Pith. sign in

REVIEW 5 major objections 5 minor 27 references

Compositional Benchmark Synthesis for Hierarchical Human Action Recognition

T0 review · 5 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper shows that a benchmark synthesized from flat action clips forces all tested model families to lose 0.13–0.17 macro-F1 on new combinations of known intention components, and that the gap is structural because a graph-aware model…

desk verdict A credible compositional-gap result on a thoughtful benchmark-synthesis protocol, but the structural-property claim needs regeneration-seed analysis before it holds. read the letter →

arxiv 2608.10765 v1 pith:HEA6WOJJ submitted 2026-08-11 cs.AI cs.CV

Farnaz Soleimani (LISSI) , Abdelghani Chibani (LISSI) , Yacine Amirat (LISSI) , Ghazaleh Khodabandelou (LISSI) This is my paper · ORCID
classification cs.AIcs.CV
keywords hierarchicalactionrecognitionbenchmarksynthesiscompositionalgeneralizationneuro-symboliclearningfirst-orderlogicconstraintstransformermultimodal
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 a way to build a four-level hierarchical-intention benchmark — actions, activities, low-level intentions, and high-level intentions — from a flat, single-label action corpus, while keeping real pre-extracted features at the action level. The central claim is that this benchmark exposes a structural limitation in current action-recognition models: when new combinations of known intention components are held out from training, every tested model family loses 0.13 to 0.17 macro-F1 at the high-level-intention tier, and a graph-aware model that otherwise recognizes best does not close the gap. Because the drop appears across architectures and is concentrated on genuinely novel compositions rather than distribution shift, the authors conclude it is a property of the compositional task itself. If correct, the work provides a regenerable testbed for controlled evaluation of hierarchical and compositional reasoning about human behavior.

What carries the argument

The load-bearing mechanism is the benchmark synthesis protocol itself: a four-level ontology (actions → activities → low-level intentions → high-level intentions) over 85 action classes, a transition model that assembles subject-consistent episodes, a coverage-aware sampler that reduces the subject-usage Gini from 0.566 to 0.248, and a validity design that keeps the sequence-generation rules disjoint from the 22 evaluation-time first-order-logic Group Y rules. Compositional held-out splits withhold one LLI→HLI association per multi-parent LLI, forcing transfer to unseen high-level contexts. An order-destroying control permutes sibling order while fixing labels, performers, and membership, serving as a generator-consistency check. Together these components make hierarchy, coverage, circularity, and compositional generalization measurable in one setting.

What would settle it

Regenerate the benchmark with an independently authored ontology and rule set produced by a different set of experts; if the compositional held-out gap shrinks to near zero on this second version, the gap is an artifact of the particular semantic standard rather than a structural property of compositional intention recognition.

Watch

Extended reading notes

Core claim

The paper's central discovery is the compositional held-out gap: a consistent drop in high-level-intention macro-F1 of 0.13 to 0.17 when LLI-to-HLI associations are withheld from training, observed across all four reference baselines (hierarchical transformer, sequential transformer, bag-of-actions MLP, and relational graph network) and every seed. The graph-aware R-GCN is the strongest recognizer yet shows the largest gap, which the authors take as evidence that the gap is structural rather than a capacity artifact. A decomposition shows the drop is far larger on episodes containing truly novel LLI→HLI pairings (0.47) than on episodes held out by ordering pattern alone (0.30), confirming genuine composition rather than distribution shift. The paper also finds that logic-free baselines violate the evaluation-only semantic rules above the 2.5% intrinsic violation rate of the ground-truth labels, arguing the benchmark cannot be solved by recovering the generator.

Load-bearing premise

The framework assumes that the experts' definition of what counts as a coherent intention, and the 22 rules used at evaluation, are a valid and stable standard independent of the generator; if that standard is idiosyncratic or shaped by the generator, the benchmark tests an arbitrary construct rather than human intentions.

Editorial extensions

If this is right

  • Benchmark scores on standard splits overstate compositional ability: models must also be evaluated on held-out intention combinations to measure genuine generalization.
  • The framework transfers to any flat single-label action corpus with per-performer metadata and sufficient class coverage, enabling hierarchy benchmarks in other domains without new recording campaigns.
  • The released ontology, transition model, and generator allow the benchmark to be regenerated and extended, supporting controlled comparisons of neuro-symbolic models that incorporate logical constraints.
  • Because order destruction leaves intention-level performance nearly unchanged, high-level intention in this benchmark is recoverable from co-occurrence of sub-behaviors; future versions can inject stronger temporal signal to test order-based reasoning.

