Heterogeneous Sequel-Aware Graph Neural Networks for Sequential Learning

Anushka Tiwari; Haimonti Dutta; Shahrzad Khanizadeh

REVIEW 4 major objections 7 minor 29 references

Heterogeneous Sequel-Aware Graph Neural Networks for Sequential Learning

T0 review · 4 major / 7 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 Explicitly modeling item-sequel links in a heterogeneous GNN improves next-item recommendation when enough items belong to series, with gains scaling with sequel density.

desk verdict A novel sequel-edge construction, but the evaluation leaks future information and the ML-100K results contradict the abstract. read the letter →

arxiv 2506.05625 v1 pith:MZ4F7U2C submitted 2025-06-05 cs.IR cs.LG

Anushka Tiwari , Haimonti Dutta , Shahrzad Khanizadeh This is my paper

classification cs.IRcs.LG

keywords sequentialrecommendationsequel-awaregraphneuralnetworksheterogeneousgraphsnext-itempredictiontemporalitemsequencesuser-itembipartite

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 asks whether next-item recommendation improves when the model is explicitly told which items belong to an ordered series, and answers yes under the right conditions. It builds a heterogeneous graph with two edge types \u2014 user-item interactions and sequel links between items in a series \u2014 and propagates messages along both. On three synthetic and three real-world datasets, the resulting HSAL-GNN beats or matches a strong graph-based recommender that ignores sequels. The size of the gain tracks the share of items that actually belong to series: substantial on Goodreads, marginal on MovieLens 1m, and absent on MovieLens 100k, where the sequel-aware model trails a graph baseline.

What carries the argument

The load-bearing mechanism is sequel-aware message propagation. For each item $i$, the model computes $h^{\mathrm{seq}}_i = \frac{1}{|N_s(i)|} \sum_{j \in N_s(i)} h^{(l-1)}_j \cdot P(j)$, where $N_s(i)$ is the set of items that follow $i$ in its series and $P(j)$ is a sinusoidal positional encoding of $j$'s position in the series. This term is what injects ordered series structure into item embeddings; it is fused with long-term and short-term interaction embeddings and the fused representation updates the item node. When this term is removed or the sequel edges are noisy, the model's advantage over a user-item GCN shrinks.

What would settle it

Run HSAL-GNN and the strongest graph baseline on a dataset with publisher-verified series metadata, holding the user-item graph fixed. If the sequel-aware model does not beat the baseline on series-rich item slices, or if randomly shuffling sequel links does not degrade its scores, then the claimed benefit of sequel information is not real.

Watch

Extended reading notes

Core claim

The central claim is that temporal item-sequence structure is a usable signal for next-item recommendation. HSAL-GNN represents each item as either standalone or sequel; for sequel items it aggregates messages from the items that come later in the same series, weighted by a sinusoidal positional embedding, and then fuses this sequel-aware representation with long-term and short-term user-item signals before predicting the next item as a link in a sampled subgraph. The paper's empirical thesis is that this fusion improves ranking quality over graph-based recommendation without sequel information. The improvement concentrates in sequel-rich data: NDCG@10 rises by 6.46% over the strongest baseline on Goodreads, by a small margin on ML-1m, and not at all on ML-100K, where the model underperforms a graph baseline.

Load-bearing premise

The whole advantage depends on the sequel edges in the data being accurate and complete; MovieLens sequels are inferred by matching titles like 'Part 1' and 'Part 2', which can miss or mislabel series entries, and the Goodreads sample is described without reporting how it was drawn.

Editorial extensions

If this is right

On platforms where items naturally form ordered series (books, films, games, courses), adding sequel edges to a recommender GNN should improve next-item ranking, with gains proportional to series density.
Below a certain fraction of series items \u2014 around 4\u20138% in the paper's data \u2014 sequel-aware modeling stops helping and can lose to a plain graph model, so metadata coverage should be checked before adopting it.
Fusion strategy matters: simple sum-based fusion of sequel and interaction signals outperformed mean, concatenation, and a more complex heterogeneous transformation on the tested datasets.
The method extends to any directed, ordered item relation, not only entertainment sequels; prerequisite chains in education and versioned products are natural candidates for the same graph construction.

