual semantics.
  • domain assumption Synthetic diffusion images generalize to real-world dining scenes
    Training is on synthetic images, testing on real images; no analysis of domain shift, Section 5.
  • ad hoc to paper TopSim and CI computed on the model's own embeddings are valid proxies for communication quality
    Appendix B defines metrics on concept embeddings and model conditionals without external ground truth.
  • ad hoc to paper Referential game success transfers to real-time assistive communication
    The paper claims low latency and tactile applicability but never measures either; limitation is acknowledged in Section 8.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Cognitively-Inspired Emergent Communication via Knowledge Graphs for Assisting the Visually Impaired." pith.science (2026). https://pith.science/paper/JBMV5K4U

@misc{pith2026250522087,
  author       = {Pith},
  title        = {Pith review of: Cognitively-Inspired Emergent Communication via Knowledge Graphs for Assisting the Visually Impaired},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JBMV5K4U}},
  note         = {Machine review of arXiv:2505.22087}
}
read the original abstract

Assistive systems for visually impaired individuals must deliver rapid, interpretable, and adaptive feedback to facilitate real-time navigation. Current approaches face a trade-off between latency and semantic richness: natural language-based systems provide detailed guidance but are too slow for dynamic scenarios, while emergent communication frameworks offer low-latency symbolic languages but lack semantic depth, limiting their utility in tactile modalities like vibration. To address these limitations, we introduce a novel framework, Cognitively-Inspired Emergent Communication via Knowledge Graphs (VAG-EC), which emulates human visual perception and cognitive mapping. Our method constructs knowledge graphs to represent objects and their relationships, incorporating attention mechanisms to prioritize task-relevant entities, thereby mirroring human selective attention. This structured approach enables the emergence of compact, interpretable, and context-sensitive symbolic languages. Extensive experiments across varying vocabulary sizes and message lengths demonstrate that VAG-EC outperforms traditional emergent communication methods in Topographic Similarity (TopSim) and Context Independence (CI). These findings underscore the potential of cognitively grounded emergent communication as a fast, adaptive, and human-aligned solution for real-time assistive technologies. Code is available at https://github.com/Anonymous-NLPcode/Anonymous_submission/tree/main.

Figures

Figures reproduced from arXiv: 2505.22087 by the authors.

Figure 1
Figure 1. Pipeline for constructing a knowledge graph from a dining image. The input image is segmented [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed VAG-EC framework. Visual scenes are segmented and converted into knowledge [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Quantitative comparison between the baseline EC and our proposed [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Example of a generated dining scenario from [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Example of a real-world dining scenario used [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

51 extracted references · 39 canonical work pages

  1. [1]

    URL: " 'urlintro :=

    ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year eprint doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRINGS urlintro eprinturl eprintpr...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Peter Anderson, Qi Wu, Damien Teney, Jake Bruce, Mark Johnson, Niko Sünderhauf, Ian Reid, Stephen Gould, and Anton van den Hengel. 2018. https://doi.org/10.1109/CVPR.2018.00387 Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments . In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition...

    arXiv 2018

  4. [4]

    Behrens, Timothy H

    Timothy E.J. Behrens, Timothy H. Muller, James C.R. Whittington, Shirley Mark, Alon B. Baram, Kimberly L. Stachenfeld, and Zeb Kurth-Nelson. 2018. https://doi.org/https://doi.org/10.1016/j.neuron.2018.10.002 What is a cognitive map? organizing knowledge for flexible behavior . Neuron, 100(2):490--509

    doi:10.1016/j.neuron.2018.10.002 2018

  5. [5]

    Irving Biederman. 1987. https://doi.org/10.1037/0033-295X.94.2.115 Recognition-by-components: A theory of human image understanding . Psychological Review, 94(2):115--147

    doi:10.1037/0033-295x.94.2.115 1987

  6. [6]

    Brendon Boldt and David Mortensen. 2024. A review of the applications of deep learning-based emergent communication. Trans. Mach. Learn. Res

    work page 2024

  7. [7]

    Boaz Carmeli, Ron Meir, and Yonatan Belinkov. 2023. https://doi.org/10.1609/aaai.v37i10.26363 Emergent quantized communication . Proceedings of the AAAI Conference on Artificial Intelligence, 37(10):11533--11541

    work page doi:10.1609/aaai.v37i10.26363 2023

  8. [8]

