Give the Long-tail More SPACE: Promoting Provider Fairness in Next POI Recommendation

Anran Zhang; Jiahui Jin; Jiaqi Jiang; Yuhan Zhao

REVIEW 4 major objections 5 minor 53 references

Give the Long-tail More SPACE: Promoting Provider Fairness in Next POI Recommendation

T0 review · 4 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 A model-agnostic training-data augmentation layer, SPACE, shifts next-POI exposure toward long-tail venues by generating feasibility- and supply-constrained virtual users, improving fairness and usually accuracy.

desk verdict A genuinely useful long-tail fairness idea held back by a missing training objective for its core generator—worth refereeing, but the central claim is not yet reproducible as written. read the letter →

arxiv 2608.07998 v1 pith:L6DEBGLU submitted 2026-08-08 cs.IR

Anran Zhang , Jiaqi Jiang , jiahui Jin , Yuhan Zhao This is my paper · ORCID

classification cs.IR

keywords nextPOIrecommendationproviderfairnesslong-tailexposurevirtualusergenerationoptimaltransportlatentdiffusionexecutionconstraintssupply

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 argues that mainstream next-POI recommenders concentrate exposure on popular venues, and that existing fairness fixes fail in this domain because users cannot always physically execute recommendations and venues have capacity limits. SPACE addresses this by generating virtual user–POI training pairs under explicit feasibility and supply control. The framework first clusters users into communities reflecting mobility constraints, then uses unbalanced optimal transport to allocate virtual-user quotas to long-tail POIs within supply budgets, then generates community-consistent user embeddings with a latent diffusion model. On three real-world datasets, adding SPACE to several backbone recommenders improves long-tail exposure and cold-POI coverage while usually also improving accuracy. The practical claim is that a recommender can be made fairer by changing its training data rather than its architecture.

What carries the argument

The load-bearing mechanism is a three-stage generation pipeline. Stage 1 learns community prototypes from dual behavioral views—a preference histogram and a mobility/constraint histogram—through soft clustering over fused user embeddings (Eq. 12–14). Stage 2 solves an unbalanced optimal-transport plan $P_{v,k}$ that decides how many virtual users each tail POI $v$ draws from community $k$, with a KL penalty pulling each POI's total allocation toward its supply budget $d_v$ (Eq. 15). Stage 3 generates each virtual user embedding $\tilde{e}_u$ by iterative residual denoising that starts from a prototype-aware state $s^{(0)}=e_{base}+\alpha \delta_k$, conditions on the POI embedding, the community prototype, and batch-level statistics, and applies an execution regularizer (Eq. 22) equal to a ReLU on $\|\tilde{e}_u-e_u\|_2^2-R_k^2$ to keep generated users inside the community's constraint manifold. The augmented pairs are then mixed into the training loss with weight $\lambda_{rec}$.

What would settle it

Collect ground-truth travel times and per-user budgets for one of the three datasets and measure the share of SPACE-generated virtual user–POI pairs whose travel time exceeds the sampled user's budget. The physical-executability claim is falsified if that share is not near zero, or if removing the community-consistency penalty leaves fairness unchanged; the supply claim is falsified if per-POI augmented demand systematically exceeds measured venue capacity.

Watch

Extended reading notes

Core claim

The paper's central claim is that provider fairness in next-POI recommendation can be improved without the usual accuracy cost by changing the training data rather than the model or the ranking. SPACE synthesizes virtual users who are conditioned on a target long-tail POI and pulled toward a learned community of real users, so the generated user–POI pairs carry the POI's exposure signal while staying inside the community's execution and supply envelope. The authors argue that this sidesteps the two failure modes of re-ranking-based fairness: recommendations that users cannot physically execute, and long-tail POIs that receive more demand than their capacity allows. Across NYC, TKY, and CA, adding SPACE to five backbone recommenders improves long-tail exposure and coverage metrics while maintaining or improving hit rate and NDCG in most configurations. The core discovery is therefore an augmentation recipe: feasibility- and supply-constrained virtual users can redistribute exposure toward the long tail without breaking the recommender's accuracy.