Reading between the lines

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

The paper's pattern of results implies a practical decision rule the authors do not state: estimate the share of sequel items and the reliability of series metadata before adopting sequel-aware modeling; if either is low, a standard user-item GCN is likely the safer choice.
A stronger test of the mechanism would use verified series metadata on MovieLens; the title-matching inference used here may understate the potential benefit of clean sequel edges.
Because the paper reports only leave-one-out evaluation on sampled data, a live deployment would reveal whether the NDCG gains translate into user-facing recommendation quality.

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.

Desk Editor's Note

read the letter

The first thing you should know: the paper's central claim—that sequel-aware GNNs are better or comparable to graph baselines—is not supported by its own experiments, and there is a plausible temporal leakage in the sequel-edge construction that would explain the positive results. Still, the core idea is genuinely new and worth a second look.

What's new: the authors inject explicit series edges with sinusoidal positional encoding into a heterogeneous GNN. That's a real departure from DGSR's item-item transitions, which don't preserve order. The problem is well-motivated, and the synthetic setup with mixed/sequential/standalone splits is a sensible way to isolate the effect. The ablation of fusion strategies and positional encodings is thorough for an architecture paper.

The soft spots are serious. In §4.1, sequel edges are built over the full series with no timestamp restriction. In §4.3.2, the model aggregates 'subsequent sequel items' for each item. Under the leave-one-out protocol, the held-out target is often a direct successor of an item already in the user's history. The user representation can therefore contain the target item's own embedding before scoring. That's leakage, not temporal modeling. The pattern in Table 3—large gains on Goodreads, smaller on ML-1m, losses on ML-100K—is exactly what you'd expect if the edge construction is giving the model access to future items. The paper does not discuss or rule this out.

Second, the ML-100K numbers are misreported. Table 3 shows HSAL-GNN worse than DGSR on every metric (NDCG@10 0.4347 vs 0.4439), but the text calls this 'little to no improvement.' That's a direct contradiction. Also missing: error bars despite five seeds, the code, and details about the Goodreads sampling.

Who is this for: people working on item-series structure in sequential recommendation. The idea has potential, but as reported, the evidence doesn't demonstrate an effect. If the authors rerun with timestamp-constrained sequel edges and show gains survive, the paper could be worth publishing. As is, I'd expect a reviewer to reject on leakage grounds.

Recommendation: I would not accept the current version. But I'd send it to a serious referee rather than desk-reject it, because the construction is novel and the leakage concern is testable. With a careful revision, there's a path to a useful contribution.

Referee Report

4 major / 7 minor

Summary. The paper proposes HSAL-GNN, a heterogeneous graph neural network for sequential recommendation that augments the user-item bipartite graph with item-item edges representing sequel relationships. The architecture combines long-term, short-term, and sequel-aware item embeddings through several fusion strategies and uses sinusoidal positional encodings. The authors evaluate on three synthetic and three real-world datasets against GRU4Rec, SASRec, BERT4Rec, SVAE, ACVAE, DiffuRec, STOSA, and DGSR, using a leave-one-out protocol. The central claim is that explicitly modeling sequel relations improves or matches state-of-the-art graph-based sequential recommenders, especially when sequel information is abundant. The paper also includes an ablation study and a sensitivity analysis of hyperparameters.

Significance. The problem of exploiting item-series structure is timely and under-explored, and the idea of encoding such structure in a heterogeneous GNN is reasonable. The paper offers a broad empirical comparison and a detailed ablation, and it makes code available. If the claimed improvements were established with a sound evaluation protocol, the contribution would be useful to the recommender-systems community. However, the current evaluation suffers from a serious temporal-leakage issue that casts doubt on the reported gains, several reported results contradict the text, and internal inconsistencies appear in the ablation table. As a result, the significance of the contribution is not established by the experiments as reported.

major comments (4)