Reading between the lines

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

  • The anti-circularity protocol — measuring violation rates on evaluation-only rules and checking that ground-truth labels violate them at a non-zero rate — could be adopted by other synthetic benchmarks as a cheap red flag for generator recovery.
  • If the ontology were rebuilt from scratch by a different expert panel, the size of the compositional gap on the new version would reveal how much of the difficulty depends on this particular semantic standard rather than on compositional structure in general.
  • The order-flexibility finding suggests a design axis for benchmark generators: tuning the transition model's temporal strength can modulate how much of the task is about sequencing versus co-occurrence, letting researchers target specific reasoning abilities.
  • A model that explicitly factorizes LLI and HLI components might close part of the gap; testing this would show whether the structural property is an irreducible interaction between component features.
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

5 major / 5 minor

Summary. The paper proposes a benchmark-generation framework that synthesizes a four-level hierarchical intention benchmark (actions, activities, low-level intentions, high-level intentions) from a flat single-label action corpus (NTU RGB+D 120), while retaining real pre-extracted action features. Episodes are assembled via a transition model under a subject-consistency constraint and a coverage-aware sampler that reduces the subject-usage Gini from 0.566 to 0.248. To avoid circular supervision, sequence-generation rules (Group X) are kept disjoint from evaluation-time first-order-logic rules (Group Y), and a compositional held-out split withholds LLI-HLI parent associations. The instantiated benchmark contains 15,002 episodes. Four reference baselines (hierarchical transformer, sequential transformer, bag-of-actions MLP, relational graph network) are evaluated, yielding a compositional held-out gap of 0.13–0.17 macro-F1 across all baselines. The paper claims this gap is a structural property of the benchmark and not a model artifact, supported by logic-violation rates and an order-destroying control. The ontology, transition model, and generator are released.

Significance. If the central result holds, the paper makes a useful contribution to compositional and hierarchical action recognition: it provides a regenerable protocol that turns a flat action corpus into a deep intention hierarchy with explicit anti-circularity safeguards, coverage balancing, and compositional splits. The released generator and ontology enable others to regenerate and extend the benchmark. The compositional gap observed across four model families, including a graph-aware model, is an interesting finding that could guide future benchmark design. However, the strength of the claim that the gap is 'structural' is currently limited by the lack of regeneration-seed variation and by uncontrolled split comparisons. The paper is a benchmark-protocol paper, not a recognition-methods paper, and its value depends on the empirical stability of the reported gap.

