Pith. sign in

REVIEW 3 major objections 6 minor 28 references

Cross-Modality Controlled Molecule Generation with Diffusion Language Model

T0 review · 3 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Hyperdimensional refinement sharpens LLM reasoning graphs for video anomaly detection

desk verdict MISSIONHD is a genuinely new HDC-based graph refinement idea with mostly positive per-threshold results, but the headline mAUC is a test-set maximum and the theory has an empirically fitted term. read the letter →

arxiv 2508.14748 v1 pith:BN4NCN7F submitted 2025-08-20 cs.LG cs.AI

Yunzhe Zhang , Yifei Wang , Khanh Vinh Nguyen , Pengyu Hong This is my paper · ORCID
classification cs.LGcs.AI
keywords hyperdimensionalcomputinggraphstructurerefinementvideoanomalydetectionLLM-generatedreasoninggraphsweaklysupervisedlearningedgedecodingmission-specific
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 proposes a new way to refine LLM-generated reasoning graphs for weakly supervised video anomaly detection and recognition. It argues that these graphs, called mission-specific graphs, are generic and lack a learnable structural distribution, so conventional graph refinement methods fail. The solution, HDC-GSR, encodes the graph into a single high-dimensional hypervector, aligns that vector with the downstream task loss, then decodes it back into an improved graph. Instantiating this as MISSIONHD, the paper reports consistent performance gains on UCF-Crime and XD-Violence, including a mAUC improvement of up to +9.66% on UCF-Crime. The central claim is that task supervision alone can drive meaningful graph structure refinement without modeling any graph distribution.

What carries the argument

The central mechanism is hyperdimensional computing (HDC), where graph structure is encoded through binding and bundling operations into a single high-dimensional graph hypervector. Binding implements message passing, bundling implements aggregation, and the resulting code is aligned with the task loss through a learnable edit hypervector. Decoding uses forward and backward path encodings under a fully-connected assumption, constructs a compositional probe for each candidate edge, and scores it by softmax-normalized cosine similarity against the trained graph code. This encode–align–decode loop is what lets structure refinement happen without learning a graph distribution.

What would settle it

A direct test would ablate the learnable edit hypervector by setting it to zero after training and re-decoding the graph: if the refined edge set is largely unchanged, then the refinement signal is not coming from task alignment. Conversely, a random-rewiring experiment that permutes edge labels between classes while keeping graph statistics fixed, and still yields large refinement gains, would suggest the gains come from sparsity artifacts rather than semantic graph structure.

Watch

Extended reading notes

Core claim

The paper's central claim is that LLM-generated reasoning graphs can be effectively refined for downstream vision tasks without learning a structural distribution, by directly optimizing a decodable graph representation in a single hyperdimensional space. The method bundles paths of the graph into a graph hypervector, adds a learnable structural edit hypervector, and optimizes the combined code with the downstream anomaly-detection loss. After training, candidate edges are scored by the cosine similarity between a compositional edge probe and the optimized graph code, and edges above a threshold form the refined graph. The paper shows consistent gains across two surveillance-video benchmarks

Load-bearing premise

The method assumes that a single learnable edit hypervector, bundled into the graph code, can capture all task-relevant structural changes, and that new edges not present in the original graph can still be scored reliably through this edit vector and noise; if the edit vector is too weak or the decoding is too noisy, the refined graph will not generalize.

Editorial extensions

If this is right

  • If HDC-GSR is correct, LLM-generated reason-ing graphs can be adapted to downstream tasks cheaply, without retraining the LLM or learning large graph datasets.
  • The method provides a principled path for graph refinement wherever graphs are generated as novel artifacts rather than sampled from a stable distribution.
  • One round of refinement often suffices, suggesting that a single learnable edit hypervector can capture most task-relevant structural changes for layered DAGs.
  • The framework extends beyond video anomaly detection to any task requiring structured, interpretable reasoning over learned knowledge graphs.
  • The theoretical connection to path-sum GNNs indicates that HDC-style encoding can serve as a lightweight substitute for message-passing layers in constrained settings.