[§4.1, §4.2, §4.3.2] The sequel-edge construction in §4.1 and the subgraph sampling in Algorithm 1 create item-item edges without any timestamp restriction. Section 4.3.2 defines hseq_i by aggregating over 'subsequent items in the sequel series.' In the leave-one-out protocol (§5.4), the held-out next item is frequently a direct successor of an item already in the user's history; because the graph contains all user interactions and sequel edges are formed over the full series, the candidate item's embedding can incorporate information from interactions that occur after the prediction time, including the target item itself. This gives HSAL-GNN access to future information that DGSR and other baselines do not have, so the reported gains in Tables 2 and 3 do not measure the value of sequel-aware representation. The authors should restrict sequel edges to pairs whose interaction timestamps are both before the prediction time, or re-run the evaluation under a strict temporal split (e.g., training only on edges with t < t_k for each user). Without such a correction, the central claim in the abstract is not supported.
[§5.5, Table 3, Table 2] Section 5.5 states that ML-100K 'sees little to no improvement' and that the Non-Sequential synthetic dataset shows 'minimal improvements.' Table 3 shows that on ML-100K HSAL-GNN is worse than DGSR on every metric, with NDCG@10 0.4347 vs 0.4439 and Hit@10 0.7241 vs 0.7288. Table 2 shows that on the Non-Sequential dataset HSAL-GNN is worse than DGSR on all metrics, including NDCG@10 0.0432 vs 0.0476. Also on the Sequential dataset, HSAL-GNN is slightly worse than DGSR on most metrics (e.g., NDCG@10 0.9529 vs 0.9553). These results contradict the text and weaken the abstract's claim of 'better (or comparable)' performance. The report should be corrected and the negative results discussed honestly.
[Table 4] In Table 4, the 'Sinusoidal' positional-embedding row reports ML-1m NDCG@10 = 0.4347, while the 'Sum' fusion row (which uses the default sinusoidal embedding) reports 0.4795 and the corresponding HSAL-GNN result in Table 3 is 0.4796. The value 0.4347 is identical to the ML-100K NDCG@10 in Table 3, suggesting a copy-paste error. This inconsistency undermines the ablation claim that sinusoidal embeddings outperform rotary embeddings on ML-1m. The ablation table should be verified and corrected.
[§5.2] Section 5.2 states that experiments are run five times and mean performance is reported, but no standard deviations or significance tests are provided. Many of the claimed improvements are small, for example ML-1m NDCG@10 0.4796 vs 0.4764, Goodreads NDCG@10 0.2074 vs 0.1948, and Mixed NDCG@10 0.4539 vs 0.4290. Without variance reporting, these differences may be due to random seed variation. The authors should report mean ± std over the five seeds and perform paired significance tests for the main comparisons.

minor comments (7)

[§5.1, Table 1] The density formula is unclear and the values appear inconsistent: for Goodreads, 200k interactions / (2000 users × 6474 items) is about 1.5%, not 0.01%, and for ML-1m, 1.0M / (6040 × 3900) is about 4.2%, not 4.79%. Please clarify the formula and correct the numbers.
[§5.7] The text says 'as illustrated in Figure 4' when discussing the effect of the number of layers, but the layer experiments are shown in Figure 3; Figure 4 is the bar-chart comparison. Please correct the cross-reference.
[§1] The abbreviation 'LTSM' is a typo and should be 'LSTM'.
[§4.1] The definition of ia and ib in Eq. (3) is confusing; please clarify that ia is the immediate predecessor and ib the immediate successor of ip in the series, and define the edge-tuple notation consistently.
[§4.3.4] The 'SeHGNN-based transformation' is described only by name; please provide the exact equation or a precise reference to the specific layer so that the fusion strategy is reproducible.
[§3.1] There are footnote markers 3 and 4 with no accompanying footnote text in the submission; please either supply the footnotes or remove the markers.
[§5.2] The maximum sequence length is set to 50, while the average sequence length in Goodreads is 260.9 (Table 1); please explain how longer sequences are truncated and whether this affects the comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: HSAL-GNN's sequel-aware design is an empirical architectural contribution evaluated against external baselines, not a derivation that reduces to its own inputs.