    Rahma Chaabouni, Eugene Kharitonov, Diane Bouchacourt, Emmanuel Dupoux, and Marco Baroni. 2020. https://doi.org/10.18653/v1/2020.acl-main.407 Compositionality and generalization in emergent languages . In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4427--4442, Online. Association for Computational Linguistics

    doi:10.18653/v1/2020.acl-main.407 2020

Show all 51 references
  1. [9]

    Rahma Chaabouni, Eugene Kharitonov, Emmanuel Dupoux, and Marco Baroni. 2019. Anti-efficient encoding in emergent communication. Curran Associates Inc., Red Hook, NY, USA

    2019

  2. [10]

    Rahma Chaabouni, Florian Strub, Florent Altch \'e , Eugene Tarassov, Corentin Tallec, Elnaz Davoodi, Kory Wallace Mathewson, Olivier Tieleman, Angeliki Lazaridou, and Bilal Piot. 2022. Emergent communication at scale. In International Conference on Learning Representations

    2022

  3. [11]

    Henry Conklin and Kenny Smith. 2023. https://openreview.net/forum?id=-Yzz6vlX7V- Compositionality with variation reliably emerges in neural networks . In The Eleventh International Conference on Learning Representations

    2023

  4. [12]

    Roberto Dess\` , Eugene Kharitonov, and Marco Baroni. 2021. Interpretable agent communication from scratch (with a generic visual processor emerging on the side). In Proceedings of the 35th International Conference on Neural Information Processing Systems, NIPS '21, Red Hook, ...

    2021

  5. [13]

    Epstein, Eva Zita Patai, Joshua B

    Russell A. Epstein, Eva Zita Patai, Joshua B. Julian, and Hugo J. Spiers. 2017. https://doi.org/10.1038/nn.4656 The cognitive map in humans: spatial navigation and beyond . Nature Neuroscience, 20(11):1504--1513

    2017 doi

  6. [14]

    Yicheng Feng, Boshi An, and Zongqing Lu. 2024. https://doi.org/10.1609/aaai.v38i16.29685 Learning multi-object positional relationships via emergent communication . Proceedings of the AAAI Conference on Artificial Intelligence, 38(16):17371--17379

    2024 doi

  7. [15]

    mental maps

    Bernard Guelton. 2023. https://doi.org/10.3389/fpsyg.2023.1142238 “mental maps”: Between memorial transcription and symbolic projection . Frontiers in Psychology, Volume 14 - 2023

    2023

  8. [16]

    Shangmin Guo, Yi Ren, Serhii Havrylov, Stella Frank, Ivan Titov, and Kenny Smith. 2019. http://arxiv.org/abs/1910.05291 The emergence of compositional languages for numeric concepts through iterated learning in neural agents . CoRR, abs/1910.05291

    2019 arXiv

  9. [17]

    Kai Han, Yunhe Wang, Jianyuan Guo, Yehui Tang, and Enhua Wu. 2022. https://openreview.net/forum?id=htM1WJZVB2I Vision GNN : An image is worth graph of nodes . In Advances in Neural Information Processing Systems

    2022

  10. [18]

    Andreas Holzinger. 2022. https://doi.org/10.1007/s10462-022-10246-w Knowledge-enhanced machine learning and the future of artificial intelligence . Artificial Intelligence Review, 55:4273--4300

    2022 doi

  11. [19]

    Toru Ishikawa. 2021. Spatial thinking, cognitive mapping, and spatial awareness. Cognitive Processing, 22:89 -- 96

    2021

  12. [20]

    Berg, Wan-Yen Lo, Piotr Doll \'a r, and Ross Girshick

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C. Berg, Wan-Yen Lo, Piotr Doll \'a r, and Ross Girshick. 2023. Segment anything. arXiv:2304.02643

    2023 arXiv

  13. [21]

    ukasz Kuci\' n ski, Tomasz Korbak, Pawe Ko odziej, and Piotr Mi o\' s . 2021. Catalytic role of noise and necessity of inductive biases in the emergence of compositional communication. In Proceedings of the 35th International Conference on Neural Information Processing Systems...