major comments (5)
  1. [Section 7, Table 6 and Abstract] The claim that the compositional gap of 0.13–0.17 macro-F1 is 'a structural property of the benchmark rather than a model artifact' is under-supported because the reported variability (three to five seeds) covers only model-initialization randomness, not the synthesis generator's own randomness (transition-model perturbations, coverage sampler, episode sampling). The paper frames the contribution as a regenerable protocol (Section 8; Data Availability), so a single instantiation of the stochastic generator cannot establish a structural property. Please report the gap's stability over multiple independent regeneration seeds (e.g., 10 regenerated benchmark instances with the same ontology and hyperparameters), or temper the claim accordingly.
  2. [Section 7, 'Compositional generalization' paragraph] The decomposition of the gap into 1,854 novel-context episodes and 321 ordering-pattern episodes is not defined anywhere in the paper: no section explains how these subsets are derived or what 'held out by ordering pattern alone' means. Without this definition, the claim that the gap is 'concentrated on genuine novel compositions' cannot be checked. Furthermore, the standard and compositional splits appear to be different episode sets; a matched test set (identical held-out episodes, differing only in which LLI-HLI associations appear in training) is needed to rule out the possibility that the gap reflects raw distribution shift rather than compositional failure.
  3. [Section 7, Table 8 and surrounding text] The text states that destroying order 'changes macro-F1 by less than one standard deviation at every level and baseline,' but the reported numbers are inconsistent with this claim: for example, the hierarchical transformer's action-level drop of +2.9 with standard deviation 2.0 is 1.45 standard deviations, and the sequential transformer's action-level drop of +5.2(5.5) is near one standard deviation but the text's 'around 0.02 to 0.04' range does not match the table's range of 0.009–0.052. Please correct the text and provide proper significance testing (e.g., a paired bootstrap or permutation test across episodes) before concluding that the order-destroying control shows no effect.
  4. [Section 7, Table 7] The claim that 'every model exceeds the 2.5% Group Y data floor, confirming genuine error rather than generator recovery' is based on mean differences of 0.3–1.7 percentage points without any significance test or confidence interval. Because the 2.5% floor is itself an estimate from a single generated dataset, please report a test (e.g., bootstrap over episodes or seeds) for whether each model's Group Y violation rate is significantly above the data floor. As written, the evidence is suggestive but not statistically established.
  5. [Section 5, 'Disjoint generation and evaluation rules'] The anti-circularity argument hinges on the assertion that the 22 Group Y rules are 'semantic constraints the generator does not use' and 'any annotator would apply independently.' Since both the generator's transition model and the Group Y rules are authored by the same researchers on the basis of the same ontology, this independence is not automatic and is not empirically demonstrated. Please provide a direct test, for example by measuring the correlation between the transition model's soft ordering preferences and the Group Y rules, or by showing that the Group Y rules cannot be derived from the ontology's transition structure. Without such a check, the validity claim is only asserted.
minor comments (5)
  1. [Section 5, plausibility check] The text says 'five independent annotators judged whether the action sequences of 150 sampled episodes form coherent behavior' but then reports 'majority-vote plausibility rate was 0.86 (43/50)'. The numbers 150 and 50 are inconsistent; please clarify how many episodes were annotated and how the 43/50 was computed.
  2. [Section 4, 'Heterogeneous graph construction'] There is a typo in the subsection heading: 'heterogeneous-graphepisode' should be 'heterogeneous-graph episode'.
  3. [Section 7, Table 6 caption] The caption says 'bootstrap 95% confidence interval on the hierarchical-transformer gap over five seeds is [0.161,0.171]' but it is not stated explicitly that this interval is over model-training seeds only and not over regeneration seeds; please state this clearly to avoid misunderstanding.
  4. [Section 7, Table 8 caption] Table 8's caption says 'mean±standard deviation in parentheses over three seeds,' while Tables 5 and 6 report HT and R-GCN over five seeds and seq-T over three. Please clarify whether Table 8 used three seeds for all baselines and why, or align the captions.
  5. [Section 8, 'Discussion and Conclusion'] The claim that the framework 'transfers to other flat single-label action corpora' is stated as a property, but no transfer demonstration is provided. Since a second corpus is explicitly left to future work, please soften the claim or add a discussion of the conditions under which transfer might fail.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central compositional-gap result is an empirical held-out measurement, and the paper's anti-circularity controls are explicit and independently checkable.

full rationale

The paper's central claim is an empirical measurement on held-out splits: macro-F1 drops by 0.13 to 0.17 on the compositional split across four baseline families (Table 6), with a bootstrap confidence interval over seeds. This is not derived from the generation rules; it is an observed gap on data the models never saw. The anti-circularity design is concrete: Group X rules (3) are shared with the generator, while Group Y rules (22) are evaluation-only, and the paper verifies empirically that ground-truth labels violate Group Y at a 2.5% intrinsic rate (Table 7), so those rules are not satisfied by construction. The order-destroying control is likewise an empirical probe with acknowledged limitations, and the paper explicitly states that order flexibility at the intention level is a designed property, not a discovered empirical claim. The limitation that intentions are imposed by construction rather than observed is stated openly and is a scope condition, not a circularity. No load-bearing self-citations or imported uniqueness theorems appear; references to grammar-based generation and compositional-generalization benchmarks are external and contextual. The skeptic's concerns about regeneration-seed variance and ontology idiosyncrasy are robustness/validity risks, not instances of a prediction reducing to its inputs. No step fits the enumerated circularity patterns.

Assumptions & free parameters 3 free parameters · 5 assumptions · 2 invented entities

The central claim rests on the expert ontology, hand-set transition and sampling parameters, and the assumption that Group Y rules are independent of the generator. The only external evidence is a small plausibility check of 150 episodes; the benchmark's validity is otherwise defined by the authors' own constructs.