Reading between the lines

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

  • The method's success on randomly augmented initial graphs suggests that HDC-GSR may be more sensitive to task-signal alignment than to the quality of the initial graph; this could make it useful for refining graphs generated by weaker or cheaper LLMs.
  • The authors' claim that one refinement round is enough hints at a capacity ceiling of the single edit hypervector; scaling the edit representation, or making it path-dependent, may unlock further gains on more complex graphs.
  • The decoding threshold acts as a sparsity knob; the observed stability around threshold 0.2 suggests that edge-score margins are fairly polarized after training, which could be tested on other tasks.
  • The framework's reliance on layered DAGs is a clear boundary; extending the encoding to cycles or cross-layer edges would require a different path-bundling scheme and is a natural next step.
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

3 major / 6 minor

Summary. The manuscript, arXiv:2508.14748 (cs.LG), presents MISSIONHD, a method for refining LLM-generated mission-specific reasoning graphs for weakly supervised video anomaly detection and recognition. The proposed HDC-GSR paradigm encodes graph paths into a hyperdimensional vector, optimizes this representation with a downstream task loss, and decodes edge contributions by thresholding softmax-normalized cosine similarities. The authors claim consistent performance gains over unrefined graphs and several baselines on UCF-Crime and XD-Violence, and they provide a theoretical analysis connecting the encoding to path-sum GNNs with a decoding accuracy bound.

Significance. If the empirical claims are reliable, the work offers a genuinely different angle on graph structure refinement: instead of learning a structural distribution, it optimizes a decodable hyperdimensional representation directly with task supervision. This could be useful for small or one-off graphs where distributional assumptions are untenable. The paper also ships a relatively detailed algorithmic description and qualitative analysis of the refined graphs. However, the evaluation as reported does not support the headline 'consistent performance gains' with the claimed precision, because the best test-set threshold is selected post hoc and no variance estimates are provided. The theoretical bound in Proposition 4.4 is partly empirical, so it does not fully compensate for the evaluation weaknesses.

major comments (3)
  1. [Section 5.2] The central empirical claim is undermined by test-set threshold selection. The text states: 'Refined graphs were collected under varying thresholds, and we selected those corresponding to the threshold that yielded the highest mAUC improvement.' Thus the reported +9.66% (UCF-Crime) and +0.56% (XD-Violence) are the best of four test-set thresholds, not fixed-procedure results. Tables 3 and 4 do show positive mAUC gains at every threshold, which weakens a pure selection-artifact reading, but the claimed magnitudes and the comparisons in Tables 1/2 are not based on a pre-registered threshold. Please report results with a validation-based threshold selection or, at minimum, present all thresholds as the primary result without selecting the best.
  2. [Section 5.1 / Tables 1-4] No error bars, confidence intervals, or multiple seeds are reported for any of the quantitative results. The only ablations (Figure 5) also appear to be single runs. Since the gains are often small (e.g., +0.56% mAUC on XD-Violence) and the baseline differences are within a few points, the conclusions require an estimate of run-to-run variability. Please provide at least 3-5 seeds with standard deviations/confidence intervals, or state if the protocol is deterministic and justify that claim.
  3. [Proposition 4.4] The decoding accuracy bound is not a fully derived guarantee. The bound contains an 'empirically estimated' damping factor d(T) and an arbitrary finite set of newly synthesized paths P added to approximate the learned edit vector w_e. The dependence on P and on an empirical constant means the result does not provide a priori error control; it is closer to a consistency argument under assumptions that are not fully specified (Assumptions A3, for instance, is a JL-type projection assumption without a concrete map). This is load-bearing because the theoretical section is invoked to support the decoding scheme. Please either strengthen the bound to remove the empirically estimated quantity or clearly state which parts of the analysis are empirical.
minor comments (6)
  1. [Title/Abstract] The manuscript title and the paper body are for 'MISSIONHD: Hyperdimensional Refinement of Distribution-Deficient Reasoning Graphs for Video Anomaly Detection,' but the abstract at the top of the file describes a different paper, 'Cross-Modality Controlled Molecule Generation with Diffusion Language Model (CMCM-DLM).' This mismatch must be corrected; in the current form the manuscript cannot be properly indexed or reviewed.
  2. [Section 5.2] The sentence 'All methods improve performance' is contradicted by Table 1, where several baseline cells show negative deltas (e.g., GCN on Arson is -2.43; ECL-GSR on Shoplifting is -8.64). Please rephrase to reflect the actual pattern, or use a different summary statistic.
  3. [Section 2.1 / Appendix B] There are numerous OCR/garbled symbols throughout (e.g., 'H\u211d', '\u2208', '�'). In particular, the HDC definitions and Algorithm 1 contain illegible characters. The authors should provide a cleanly compiled version.
  4. [References] References [26] and [27] are the same paper (Zeng et al., AAAI 2025). Please deduplicate.
  5. [Algorithm 1] Algorithm 1 lists 'Graph edit hypervectors: w_e' as an input but never uses w_e in the encoding loop; w_e is introduced later in Eq. (2) as part of the trained graph vector. The pseudocode should be consistent with the text.
  6. [Table 3 / Table 4] The 'Normal' class is not an anomaly class; including it in the mAUC is nonstandard. Please clarify whether mAUC is computed over all classes including 'Normal' or over anomaly classes only, and justify the choice.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the encode-align-decode derivation is self-contained and the empirical gains are measured outcomes, not constructions from the inputs.