    2021

  14. [22]

    Shuhei Kurita and Kyunghyun Cho. 2020. Generative language-grounded policy in vision-and-language navigation with bayes' rule. ArXiv, abs/2009.07783

    2020 arXiv

  15. [23]

    Angeliki Lazaridou, Karl Moritz Hermann, Karl Tuyls, and Stephen Clark. 2018. Emergence of linguistic communication from referential games with symbolic and pixel input. In International Conference on Learning Representations

    2018

  16. [24]

    Heeyoung Lee. 2024. https://doi.org/10.18653/v1/2024.emnlp-main.1157 One-to-many communication and compositionality in emergent communication . In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 20794--20811, Miami, Florida, USA. A...

    2024 doi

  17. [25]

    David Lewis. 1986. Convention: A philosophical study

    1986

  18. [26]

    Jinhui Li, Qunjun Liang, Jiajun Liao, Senning Zheng, Kemeng Chen, and Ruiwang Huang. 2023. https://doi.org/https://doi.org/10.1002/hbm.26309 Representation of the inferred relationships in a map-like space . Human Brain Mapping, 44(9):3744--3757

    2023 doi

  19. [27]

    Chunxiao Liu, Zhendong Mao, Tianzhu Zhang, Hongtao Xie, Bin Wang, and Yongdong Zhang. 2020. https://doi.org/10.1109/CVPR42600.2020.01093 Graph structured network for image-text matching . In 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10918--10927

    2020

  20. [28]

    Jesse Mu and Noah Goodman. 2021. Emergent communication of generalizations. In Advances in Neural Information Processing Systems, volume 34, pages 17994--18007

    2021

  21. [29]

    Mustafa Munir, William Avery, and Radu Marculescu. 2023. Mobilevig: Graph-based sparse attention for mobile vision applications. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 2211--2219

    2023

  22. [30]

    Mustafa Munir, William Avery, Md Mostafijur Rahman, and Radu Marculescu. 2024. Greedyvig: Dynamic axial graph construction for efficient vision gnns. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 6118--6127

    2024

  23. [31]

    Quang Ho Nguyen, Truong Tuan Vu, Anh Tuan Tran, and Khoi Nguyen. 2023. https://openreview.net/forum?id=StD4J5ZlI5 Dataset diffusion: Diffusion-based synthetic data generation for pixel-level semantic segmentation . In Thirty-seventh Conference on Neural Information Processing Systems

    2023

  24. [32]

    Mitja Nikolaus. 2024. https://openreview.net/forum?id=Sy8upuD6Bw Emergent communication with conversational repair . In The Twelfth International Conference on Learning Representations

    2024

  25. [33]

    Makanjuola Adekunmi Ogunleye, Chase Vickery, and Ismini Lourentzou. 2024. https://openreview.net/forum?id=vSkcS3qnZk Emergent corpus pretraining benefits vision language modeling

    2024

  26. [34]

    Xenia Ohmer, Marko Duda, and Elia Bruni. 2022. https://aclanthology.org/2022.coling-1.501/ Emergence of hierarchical reference systems in multi-agent communication . In Proceedings of the 29th International Conference on Computational Linguistics, pages 5689--5706, Gyeongju, R...

    2022

  27. [35]

    Brunec, Nora S

    Michael Peer, Iva K. Brunec, Nora S. Newcombe, and Russell A. Epstein. 2021. https://doi.org/https://doi.org/10.1016/j.tics.2020.10.004 Structuring knowledge with cognitive maps and cognitive graphs . Trends in Cognitive Sciences, 25(1):37--54

    2021 doi

  28. [36]

    Jannik Peters, Constantin Waubert de Puiseau, Hasan Tercan, Arya Gopikrishnan, Gustavo Adolpho Lucas de Carvalho, Christian Bitter, and Tobias Meisen. 2025. https://doi.org/10.1007/s10458-025-09691-y Emergent language: a survey and taxonomy . Autonomous Agents and Multi-Agent ...

    2025 doi

  29. [37]

    Pope, Soheil Kolouri, Mohammad Rostami, Charles E

    Phillip E. Pope, Soheil Kolouri, Mohammad Rostami, Charles E. Martin, and Heiko Hoffmann. 2019. Explainability methods for graph convolutional neural networks. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    2019

  30. [38]

    Ilya Reutov. 2023. https://doi.org/https://doi.org/10.1016/j.procs.2023.12.036 Generating of synthetic datasets using diffusion models for solving computer vision tasks in urban applications . Procedia Computer Science, 229:335--344. 12th International Young Scientists Confere...