full rationale

The paper does not claim to derive a result from first principles; it proposes a heterogeneous GNN architecture that incorporates sequel-aware item-item edges and evaluates it against standard baselines on synthetic and real-world datasets. The sequel-aware message propagation hseq_i = (1/|Ns(i)|) sum_{j in Ns(i)} (h_j * P(j)) is a modeling choice, not a fitted parameter that is later renamed as a prediction. The evaluation uses a leave-one-out protocol with training, validation, and test splits, and the reported Hit/NDCG gains are measured against external baselines such as DGSR, SASRec, and BERT4Rec. The ablation study tests fusion strategies and positional embeddings, but none of these variants is fitted to the test labels and then reported as an independent prediction. The paper contains no load-bearing self-citations: all references are to external prior work. A reader might worry that sequel edges built without timestamp constraints could leak future interactions, and that the gains on Goodreads and the losses on ML-100K are consistent with such leakage; however, that is a correctness or validity concern about the experimental setup, not a circularity in which the claimed result is equivalent to the input by construction. No equation in the paper defines the predicted next item in terms of the evaluation labels, and no fitted constant is passed off as an architectural result. Therefore the appropriate circularity score is 0.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The central claim relies on reliable extraction of sequel structure and on the assumption that ordered series consumption is a useful signal; both are domain assumptions rather than derived facts. The hyperparameters are conventional and not fitted to the target result, so the circularity burden is low.

free parameters (7)

Embedding dimension d = 50
Set to 50 in Section 5.2; not tuned, but determines representation size and model capacity.
Sub-graph sampling order m = 4
Set to 4 in Section 5.2; controls neighborhood depth and is chosen without a reported sensitivity analysis.
Number of GNN layers L = tuned per dataset (final values not listed)
Section 5.2 says L is tuned per dataset; individual values are not reported and Figure 3 shows different optimal depths across datasets.
Maximum sequence length = 50
Set to 50 in Section 5.2; truncates user sequences and affects which interactions are modeled.
Learning rate = 0.01
Adam optimizer setting in Section 5.2, chosen by hand.
Batch size = 50
Adam optimizer setting in Section 5.2, chosen by hand.
Weight decay lambda = 1e-4
L2 regularization strength in Section 5.2, chosen by hand.

assumptions (4)

domain assumption Sequel relationships are static and known a priori from series metadata or title matching.
Section 4.1 constructs sequel-aware edges from a given sequence identifier s; for MovieLens, sequels are inferred by title matching (Section 5.1), and errors in this inference would corrupt the edges.
domain assumption Items in a sequel series have an ordered progression that users tend to consume in order.
The positional encoding P(i) in Section 3.1 assumes a monotonic order; the method's benefit relies on this behavioral regularity.
standard math Standard GNN message-passing converges to useful representations for recommendation.
The paper uses GCN-style aggregation (Section 4.3.1) without proof of convergence or expressiveness; this is a common assumption in the field.
domain assumption The m-order sub-graph sampling preserves enough signal for link prediction.
Section 4.2 restricts to m-order neighborhoods; if relevant users or items are outside this radius, the prediction is degraded.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Heterogeneous Sequel-Aware Graph Neural Networks for Sequential Learning." pith.science (2026). https://pith.science/paper/MZ4F7U2C

@misc{pith2026250605625,
  author       = {Pith},
  title        = {Pith review of: Heterogeneous Sequel-Aware Graph Neural Networks for Sequential Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MZ4F7U2C}},
  note         = {Machine review of arXiv:2506.05625}
}

read the original abstract

Graph-based recommendation systems use higher-order user and item embeddings for next-item predictions. Dynamically adding collaborative signals from neighbors helps to use similar users' preferences during learning. While item-item correlations and their impact on recommendations have been studied, the efficacy of temporal item sequences for recommendations is much less explored. In this paper, we examine temporal item sequence (sequel-aware) embeddings along with higher-order user embeddings and show that sequel-aware Graph Neural Networks have better (or comparable) recommendation performance than graph-based recommendation systems that do not consider sequel information. Extensive empirical results comparing Heterogeneous Sequel-aware Graph Neural Networks (HSAL-GNNs) to other algorithms for sequential learning (such as transformers, graph neural networks, auto-encoders) are presented on three synthetic and three real-world datasets. Our results indicate that the incorporation of sequence information from items greatly enhances recommendations.

Figures

Figures reproduced from arXiv: 2506.05625 by the authors.