full rationale

The claimed derivation chain is an encode–align–decode procedure: Algorithm 1 builds the graph hypervector G from path bundles, Eq. (2)–(3) train the edit hypervector w_e under the task loss, and Algorithm 3 decodes edge scores via cosine similarity to G⊕w_e. None of these steps defines the output in terms of the reported mAUC or imports an unverified self-citation as the load-bearing premise. Theorems 4.1 and 4.2 are exact algebraic equivalences (diagonal linear maps and path-sum unfolding), not results that presuppose the empirical gains. Proposition 4.4 is an error decomposition whose residual terms are explicitly acknowledged approximations; it does not rename a fitted value as a prediction. Self-citations to MissionGNN [24], GrapHD [13], and related HDC work supply the MSG generator and background operations, but the refinement result is evaluated against external baselines and against the unrefined MSG, so the central claim has independent content. The only notable methodological issue is Section 5.2's test-set threshold selection ('Refined graphs were collected under varying thresholds, and we selected those corresponding to the threshold that yielded the highest mAUC improvement'), which is a statistical validity/overfitting concern rather than circularity: the reported '+9.66% mAUC' is the maximum over four thresholds, although Tables 3 and 4 show positive deltas at every threshold. The paper itself acknowledges 'reliance on thresholding strategies' in the Limitations section. Thus no circular step is present.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The main free parameter is the edge threshold, which is fit to the test set. The damping factor in the theoretical bound is also empirical. The axioms are standard HDC assumptions and domain assumptions about the graph structure and the nature of LLM-generated graphs.

free parameters (3)
  • edge threshold T = 0.2 (selected as best on test set)
    The threshold used for the main reported result (Tables 1 and 2) was selected as the one yielding the highest mAUC on the test set, as stated in Section 5.2. This is a post-hoc selection that inflates results.
  • damping factor d(T) = estimated from observed softmax margins
    In Proposition 4.4, the theoretical bound includes a threshold-dependent damping factor 'estimated from observed softmax margins', which is an empirical quantity fitted to data rather than derived from first principles.
  • hyperspace dimensionality d = 10000 (stated, though rendering is garbled)
    The dimensionality of the hyperdimensional space is chosen by hand; it is a standard hyperparameter but is nonetheless a free choice that affects decoding accuracy.
assumptions (5)
  • standard math Near-orthogonality of random hypervectors (A1)
    Assumed in Appendix A, used to bound residual and crosstalk terms in edge decoding.
  • domain assumption Similarity preservation under binding (A2)
    Assumed in Appendix A, needed for the edge score decomposition in Proposition 4.3.
  • standard math Johnson-Lindenstrauss preservation of projection phi (A3)
    Assumed in Appendix A, although the projection is learned, the JL property is invoked as a given.
  • domain assumption MSGs are layered DAGs with edges only between adjacent layers
    The encoding and decoding algorithms are defined only for this graph type, as stated in Section 2.2 and used throughout Section 3.
  • domain assumption LLM-generated graphs lack a learnable structural distribution
    This motivates the entire method in the introduction; it rules out conventional GSR approaches and is a premise for why HDC-GSR is needed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Cross-Modality Controlled Molecule Generation with Diffusion Language Model." pith.science (2026). https://pith.science/paper/BN4NCN7F