    2023 doi

  31. [39]

    Ryokan Ri, Ryo Ueda, and Jason Naradowsky. 2023. https://escholarship.org/uc/item/7dg8r8zk Emergent communication with attention . In CogSci

    2023

  32. [40]

    Mathieu Rita, Corentin Tallec, Paul Michel, Jean-Bastien Grill, Olivier Pietquin, Emmanuel Dupoux, and Florian Strub. 2022 a . https://proceedings.neurips.cc/paper_files/paper/2022/file/093b08a7ad6e6dd8d34b9cc86bb5f07c-Paper-Conference.pdf Emergent communication: Generalizatio...

    2022

  33. [41]

    Mathieu Rita, Corentin Tallec, Paul Michel, Jean-Bastien Grill, Olivier Pietquin, Emmanuel Dupoux, and Florian Strub. 2022 b . https://openreview.net/forum?id=qqHMvHbfu6 Emergent communication: Generalization and overfitting in lewis games . In Advances in Neural Information P...

    2022

  34. [42]

    Damien Teney, Lingqiao Liu, and Anton van den Hengel. 2017. http://arxiv.org/abs/1609.05600 Graph-structured representations for visual question answering

    2017 arXiv

  35. [43]

    Ryo Ueda and Tadahiro Taniguchi. 2024. http://arxiv.org/abs/2311.04453 Lewis's signaling game as beta-vae for natural word lengths and segments

    2024 arXiv

  36. [44]

    Ryo Ueda and Koki Washio. 2021. https://doi.org/10.18653/v1/2021.acl-srw.6 On the relationship between Z ipf`s law of abbreviation and interfering noise in emergent languages . In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 1...

    2021 doi

  37. [45]

    Ruben van Bergen, Justus Hübotter, and Pablo Lanillos. 2025. http://arxiv.org/abs/2411.17438 Object-centric proto-symbolic behavioural reasoning from pixels

    2025

  38. [46]

    Taylor Webb, Shanka Subhra Mondal, and Jonathan D Cohen. 2023. https://proceedings.neurips.cc/paper_files/paper/2023/file/e3cdc587873dd1d00ac78f0c1f9aa60c-Paper-Conference.pdf Systematic visual reasoning through object-centric relational abstraction . In Advances in Neural Inf...

    2023

  39. [47]

    Martin Weiss, Simon Chamorro, Roger Girgis, Margaux Luck, Samira Ebrahimi Kahou, Joseph Paul Cohen, Derek Nowrouzezahrai, Doina Precup, Florian Golemo, and Christopher Joseph Pal. 2019. Navigation agents for the visually impaired: A sidewalk simulator and experiments. In Confe...

    2019

  40. [48]

    World Health Organization . 2023. https://www.who.int/news-room/fact-sheets/detail/blindness-and-visual-impairment Blindness and vision impairment

    2023

  41. [49]

    Zhenlin Xu, Marc Niethammer, and Colin Raffel. 2022. https://openreview.net/forum?id=ZEQ5Gf8DiD Compositional generalization in unsupervised compositional representation learning: A study on disentanglement and emergent language . In Advances in Neural Information Processing Systems

    2022

  42. [50]

    Ling Yang, Zhilong Zhang, Yang Song, Shenda Hong, Runsheng Xu, Yue Zhao, Wentao Zhang, Bin Cui, and Ming-Hsuan Yang. 2023. https://doi.org/10.1145/3626235 Diffusion models: A comprehensive survey of methods and applications . ACM Comput. Surv., 56(4)

    2023 doi

  43. [51]

    Enshuai Zhou, Yifan Hao, Rui Zhang, Yuxuan Guo, Zidong Du, Xishan Zhang, Xinkai Song, Chao Wang, Xuehai Zhou, Jiaming Guo, Qi Yi, Shaohui Peng, Di Huang, Ruizhi Chen, Qi Guo, and Yunji Chen. 2024. https://doi.org/10.1609/aaai.v38i16.29712 Emergent communication for numerical c...

    2024 doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.