**Figure 1.** Illustrative Example 1. (Left Figure) A sequential recommendation algorithm from literature.The goal is to predict the next item after i4, using user-item interaction history. (Right Figure) A sequel-aware sequential recommendation problem. The goal is to predict the next item after i4 using user-item interaction history and sequel information. The green nodes represent items a user has interacted with while the y… view at source ↗

**Figure 2.** Interaction of user-item graph at times t1, t2, t3 for the illustrative example discussed in the text. Upper graphs illustrate traditional, dynamic sequential recommendation; Lower graphs illustrate sequel-aware, dynamic sequential recommendation. The graphs are heterogeneous at each time – nodes belong to any of the following sets: U, iSA, or iSQ; edges can be user-iSA, user-iSQ, iSA-iSA, iSQ-iSQ, iSA-iSQ. 4.2 Su… view at source ↗

**Figure 3.** Effect of propagation layer numbers on different datasets. [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Comparison of Hit@10 and NDCG@10 scores across different algorithms [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]

**Figure 5.** Figure 5: Effect of (a) number of sequence and (b) length of sequence [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

**Figure 1.** Figure 1: presents the overall framework for HSAL-GNN. While all the details have been described in the main paper, the appendix presents a visual representation of the algorithm framework. Also, a schematic view of the sub-sampling algorithm and the forward propagation of HSAL…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

29 extracted references · 27 canonical work pages

[1]

Session-based recommen- dations with recurrent neural networks

Hidasi, B., Karatzoglou, A., Baltrunas, L. and Tikk, D., “Session-based recommen- dations with recurrent neural networks". arXiv preprint arXiv:1511.06939, 2015

arXiv 2015
[2]

Dynamic Graph Neural Networks for Sequential Recommendation

Mengqi Zhang, Shu Wu, Xueli Yu, Qiang Liu, and Liang Wang, “Dynamic Graph Neural Networks for Sequential Recommendation", IEEE Trans. on Knowl. and Data Eng. 35 (5), 4741–4753,2023

work page 2023
[3]

DiﬀuRec: A Diﬀusion Model for Sequential Recommendation

Zihao Li, Aixin Sun, and Chenliang Li, “DiﬀuRec: A Diﬀusion Model for Sequential Recommendation", ACM Trans. Inf. Syst. 42 (3),1–28, 2023

work page 2023
[4]

Modeling Se- quences as Distributions with Uncertainty for Sequential Recommendation

Ziwei Fan, Zhiwei Liu, Shen Wang, Lei Zheng, and Philip S. Yu, “Modeling Se- quences as Distributions with Uncertainty for Sequential Recommendation", In Proceedings of the International Conference on Information and Knowledge Man- agement (CIKM), pp 3019–3023, 2021

work page 2021
[5]

Sequential Recommendation via Stochastic Self-Attention

Ziwei Fan, Zhiwei Liu, Yu Wang, Alice Wang, Zahra Nazari, Lei Zheng, Hao Peng, and Philip S. Yu, “Sequential Recommendation via Stochastic Self-Attention", In Proceedings of the ACM Web Conference (WWW ’22). Association for Computing Machinery, New York, NY, USA, 2036–2047, 2022

work page 2022
[6]

Factorizing personalized Markov Chains for next-basket recommendation

S. Rendle, C. Freudenthaler, and L. Schmidt-Thieme, “Factorizing personalized Markov Chains for next-basket recommendation", In Proceedings of the Interna- tional conference on World Wide Web, pp. 811–820, 2010

work page 2010
[7]

Personalizing session- based recommendations with hierarchical recurrent neural networks

M. Quadrana, A. Karatzoglou, B. Hidasi, and P. Cremonesi, “Personalizing session- based recommendations with hierarchical recurrent neural networks", In Proceed- ings of the Eleventh ACM Conference on Recommender Systems, pp. 130–137, 2017

work page 2017
[8]

Recurrent neural networks with top-k gains for session-based recommendations

B. Hidasi and A. Karatzoglou, “Recurrent neural networks with top-k gains for session-based recommendations", In Proceedings of the 27th ACM International Conference on Information and Knowledge Management, 2018, pp. 843–852

work page 2018

Show all 29 references

[9]

A vatar, https://en.wikipedia.org/wiki/A vatar
[10]

Long short-term memory recurrent neural network architectures for large scale acoustic modeling

H. Sak, A. W. Senior, and F. Beaufays, “Long short-term memory recurrent neural network architectures for large scale acoustic modeling", 2014

2014
[11]