@misc{pith2026250814748,
  author       = {Pith},
  title        = {Pith review of: Cross-Modality Controlled Molecule Generation with Diffusion Language Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BN4NCN7F}},
  note         = {Machine review of arXiv:2508.14748}
}
read the original abstract

Current SMILES-based diffusion models for molecule generation typically support only unimodal constraint. They inject conditioning signals at the start of the training process and require retraining a new model from scratch whenever the constraint changes. However, real-world applications often involve multiple constraints across different modalities, and additional constraints may emerge over the course of a study. This raises a challenge: how to extend a pre-trained diffusion model not only to support cross-modality constraints but also to incorporate new ones without retraining. To tackle this problem, we propose the Cross-Modality Controlled Molecule Generation with Diffusion Language Model (CMCM-DLM), demonstrated by two distinct cross modalities: molecular structure and chemical properties. Our approach builds upon a pre-trained diffusion model, incorporating two trainable modules, the Structure Control Module (SCM) and the Property Control Module (PCM), and operates in two distinct phases during the generation process. In Phase I, we employs the SCM to inject structural constraints during the early diffusion steps, effectively anchoring the molecular backbone. Phase II builds on this by further introducing PCM to guide the later stages of inference to refine the generated molecules, ensuring their chemical properties match the specified targets. Experimental results on multiple datasets demonstrate the efficiency and adaptability of our approach, highlighting CMCM-DLM's significant advancement in molecular generation for drug discovery applications.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 22 canonical work pages

  1. [1]

    Hdreason: Algorithm-hardware codesign for hyperdimensional knowledge graph reasoning.arXiv preprint arXiv:2403.05763, 2024

    Hanning Chen, Yang Ni, Ali Zakeri, Zhuowen Zou, Sanggeon Yun, Fei Wen, Behnam Khaleghi, Narayan Srinivasa, Hugo Latapie, and Mohsen Imani. Hdreason: Algorithm-hardware codesign for hyperdimensional knowledge graph reasoning.arXiv preprint arXiv:2403.05763, 2024

    arXiv 2024

  2. [2]

    Hyperdimensional representation learning for node classification and link prediction

    Abhishek Dalvi and Vasant Honavar. Hyperdimensional representation learning for node classification and link prediction. InProceedings of the Eighteenth ACM International Conference on Web Search and Data Mining, pages 88–97, 2025

    work page 2025

  3. [3]

    Imagebind: One embedding space to bind them all

    Rohit Girdhar, Alaaeldin El-Nouby, Zhuang Liu, Mannat Singh, Kalyan Vasudev Alwala, Armand Joulin, and Ishan Misra. Imagebind: One embedding space to bind them all. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 15180–15190, 2023

    2023

  4. [4]

    Building a mind palace: Structuring environment-grounded semantic graphs for effective long video analysis with llms

    Zeyi Huang, Yuyang Ji, Xiaofang Wang, Nikhil Mehta, Tong Xiao, Donghyun Lee, Sigmund Vanvalkenburgh, Shengxin Zha, Bolin Lai, Licheng Yu, et al. Building a mind palace: Structuring environment-grounded semantic graphs for effective long video analysis with llms. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 24169–24179, 2025

    work page 2025

  5. [5]

    Exploiting boosting in hyperdimensional computing for enhanced reliability in healthcare

    SungHeon Jeong, Hamza Errahmouni Barkam, Sanggeon Yun, Yeseong Kim, Shaahin Angizi, and Mohsen Imani. Exploiting boosting in hyperdimensional computing for enhanced reliability in healthcare. In2025 Design, Automation & Test in Europe Conference (DATE), pages 1–7. IEEE, 2025

    work page 2025

  6. [6]