Load-bearing premise

The load-bearing premise is that the community prototypes learned from users' activity-time and movement histograms faithfully encode each user's real mobility budget, so keeping generated virtual users near those prototypes guarantees that their recommended POIs are physically executable; the paper does not directly measure travel cost or budget.

Editorial extensions

If this is right

SPACE is model-agnostic: the same synthesized user–POI pairs can be fed into FPMC, LSTPM, GETNext, MTNet, or DiffuRec without changing their architectures.
On three real-world city datasets, adding SPACE improves long-tail exposure and coverage—for example GETNext on CA raises long-tail coverage from 0.3120 to 0.4672—while keeping or improving hit rate and NDCG.
Fairness gains are not confined to one metric: cold-warm group fairness, cold-POI exposure, long-tail coverage, NC@K, and GINI all move in the fairer direction.
The extra training cost is small in the reported experiments, so the augmentation layer is practical to deploy.

Reading between the lines

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

Editorial extension: because SPACE changes only the training data, its gains could be stacked with post-processing or re-ranking fairness methods, potentially pushing exposure further toward the tail.
Editorial extension: the supply budget $d_v$ is set from historical visit behavior; estimating it from actual capacity data and testing sensitivity to that estimate would make the supply guarantee more direct.
Editorial extension: the same community-quota-generation recipe could transfer to other two-sided physical markets, such as ride-hailing or restaurant reservations, where user feasibility and provider capacity both bind.

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.

Desk Editor's Note

read the letter

The paper's core idea is good: instead of re-ranking or adding fairness losses, generate virtual users that respect both user-side execution budgets and POI-side supply caps, then train any backbone on the augmented data. That framing is new to POI fairness, and the three-stage design—community inference, unbalanced OT allocation, constraint-guided latent diffusion—is a plausible division of labor. The empirical sweep is broad: three real-world datasets, five backbones, and consistent fairness gains plus frequent accuracy gains. The theoretical preamble (NP-hardness, unbounded price of fairness) is simple but correctly motivates why soft-penalty approaches are insufficient.

Now the soft spots. The load-bearing one is exactly what the stress-test note says: there is no training objective for the denoising network epsilon_theta. Equations (19)–(20) give a residual update rule, and L_exec in (22) only penalizes distance to a community prototype. No reconstruction loss, no denoising objective, no distribution-matching term. As written, epsilon_theta is an unspecified function that may never be trained at all. The downstream recommender loss (24) is applied after generation, so it cannot shape epsilon_theta unless the authors intend end-to-end backprop through the latent refinement—which the text contradicts by saying 'after generation.' The w/ogen ablation cannot resolve this, because removing generation also removes the initialization and the constraints. This is not a cosmetic omission; it means the mechanism credited for the gains—trained constraint-guided diffusion—is not actually described. The public code may clarify, but the paper must state the objective and optimizer explicitly.

Other issues are secondary but real: no standard deviations or test repetitions; the 'p <= 0.05' in Table 2 is unsupported; the aggregate improvement percentages are undefined; and no comparison to existing provider-fairness baselines (FairAgent, regularization-based methods, re-ranking). The 'physically executable' claim rests on activity-time and movement histograms rather than direct cost measurements, which is a reasonable proxy but should be acknowledged as such. These are fixable in revision.

If the missing objective is resolved and the fairness baselines added, this could become a solid contribution. As it stands, it is a promising idea with suggestive results, but the central mechanism is underspecified. I'd send it to peer review—the idea deserves referee time—and require the authors to clarify the training procedure, add comparisons to prior fairness methods, and report variance. If the diffusion network turns out to be untrained, the paper should be reframed as a perturbation-based augmentation method, which may still work but is a different claim.

Referee Report

4 major / 5 minor

Summary. SPACE is a model-agnostic training-data augmentation framework for next-POI recommendation. It first learns user communities from preference and constraint views via a soft-clustering auxiliary loss (Section 3.1), then allocates virtual-user quotas to long-tail POIs with an unbalanced optimal-transport objective that penalizes deviation from per-POI supply budgets (Section 3.2), and finally generates virtual user embeddings by iterative residual refinement of a prototype-anchored initial state (Section 3.3). The generated user-POI pairs are added to the training set, and the recommendation loss is minimized jointly with a weighting hyperparameter (Eq. 24). Experiments on NYC, TKY, and CA with five backbone recommenders report improved long-tail exposure and competitive or better accuracy.