Empirical evaluation of gated recurrent neural networks on sequence modeling

J. Chung, C. Gulcehre, K. Cho, and Y. Bengio, “Empirical evaluation of gated recurrent neural networks on sequence modeling", arXiv preprint arXiv:1412.3555, 2014

2014 arXiv
[12]

Personalized top-n sequential recommendation via convo- lutional sequence embedding

J. Tang and K. Wang, “Personalized top-n sequential recommendation via convo- lutional sequence embedding", In Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining, 2018, pp. 565–573

2018
[13]

Self-attentive sequential recommendation

W.-C. Kang and J. McAuley, “Self-attentive sequential recommendation", In IEEE International Conference on Data Mining(ICDM). IEEE, pp. 197–206, 2018

2018
[14]

Stamp: short-term attention/memory priority model for session-based recommendation

Q. Liu, Y. Zeng, R. Mokhosi, and H. Zhang, “Stamp: short-term attention/memory priority model for session-based recommendation", In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Min- ing, 2018, pp. 1831–1839

2018
[15]

Variational autoen- coders for collaborative ﬁltering

D. Liang, R. G. Krishnan, M. D. Hoﬀman, and T. Jebara, “Variational autoen- coders for collaborative ﬁltering", In Proceedings of the World Wide Web Confer- ence, pp 689–698, 2018

2018
[16]

Sequential variational autoen- coders for collaborative ﬁltering

N. Sachdeva, G. Manco, E. Ritacco, and V. Pudi, “Sequential variational autoen- coders for collaborative ﬁltering", In Proceedings of the ACM International Con- ference on Web Search and Data Mining, pp 600–608, 2019

2019
[17]

Adversarial and contrastive variational autoencoder for sequential recommendation

Z. Xie, C. Liu, Y. Zhang, H. Lu, D. Wang, and Y. Ding, “Adversarial and contrastive variational autoencoder for sequential recommendation", In Proceedings of the Web Conference, pp 449–459, 2021. 16 Anushka Tiwari, Haimonti Dutta( ), and Shahrzad Khanizadeh

2021
[18]

An MDP-based recommender system

Guy Shani, David Heckerman, Ronen I. Brafman, and Craig Boutilier, “An MDP-based recommender system", Journal of Machine Learning Research, 6, 9, 1265–1295, 2005

2005
[19]

BERT4Rec: Sequential recommendation with bidirectional encoder representations from Transformer

F. Sun, J. Liu, J. Wu, C. Pei, X. Lin, W. Ou, and P. Jiang, “BERT4Rec: Sequential recommendation with bidirectional encoder representations from Transformer", In Proceedings of the ACM International Conference on Information and Knowledge Management, pp 1441–1450, 2019

2019
[20]

Leveraging two types of global graph for sequential fashion recommendation

Y. Ding, Y. Ma, W. K. Wong, and T.S. Chua, “Leveraging two types of global graph for sequential fashion recommendation", In Proceedings of the International Conference on Multimedia Retrieval. pp 73–81, 2021

2021
[21]

Session-based recommen- dation with graph neural networks

S. Wu, Y. Tang, Y. Zhu, L. Wang, X. Xie, and T. Tan, “Session-based recommen- dation with graph neural networks", In Proceedings of the AAAI Conference on Artiﬁcial Intelligence, Vol. 33, pp 346–353, 2019

2019
[22]

The Graph Neural Network Model,

F. Scarselli, M. Gori, A. C. Tsoi, M. Hagenbuchner and G. Monfardini, “The Graph Neural Network Model," in IEEE Transactions on Neural Networks, vol. 20, no. 1, pp. 61-80, Jan. 2009

2009
[23]

and Fan, D., 2023, June

Yang, X., Yan, M., Pan, S., Ye, X. and Fan, D., 2023, June. Simple and eﬃcient heterogeneous graph neural network. In Proceedings of the AAAI conference on artiﬁcial intelligence (Vol. 37, No. 9, pp. 10816-10824)

2023
[24]

and McAuley, J., 2018, November

Kang, W.C. and McAuley, J., 2018, November. Self-attentive sequential recom- mendation. In 2018 IEEE international conference on data mining (ICDM) (pp. 197-206). IEEE