    Hyperdimensional computing: An introduction to computing in distributed representation with high-dimensional random vectors.Cognitive computation, 1:139–159, 2009

    Pentti Kanerva. Hyperdimensional computing: An introduction to computing in distributed representation with high-dimensional random vectors.Cognitive computation, 1:139–159, 2009

    work page 2009

  7. [7]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. Semi-supervised classification with graph convolutional networks. In International Conference on Learning Representations (ICLR), 2017

    2017

  8. [8]

    A survey on hyperdimensional computing aka vector symbolic architectures, part i: Models and data transformations.ACM Computing Surveys, 55(6):1–40, 2022

    Denis Kleyko, Dmitri A Rachkovskij, Evgeny Osipov, and Abbas Rahimi. A survey on hyperdimensional computing aka vector symbolic architectures, part i: Models and data transformations.ACM Computing Surveys, 55(6):1–40, 2022

    work page 2022

Show all 28 references
  1. [9]

    GraphOTTER: Evolving LLM-based graph reasoning for complex table question answering

    Qianlong Li, Chen Huang, Shuai Li, Yuanxin Xiang, Deng Xiong, and Wenqiang Lei. GraphOTTER: Evolving LLM-based graph reasoning for complex table question answering. In Owen Rambow, Leo Wanner, Marianna Apidianaki, Hend Al-Khalifa, Barbara Di Eugenio, and Steven Schockaert, edi...

    2025

  2. [10]

    Reasoning on graphs: Faithful and interpretable large language model reasoning

    Linhao Luo, Yuan-Fang Li, Gholamreza Haffari, and Shirui Pan. Reasoning on graphs: Faithful and interpretable large language model reasoning. InInternational Conference on Learning Representations, 2024

    2024

  3. [11]

    Unbiased multiple instance learning for weakly supervised video anomaly detection

    Hui Lv, Zhongqi Yue, Qianru Sun, Bin Luo, Zhen Cui, and Hanwang Zhang. Unbiased multiple instance learning for weakly supervised video anomaly detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8022–8031, 2023

    2023

  4. [12]

    Pv-vtt: A privacy-centric dataset for mission-specific anomaly detection and natural language interpretation

    Ryozo Masuakwa, Sanggeon Yun, Yoshiki Yamaguchi, and Mohsen Imani. Pv-vtt: A privacy-centric dataset for mission-specific anomaly detection and natural language interpretation. In2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 6415–6424. IEEE, 2025

    2025

  5. [13]

    Graphd: Graph-based hyperdimensional memorization for brain-like cognitive learning.Frontiers in Neuroscience, 16:757125, 2022

    Prathyush Poduval, Haleh Alimohamadi, Ali Zakeri, Farhad Imani, M Hassan Najafi, Tony Givargis, and Mohsen Imani. Graphd: Graph-based hyperdimensional memorization for brain-like cognitive learning.Frontiers in Neuroscience, 16:757125, 2022

    2022

  6. [14]

    Step: Enhancing video-llms’ compositional reasoning by spatio-temporal graph- guided self-training

    Haiyi Qiu, Minghe Gao, Long Qian, Kaihang Pan, Qifan Yu, Juncheng Li, Wenjie Wang, Siliang Tang, Yueting Zhuang, and Tat-Seng Chua. Step: Enhancing video-llms’ compositional reasoning by spatio-temporal graph- guided self-training. InProceedings of the Computer Vision and Patt...

    2025

  7. [15]

    A vision check-up for language models

    Pratyusha Sharma, Tamar Rott Shaham, Manel Baradad, Stephanie Fu, Adrian Rodriguez-Munoz, Shivam Duggal, Phillip Isola, and Antonio Torralba. A vision check-up for language models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14410...