free parameters (3)
  • Perturbation probabilities = 0.1
    Activity substitution, order perturbation, and filler insertion each occur with probability 0.1 per episode (Section 4). This is a hand-set hyperparameter affecting episode diversity.
  • Branching ranges = 2-5 LLIs per HLI, 2-4 activities per LLI, 1-2 actions per activity
    Episode tree structure ranges are chosen by the authors in Section 4. They constrain the benchmark's structural statistics and are not derived from data.
  • Eligibility floor = 2
    The coverage-aware sampler relaxes to at least 2 eligible performers when needed (Section 4). This is a hand-set threshold that affects subject balance and episode composition.
assumptions (5)
  • domain assumption NTU RGB+D 120 pre-extracted features correctly represent atomic actions
    The benchmark reuses PoseC3D and VideoMAE features without revalidation; errors in these features propagate into the generated episodes.
  • domain assumption The expert ontology is semantically meaningful and covers the action space
    Section 3 states semantic realism; there is no external validation except a small plausibility check of 150 episodes.
  • domain assumption Group Y rules are independent of the generator and annotator-universal
    Section 5 asserts these rules would be applied by any annotator; the intrinsic 2.5 percent violation rate is the only evidence offered.
  • domain assumption Subject consistency is a valid proxy for real performer identity
    All clips in an episode share one performer; two-person source actions are anchored to one performer, a simplification stated in Section 8.
  • domain assumption The four baselines are representative of model families
    Section 7 generalizes from four architectures to a structural claim about the benchmark, which is a stronger conclusion than the evidence supports.
invented entities (2)
  • LLI and HLI intention labels
    purpose: Provide two semantic supervision levels above activities
    These categories are expert-authored and imposed by construction; the paper acknowledges in Section 8 that intentions are not observed.
  • 22 Group Y first-order-logic rules
    purpose: Evaluation-time validity instrument to measure anti-circularity
    The rules are defined by the authors and used to assess whether models recover the generator; they are not grounded in an external standard.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Compositional Benchmark Synthesis for Hierarchical Human Action Recognition." pith.science (2026). https://pith.science/paper/HEA6WOJJ

@misc{pith2026260810765,
  author       = {Pith},
  title        = {Pith review of: Compositional Benchmark Synthesis for Hierarchical Human Action Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HEA6WOJJ}},
  note         = {Machine review of arXiv:2608.10765}
}
read the original abstract

Recognizing human behavior across levels of abstraction, from atomic actions to long-horizon intentions, requires data annotated along a semantic hierarchy. Large corpora provide isolated, atomically labeled clips without temporal composition, whereas recorded composite-activity corpora offer shallow, domain-narrow, fixedhierarchies. A benchmark-generation and evaluation frameworkis proposed that synthesizes a four-level hierarchical-intention benchmark, spanning actions, activities, low-level intentions (LLIs), and high-level intentions (HLIs), from a flat single-label action corpus while retaining real pre-extracted features at the action level. Episodes are assembled by a transition model under a subject-consistency constraint, and a coverage-aware sampler reduces the subject usage Gini from 0.566 to 0.248. Synthesizing such a benchmark raises a circular-supervision risk that recorded datasets avoid: if the rules generating the episodes also govern the evaluation, models can succeed by recovering the generator rather than through genuine reasoning. Validity is addressed by design, holding sequence-generation rules disjoint from the first-order-logic rules used at evaluation. The instantiation yields 15,002 episodes. Four reference baselines from different model families characterize difficulty, not as recognition methods. A compositional held-out gap of 0.13 to 0.17 macro-F1 appears across all baselines, including a graph-aware model that recognizes best yet does not close the gap, indicating a structural property of the benchmark rather than a model artifact. A logic-free baseline still violates the held-out semantic rules above their intrinsic data rate, and the order-destroying control changes macro-F1 within seed variation, serving as a generator-consistency check. Theontology, transition model, and generator are released so the benchmark can beregenerated and extended.

Figures

Figures reproduced from arXiv: 2608.10765 by the authors.

Figure 1
Figure 1. Overview of the experiment. (A) Benchmark synthesis: a flat, single-label action corpus is composed into four-level hierarchical-intention episodes through an ontology and transition model, a coverage-aware subject-consistent sampler that reduces the subject-usage Gini from 0.566 to 0.248, and a heterogeneous-graph representation with decoupled subject-disjoint and compositional splits. (B) Evaluation protocol: refe… view at source ↗
Figure 2
Figure 2. Compact benchmark statistics. The panels show (a) source subject-by-class availability, (b) Lorenz curves of source availability versus realized subject usage, and (c) per-level branching distributions of the synthesized episodes. The coverage-aware sampler reduces the realized subject￾usage Gini to 0.248 (most-used performer in 364 episodes, least in 64). Eligible-subject pools per HLI range from 40 to 60, reflecti… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 25 canonical work pages

  1. [1]

    Logic tensor networks

    Badreddine, S., Garcez, A.d., Serafini, L., Spranger, M., 2022. Logic tensor networks. Artificial Intelligence 303, 103649

    work page 2022

  2. [2]