Significance. If the proposed framework works as described, the paper contributes a useful alternative to re-ranking and post-processing for POI provider fairness: it is model-agnostic, explicitly encodes supply budgets, and avoids changing recommender architectures. The public code release, the breadth of backbones (FPMC, LSTPM, GETNext, MTNet, DiffuRec), and the three datasets are concrete strengths. However, the central generation mechanism is not fully specified, the feasibility proxy is indirect, and the statistical support is incomplete; these issues must be resolved before the empirical claims can be fully credited.

major comments (4)

[§3.3.2, Eqs. (19)–(21)] Equations (19)–(21) define an iterative residual update but no training objective for the denoising network ε_θ. There is no reconstruction loss, denoising loss, distribution-matching term, or end-to-end gradient path specified; the only losses in the paper are L_infer (Eq. 14), L_exec (Eq. 22), and downstream L_rec (Eq. 24), and Section 3.4 states that L_rec is applied after generation. As written, ε_θ is therefore an untrained network applied to s(0), and the claimed 'constraint-guided latent diffusion' cannot be credited with the observed improvements. The w/ogen ablation does not resolve this, because removing generation also removes the prototype-anchored initialization and the constraint regularizer. Please specify the objective and optimizer for ε_θ, or state explicitly if ε_θ is trained end-to-end through Eq. (24) together with the recommender, and describe the resulting procedure.
[§2.3, Theorem 2.2] The NP-hardness proof is not valid as stated. The reduced instance in Eqs. (3)–(5) is a capacitated assignment problem with unit demands and integer capacities, which is a maximum-weight bipartite b-matching / min-cost flow problem and is solvable in polynomial time, not NP-hard in general. Consequently, the claim that the original constrained allocation problem is NP-hard is unproven; the later assertion that continuous optimization cannot give feasibility guarantees may still be true, but it needs a correct argument or a different reduction.
[§4.1.3 and Table 2] The caption of Table 2 states 'p <= 0.05', but the paper describes no statistical test, no number of random seeds, and no standard deviation or confidence interval for any reported metric. The rows 'Accuracy Improvement' and 'Fairness Improvement' are also not defined: the reader cannot tell whether these are averages of relative improvements over HR@1, HR@10, NDCG@10 or over the fairness metrics, and the calculation sometimes appears to combine metrics with different scales. Please report variance, specify the significance test and null hypothesis, and define the aggregate improvement formulas.
[§3.3.2 and §4] The 'physically executable' claim is not directly validated. The paper defines execution feasibility via c_{u,v} and B_u in Section 2.2, but the generation constraint in Eq. (22) only keeps the full virtual embedding close to a community center in latent space, and the community centers are learned from activity-time and movement histograms. No experiment measures whether generated users can actually reach their target POIs within budget, nor compares against actual travel costs. Please state that executability is enforced through learned proxy constraints, and provide, if possible, a direct evaluation of generated trajectories against travel-time or distance budgets.

minor comments (5)

[Figure 4] The x-axis label 'DiffuPOI' is inconsistent with the backbone name 'DiffuRec' used in Table 2 and the related work; please align the terminology.
[§4.4 and Eq. (14)] The hyperparameter λ_virtual is discussed in the hyperparameter study and Figure 3, but it is never defined in an equation; if it weights the L_infer term, please include it explicitly in Eq. (14).
[§4.1.4 and §2.1] The long-tail split threshold (top 20% popular) in Section 4.1.4 should be reconciled with the reference to the long-tail definition in [38] in Section 2.1, since the formal definition in Eq. (1) does not state how V_tail is determined.
[§4.3] There is a typo in 'denots' in the description of the w/os(0) variant; please correct it to 'denotes'.
[§4.1.3] The fairness metrics CGF@K, CE@K, and LTC are introduced by name but no formal definitions are given; for reproducibility, please provide the exact computations used for these metrics.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: SPACE's fairness gains are empirical, with only contextual self-citations.