2018
[25]

and Li, C., 2023

Li, Z., Sun, A. and Li, C., 2023. Diﬀurec: A diﬀusion model for sequential recom- mendation. ACM Transactions on Information Systems, 42(3), pp.1-28

2023
[26]

and Yu, P.S., 2022, April

Fan, Z., Liu, Z., Wang, Y., Wang, A., Nazari, Z., Zheng, L., Peng, H. and Yu, P.S., 2022, April. Sequential recommendation via stochastic self-attention. In Proceed- ings of the ACM web conference 2022 (pp. 2036-2047)

2022
[27]

Item Recommendation on Monotonic Behavior Chains

Wan, M. and McAuley, J. “Item Recommendation on Monotonic Behavior Chains", RecSys, 2018

2018
[28]

Fine-Grained Spoiler Detection from Large-Scale Review Corpora

Wan, M., Misra, R., Nakashole, N., McAuley, J., “Fine-Grained Spoiler Detection from Large-Scale Review Corpora", ACL, 2019

2019
[29]

While all the details have been described in the main paper, the appendix presents a visual represen- tation of the algorithm framework

MovieLens, https://grouplens.org/datasets/movielens/ Heterogeneous Sequel-Aware Graph Neural Networks for Sequential Learning Author information scrubbed for double-blind reviewing No Institute Given A Appendix A Figure 1 presents the overall framework for HSAL-GNN. While all ...

Pith tools

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

[1] [1]

Session-based recommen- dations with recurrent neural networks

Hidasi, B., Karatzoglou, A., Baltrunas, L. and Tikk, D., “Session-based recommen- dations with recurrent neural networks". arXiv preprint arXiv:1511.06939, 2015

arXiv 2015

[2] [2]

Dynamic Graph Neural Networks for Sequential Recommendation

Mengqi Zhang, Shu Wu, Xueli Yu, Qiang Liu, and Liang Wang, “Dynamic Graph Neural Networks for Sequential Recommendation", IEEE Trans. on Knowl. and Data Eng. 35 (5), 4741–4753,2023

work page 2023

[3] [3]

DiﬀuRec: A Diﬀusion Model for Sequential Recommendation

Zihao Li, Aixin Sun, and Chenliang Li, “DiﬀuRec: A Diﬀusion Model for Sequential Recommendation", ACM Trans. Inf. Syst. 42 (3),1–28, 2023

work page 2023

[4] [4]

Modeling Se- quences as Distributions with Uncertainty for Sequential Recommendation

Ziwei Fan, Zhiwei Liu, Shen Wang, Lei Zheng, and Philip S. Yu, “Modeling Se- quences as Distributions with Uncertainty for Sequential Recommendation", In Proceedings of the International Conference on Information and Knowledge Man- agement (CIKM), pp 3019–3023, 2021

work page 2021

[5] [5]

Sequential Recommendation via Stochastic Self-Attention

Ziwei Fan, Zhiwei Liu, Yu Wang, Alice Wang, Zahra Nazari, Lei Zheng, Hao Peng, and Philip S. Yu, “Sequential Recommendation via Stochastic Self-Attention", In Proceedings of the ACM Web Conference (WWW ’22). Association for Computing Machinery, New York, NY, USA, 2036–2047, 2022

work page 2022

[6] [6]

Factorizing personalized Markov Chains for next-basket recommendation

S. Rendle, C. Freudenthaler, and L. Schmidt-Thieme, “Factorizing personalized Markov Chains for next-basket recommendation", In Proceedings of the Interna- tional conference on World Wide Web, pp. 811–820, 2010

work page 2010

[7] [7]

Personalizing session- based recommendations with hierarchical recurrent neural networks

M. Quadrana, A. Karatzoglou, B. Hidasi, and P. Cremonesi, “Personalizing session- based recommendations with hierarchical recurrent neural networks", In Proceed- ings of the Eleventh ACM Conference on Recommender Systems, pp. 130–137, 2017

work page 2017

[8] [8]

Recurrent neural networks with top-k gains for session-based recommendations

B. Hidasi and A. Karatzoglou, “Recurrent neural networks with top-k gains for session-based recommendations", In Proceedings of the 27th ACM International Conference on Information and Knowledge Management, 2018, pp. 843–852

work page 2018