    Enhancing action recognition by leveraging the hierarchical structure of actions and textual context

    Benavent-Lledo, M., Mulero-Pérez, D., Ortiz-Perez, D., Garcia- Rodriguez, J., Argyros, A., 2025. Enhancing action recognition by leveraging the hierarchical structure of actions and textual context. Computer Vision and Image Understanding , 104560

    work page 2025

  3. [3]

    Vision and Intention Boost Large Language Model in Long-Term Action Anticipation

    Cao, C., Hu, L., Yu, Y., Zhang, Y., 2025. Vision and intention boost largelanguagemodelinlong-termactionanticipation. arXivpreprint arXiv:2505.01713

    work page Pith review arXiv 2025

  4. [4]

    Cruz, R.S., Cherian, A., Fernando, B., Campbell, D., Gould, S.,

    work page

  5. [5]

    Procedu- ralgenerationofvideostotraindeepactionrecognitionnetworks,in: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp

    DeSouza,C.R.,Gaidon,A.,Cabon,Y.,López,A.M.,2017. Procedu- ralgenerationofvideostotraindeepactionrecognitionnetworks,in: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2594–2604. doi:10.1109/CVPR.2017.278

    work page doi:10.1109/cvpr.2017.278 2017

  6. [6]

    Datasheetsfordatasets

    Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J.W., Wallach, H.,Iii,H.D.,Crawford,K.,2021. Datasheetsfordatasets. Communi- cations of the ACM 64, 86–92

    work page 2021

  7. [7]

    What has been lost with synthetic evaluation? arXiv preprint arXiv:2505.22830

    Gill, A., Ravichander, A., Marasović, A., 2025. What has been lost with synthetic evaluation? arXiv preprint arXiv:2505.22830

    arXiv 2025

  8. [8]

    Heterogeneous graph transformer, in: Proceedings of the web conference 2020, pp

    Hu, Z., Dong, Y., Wang, K., Sun, Y., 2020. Heterogeneous graph transformer, in: Proceedings of the web conference 2020, pp. 2704– 2710

    2020

Show all 27 references
  1. [9]

    Cogs: A compositional generalization challenge based on semantic interpretation, in: Proceedings of the 2020conferenceonempiricalmethodsinnaturallanguageprocessing (emnlp), pp

    Kim, N., Linzen, T., 2020. Cogs: A compositional generalization challenge based on semantic interpretation, in: Proceedings of the 2020conferenceonempiricalmethodsinnaturallanguageprocessing (emnlp), pp. 9087–9105

    2020

  2. [10]

    C2c: Component-to-composition learning for zero- shot compositional action recognition, in: European Conference on Computer Vision, Springer

    Li, R., Feng, Z., Xu, T., Li, L., Wu, X.J., Awais, M., Atito, S., Kittler, J., 2024. C2c: Component-to-composition learning for zero- shot compositional action recognition, in: European Conference on Computer Vision, Springer. pp. 369–388

    2024

  3. [11]

    Liu, J., Shahroudy, A., Perez, M., Wang, G., Duan, L.Y., Kot, A.C.,

  4. [12]

    Fineaction: A fine-grained video dataset for temporal action localization

    Liu, Y., Wang, L., Wang, Y., Ma, X., Qiao, Y., 2022. Fineaction: A fine-grained video dataset for temporal action localization. IEEE transactions on image processing 31, 6937–6950

    2022

  5. [13]

    Intention-conditioned long- term human egocentric action anticipation, in: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp

    Mascaró, E.V., Ahn, H., Lee, D., 2023. Intention-conditioned long- term human egocentric action anticipation, in: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 6048–6057

    2023

  6. [14]

    1049– 1059

    Materzynska, J., Xiao, T., Herzig, R., Xu, H., Wang, X., Darrell, T., 2020.Something-else:Compositionalactionrecognitionwithspatial- temporal interaction networks, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 1049– 1059

    2020

  7. [15]

    Gaze-guided graph neural network for action anticipation conditioned on intention, in: Proceedings of the 2024 Symposium on Eye Tracking Research and Applications, pp

    Özdel, S., Rong, Y., Albaba, B.M., Kuo, Y.L., Wang, X., Kasneci, E., 2024. Gaze-guided graph neural network for action anticipation conditioned on intention, in: Proceedings of the 2024 Symposium on Eye Tracking Research and Applications, pp. 1–9

    2024

  8. [16]