full rationale

The central claim is not derived from its own inputs. SPACE improves tail-POI exposure by generating virtual users and training existing recommenders; the evaluation (Tables 2-5) measures held-out HR/NDCG and fairness (CGF, CE, LTC) across five backbone models on three datasets, so the reported gains are externally falsifiable rather than constructed by the objective. The supply budget d_v is set from historical maximum daily visits, and the UOT allocation (Eq. 15) minimizes a cost plus KL penalty; neither term is the fairness metric being reported. Community inference (Eqs. 9-14) and the generation regularizer (Eqs. 22-23) are optimization constraints, not redefinitions of the evaluation metrics. The NP-hardness and price-of-fairness theorems (Theorems 2.2, 2.3) are standard reductions and constructions independent of the method. Two self-citations appear ([40] as inspiration for community inference, [43] as an example of digital-only settings), but neither is load-bearing. Two non-circular limitations exist: physical executability is proxied by behavioral histograms rather than measured c_{u,v} and B_u, and no explicit training objective is given for epsilon_theta; these affect correctness and reproducibility, not circularity, because no equation reduces the claimed result to a fitted parameter. The w/ogen ablation removes generation together with initialization and constraints, so it cannot isolate epsilon_theta; that is an experimental attribution gap, not an equation-level reduction. Under the rubric, the minor contextual self-citations yield score 2.

Assumptions & free parameters 8 free parameters · 6 assumptions · 1 invented entities

The method relies on a number of tuned scalars and data-derived budgets, not all reported. Its physical-feasibility guarantee is indirect: it assumes behavioral histograms encode execution constraints and that closeness to a community prototype implies executability. No new physical entities are posited.

free parameters (8)

Number of communities K = not reported
The soft clustering in Eq.12 requires the community count K; no value or selection criterion is given, and results likely depend on it.
Temperature tau for community softmax = not reported
Eq.12 uses temperature tau in the soft assignment; the value is not listed.
Exploration strength alpha = tuned, values not reported
Eq.18 scales the real-user deviation delta_k; Figure 3 shows sensitivity but no numeric values.
Supply budget vector d_v = e.g., historical max daily visits, per-POI values not reported
Eq.15 requires per-tail-POI supply budgets; the text suggests historical maximum daily visits but no exact construction or values are given.
Diffusion refinement steps T = not reported
Eq.20 sets step size eta_t=1/T, but T is not reported.
Loss weights lambda_virtual, lambda_rec, lambda_exec = tuned, values not reported
Figure 3 shows sensitivity for lambda_virtual, alpha, and lambda_rec, but numeric ranges or final choices are not in the paper.
Quantile q for community radius R_k = not reported
Eq.23 computes R_k using quantile q, which is not specified.
Long-tail split threshold (top 20% popular) = 0.20
Section 4.1.4 defines long-tail as the bottom 80% of POIs; this choice determines Vtail and all fairness metrics.

assumptions (6)

standard math The generalized assignment problem is NP-hard.
Theorem 2.2 reduces to a capacitated assignment problem and relies on this standard result.
standard math NP-hardness of a special case transfers to the general problem.
The K=1, fairness-free case is a subproblem of the original allocation problem, so proving it NP-hard implies the general problem is NP-hard.
domain assumption Behavioral histograms (activity time, movement statistics) are proxies for execution constraints.
Section 3.1.1 defines the constraint view x^c_u; physical budgets B_u are never measured directly.
ad hoc to paper Closeness to a community center implies physical executability.
Eq.22 uses a distance penalty to the community center as a surrogate for feasibility, but no direct feasibility check is used.
domain assumption Virtual user training pairs transfer to improved real-user ranking.
The central mechanism of SPACE is that augmented synthetic interactions improve downstream recommender performance; this is supported only by the experiments.
standard math The unbalanced optimal transport problem in Eq.15 is tractable via standard solvers.
The paper assumes the KL-regularized OT objective can be minimized, but does not specify the solver or convergence conditions.

invented entities (1)

Virtual user embeddings
purpose: Synthetic training examples paired with tail POIs to boost their exposure signals
They are internal data-augmentation constructs with no direct falsifiable prediction outside the method; their utility is only indirectly evaluated through downstream recommendation metrics.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Give the Long-tail More SPACE: Promoting Provider Fairness in Next POI Recommendation." pith.science (2026). https://pith.science/paper/L6DEBGLU