    2024

  8. [16]

    Real-world anomaly detection in surveillance videos

    Waqas Sultani, Chen Chen, and Mubarak Shah. Real-world anomaly detection in surveillance videos. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 6479–6488, 2018

    2018

  9. [17]

    A theoretical perspective on hyperdimensional computing

    Anthony Thomas, Sanjoy Dasgupta, and Tajana Rosing. A theoretical perspective on hyperdimensional computing. Journal of Artificial Intelligence Research, 72:215–249, 2021. 12 MISSIONHD: Hyperdimensional Refinement of Distribution-Deficient Graphs for V ADA PREPRINT

    2021

  10. [18]

    Weakly- supervised video anomaly detection with robust temporal feature magnitude learning

    Yu Tian, Guansong Pang, Yuanhong Chen, Rajvinder Singh, Johan W Verjans, and Gustavo Carneiro. Weakly- supervised video anomaly detection with robust temporal feature magnitude learning. InProceedings of the IEEE/CVF international conference on computer vision, pages 4975–4986, 2021

    2021

  11. [19]

    Graph attention networks.arXiv preprint arXiv:1710.10903, 2017

    Petar Veliˇckovi´c, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, and Yoshua Bengio. Graph attention networks.arXiv preprint arXiv:1710.10903, 2017

    2017 arXiv

  12. [20]

    Vqa-gnn: Reasoning with multimodal knowledge via graph neural networks for visual question answering

    Yanan Wang, Michihiro Yasunaga, Hongyu Ren, Shinya Wada, and Jure Leskovec. Vqa-gnn: Reasoning with multimodal knowledge via graph neural networks for visual question answering. InProceedings of the IEEE/CVF international conference on computer vision, pages 21582–21592, 2023

    2023

  13. [21]

    Not only look, but also listen: Learning multimodal violence detection under weak supervision

    Peng Wu, jing Liu, Yujia Shi, Yujia Sun, Fangtao Shao, Zhaoyang Wu, and Zhiwei Yang. Not only look, but also listen: Learning multimodal violence detection under weak supervision. InEuropean Conference on Computer Vision (ECCV), 2020

    2020

  14. [22]

    Scene graph generation by iterative message passing

    Danfei Xu, Yuke Zhu, Christopher Choy, and Li Fei-Fei. Scene graph generation by iterative message passing. In Computer Vision and Pattern Recognition (CVPR), 2017

    2017

  15. [23]

    Text prompt with normality guidance for weakly supervised video anomaly detection

    Zhiwei Yang, Jing Liu, and Peng Wu. Text prompt with normality guidance for weakly supervised video anomaly detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 18899–18908, June 2024

    2024

  16. [24]

    Missiongnn: Hierarchical multimodal gnn-based weakly supervised video anomaly recognition with mission-specific knowledge graph generation

    Sanggeon Yun, Ryozo Masukawa, Minhyoung Na, and Mohsen Imani. Missiongnn: Hierarchical multimodal gnn-based weakly supervised video anomaly recognition with mission-specific knowledge graph generation. In 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV...

    2025

  17. [25]

    Delving into clip latent space for video anomaly recognition.Computer Vision and Image Understanding, 249:104163, 2024

    Luca Zanella, Benedetta Liberatori, Willi Menapace, Fabio Poiesi, Yiming Wang, and Elisa Ricci. Delving into clip latent space for video anomaly recognition.Computer Vision and Image Understanding, 249:104163, 2024

    2024

  18. [27]

    Graph structure refinement with energy-based contrastive learning

    Xianlin Zeng, Yufeng Wang, Yuqi Sun, Guodong Guo, Wenrui Ding, and Baochang Zhang. Graph structure refinement with energy-based contrastive learning. InProceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 22326–22335, 2025

    2025

  19. [28]

    Self-supervised graph structure refinement for graph neural networks

    Jianan Zhao, Qianlong Wen, Mingxuan Ju, Chuxu Zhang, and Yanfang Ye. Self-supervised graph structure refinement for graph neural networks. InProceedings of the sixteenth ACM international conference on web search and data mining, pages 159–167, 2023

    2023

  20. [29]

    Graph convolutional label noise cleaner: Train a plug-and-play action classifier for anomaly detection

    Jia-Xing Zhong, Nannan Li, Weijie Kong, Shan Liu, Thomas H Li, and Ge Li. Graph convolutional label noise cleaner: Train a plug-and-play action classifier for anomaly detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1237–124...

    2019

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.