    Ad- versarial generative grammars for human activity prediction, in: Eu- ropean Conference on Computer Vision, Springer

    Piergiovanni, A., Angelova, A., Toshev, A., Ryoo, M.S., 2020. Ad- versarial generative grammars for human activity prediction, in: Eu- ropean Conference on Computer Vision, Springer. pp. 507–523

    2020

  9. [17]

    Predicting human activities using stochastic grammar, in: Proceedings of the IEEE International Conference on Computer Vision, pp

    Qi, S., Huang, S., Wei, P., Zhu, S.C., 2017. Predicting human activities using stochastic grammar, in: Proceedings of the IEEE International Conference on Computer Vision, pp. 1164–1172

    2017

  10. [18]

    Assembly101: A large-scale multi-view video datasetforunderstandingproceduralactivities,in:Proceedingsofthe IEEE/CVFConferenceonComputerVisionandPatternRecognition, pp

    Sener, F., Chatterjee, D., Shelepov, D., He, K., Singhania, D., Wang, R., Yao, A., 2022. Assembly101: A large-scale multi-view video datasetforunderstandingproceduralactivities,in:Proceedingsofthe IEEE/CVFConferenceonComputerVisionandPatternRecognition, pp. 21096–21106

    2022

  11. [19]

    Humanactivitiesasstochastickroneckergraphs, in:EuropeanConferenceonComputerVision,Springer.pp.130–143

    Todorovic,S.,2012. Humanactivitiesasstochastickroneckergraphs, in:EuropeanConferenceonComputerVision,Springer.pp.130–143

    2012

  12. [20]

    A semantic loss function for deep learning with symbolic knowledge, in: International conference on machine learning, PMLR

    Xu, J., Zhang, Z., Friedman, T., Liang, Y., Broeck, G., 2018. A semantic loss function for deep learning with symbolic knowledge, in: International conference on machine learning, PMLR. pp. 5502– 5511

    2018

  13. [21]

    Logicmp: A neuro-symbolic approach for encodingfirst-orderlogicconstraints,in:InternationalConferenceon Learning Representations, pp

    Xu,W.,Wang,J.,Xie,L.,He,J.,Zhou,H.,Wang,T.,Wan,X.,Chen, J., Qu, C., Chu, W., 2024. Logicmp: A neuro-symbolic approach for encodingfirst-orderlogicconstraints,in:InternationalConferenceon Learning Representations, pp. 4181–4209

    2024

  14. [22]

    Skelformer:Anadaptivehi- erarchical transformer-based approach on skeleton graphs for human action recognition in video sequences

    Yan,J.,Zhang,X.,Tan,C.,Li,D.,2026. Skelformer:Anadaptivehi- erarchical transformer-based approach on skeleton graphs for human action recognition in video sequences. PloS one 21, e0340390

    2026

  15. [23]

    Zero-shot composi- tionalactionrecognitionwithneurallogicconstraints,in:Proceedings ofthe33rdACMInternationalConferenceonMultimedia,pp.3625– 3634

    Ye, G., Li, L., Li, K., Xiao, J., Chen, L., 2025. Zero-shot composi- tionalactionrecognitionwithneurallogicconstraints,in:Proceedings ofthe33rdACMInternationalConferenceonMultimedia,pp.3625– 3634

    2025

  16. [24]

    Action anticipation with goal consis- tency, in: 2023 IEEE International Conference on Image Processing (ICIP), IEEE

    Zatsarynna, O., Gall, J., 2023. Action anticipation with goal consis- tency, in: 2023 IEEE International Conference on Image Processing (ICIP), IEEE. pp. 1630–1634

    2023

  17. [25]

    Zhu, Q., Zhang, L., Xu, Q., Liu, K., Long, C., Wang, X., 2025. Hhgt: hierarchical heterogeneous graph transformer for heteroge- neous graph representation learning, in: Proceedings of the Eigh- teenth ACM International Conference on Web Search and Data Min- ing, pp. 318–326. S...

    2025

  18. [2019]

    IEEE transactions on pattern analysis and machine intelligence 42, 2684–2701

    Nturgb+d120:Alarge-scalebenchmarkfor3dhumanactivity understanding. IEEE transactions on pattern analysis and machine intelligence 42, 2684–2701

  19. [2020]

    Inferring temporal compositions of actions using probabilistic automata,in:ProceedingsoftheIEEE/CVFConferenceonComputer Vision and Pattern Recognition Workshops, pp. 368–369

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.