@misc{pith2026260807998,
  author       = {Pith},
  title        = {Pith review of: Give the Long-tail More SPACE: Promoting Provider Fairness in Next POI Recommendation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L6DEBGLU}},
  note         = {Machine review of arXiv:2608.07998}
}

read the original abstract

Next point-of-interest (POI) recommendation predicts users' future destinations from historical mobility sequences and has become a key component of location-based services. However, mainstream models often concentrate exposure on a small set of popular POIs, leaving long-tail merchants systematically under-exposed. While provider fairness has recently attracted increasing attention, directly applying existing provider-fairness techniques to POI recommendation is problematic: (i) users face execution constraints; and (ii) POIs face resource supply constraints. To address this, we propose SPACE (Supply- and Physics-Aware Conditional Embedding generation), a model-agnostic framework that improves long-tail POI exposure via virtual user generation under explicit feasibility and supply control. SPACE consists of three stages: (1) community inference to capture heterogeneous user execution constraints; (2) unbalanced optimal-transport allocation to decide how many virtual users each tail POI should receive from which communities under POI-specific supply budgets; and (3) constraint-guided latent diffusion to generate POI-conditional, community-consistent virtual user embeddings. The generated user-POI pairs can be seamlessly used to train existing recommenders without modifying their architectures. Extensive experiments on three real-world datasets demonstrate that SPACE substantially improves provider fairness while maintaining and often improving recommendation accuracy across multiple backbone models. Our code is publicly available at https://github.com/Anniran1/SPACE-main.

Figures

Figures reproduced from arXiv: 2608.07998 by the authors.

**Figure 2.** 3.1 Latent Constraint Discovery through Community Inference Inspired by prior community inference studies [2, 6, 40], we divide users into different communities according to their inherent constraints. Intuitively, we expect this module to learn communities with the following properties: (1) users assigned to the same community are relatively consistent in constraints; (2) we allow correlations between preference … view at source ↗

**Figure 2.** The illustration of our proposed SPACE. constraint view x 𝑐 𝑢 , which captures the user’s mobility pattern using a normalized hourly activity histogram, a day-of-week histogram, geographic statistics of visited locations, and the average trajectory length. The former mainly reflects historical POI choices (preference), while the latter mainly reflects mobility regularities (constraints). Then, we map the two views … view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Training time of models on the NYC dataset, for [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 27 canonical work pages

[1]

Robin Burke, Nasim Sonboli, and Aldo Ordonez-Gauger. 2018. Balanced neigh- borhoods for multi-sided fairness in recommendation. InConference on fairness, accountability and transparency. PMLR, 202–214

work page 2018
[2]

Jiazun Chen, Yikuan Xia, Jun Gao, Zhao Li, and Hongyang Chen. 2025. Com- munitydf: A guided denoising diffusion approach for community search. In2025 IEEE 41st International Conference on Data Engineering (ICDE). IEEE, 460–473

work page 2025
[3]

Weixin Chen, Yuhan Zhao, Li Chen, and Weike Pan. 2025. Leave no one behind: Fairness-aware cross-domain recommender systems for non-overlapping users. InProceedings of the Nineteenth ACM Conference on Recommender Systems. 226– 236

work page 2025
[4]

Salim, Yonchanok Khaokaew, and Jeffrey Chan

Francis Zac Dela Cruz, Flora D. Salim, Yonchanok Khaokaew, and Jeffrey Chan

work page
[5]

Marco Cuturi. 2013. Sinkhorn distances: Lightspeed computation of optimal transport.Advances in neural information processing systems26 (2013)

2013
[6]

Jinfei Gao, Xiao Wang, Tian Gan, Jianhua Yin, Chuanchen Luo, and Liqiang Nie. 2025. Social Context-Aware Community-Level Propagation Prediction. In Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval. 1995–2005

work page 2025
[7]

Huizhong Guo, Zhu Sun, Dongxia Wang, Tianjun Wei, Jinfeng Li, and Jie Zhang
[8]

Huizhong Guo, Dongxia Wang, Zhu Sun, Haonan Zhang, Jinfeng Li, and Jie Zhang. 2024. Configurable fairness for new item recommendation considering entry time of items. InProceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. 437–447

work page 2024

Show all 53 references

[9]

Jing He, Xin Li, Lejian Liao, Dandan Song, and William Cheung. 2016. Inferring a personalized next point-of-interest recommendation model with latent behavior patterns. InProceedings of the AAAI conference on artificial intelligence, Vol. 30

2016
[10]

Jonathan Ho, Ajay Jain, and Pieter Abbeel. 2020. Denoising diffusion probabilistic models.Advances in neural information processing systems33 (2020), 6840–6851

2020
[11]

Siyuan Huang, Jiahui Jin, Xin Lin, Xigang Sun, and Yukun Ban. 2025. IM-POI: Bridging ID and Multi-modal Gaps in Next POI Recommendation. InProceedings of the 33rd ACM International Conference on Multimedia. 5979–5987

2025
[12]

Tianhao Huang, Xuan Pan, Xiangrui Cai, Ying Zhang, and Xiaojie Yuan. 2024. Learning time slot preferences via mobility tree for next poi recommendation. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38. 8535–8543

2024
[13]

Wang-Cheng Kang and Julian McAuley. 2018. Self-attentive sequential recom- mendation. In2018 IEEE international conference on data mining (ICDM). IEEE, 197–206

2018
[14]

Thomas N Kipf and Max Welling. 2016. Semi-supervised classification with graph convolutional networks.arXiv preprint arXiv:1609.02907(2016)

2016 arXiv
[15]

Dejiang Kong and Fei Wu. 2018. HST-LSTM: A hierarchical spatial-temporal long-short term memory network for location prediction.. InIjcai, Vol. 18. 2341– 2347

2018
[16]

Peibo Li, Maarten de Rijke, Hao Xue, Shuang Ao, Yang Song, and Flora D Salim
[17]

Zihao Li, Aixin Sun, and Chenliang Li. 2023. Diffurec: A diffusion model for sequential recommendation.ACM Transactions on Information Systems42, 3 (2023), 1–28

2023
[18]

Bo Liu, Jun Zeng, Haoran Tang, Junhao Wen, Min Gao, and Wei Zhou. 2025. AGSRec: Mitigating Popularity Bias via Attraction-Aware Graph and Semantic Enhancement for POI Recommendations.Expert Systems with Applications(2025), 130298

2025
[19]

In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval

Large language models for next point-of-interest recommendation. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. 1463–1472
[20]

Yingtao Luo, Qiang Liu, and Zhaocheng Liu. 2021. Stan: Spatio-temporal attention network for next location recommendation. InProceedings of the web conference

2021
[21]

Yifang Qin, Hongjun Wu, Wei Ju, Xiao Luo, and Ming Zhang. 2023. A diffusion model for poi recommendation.ACM Transactions on Information Systems42, 2 (2023), 1–27

2023
[22]

Zhao Liu, Yichen Zhu, Yiqing Yang, Xiao Lv, Guoping Tang, Rui Huang, Qiang Luo, Ruiming Tang, and Guorui Zhou. 2026. Diffgrm: Diffusion-based generative recommendation model. InProceedings of the ACM Web Conference 2026. 5853– 5864

2026
[23]

Steffen Rendle, Christoph Freudenthaler, and Lars Schmidt-Thieme. 2010. Factor- izing personalized markov chains for next-basket recommendation. InProceedings of the 19th international conference on World wide web. 811–820

2010
[24]

Wondo Rhee, Sung Min Cho, and Bongwon Suh. 2022. Countering popularity bias by regularizing score differences. InProceedings of the 16th ACM conference on recommender systems. 145–155

2022
[25]

Theresia Veronika Rampisela, Maria Maistro, Tuukka Ruotsalo, Falk Scholer, and Christina Lioma. 2025. Stairway to fairness: Connecting group and individ- ual fairness. InProceedings of the Nineteenth ACM Conference on Recommender Systems. 677–683

2025
[26]

Tobias Schnabel, Adith Swaminathan, Ashudeep Singh, Navin Chandak, and Thorsten Joachims. 2016. Recommendations as treatments: Debiasing learning and evaluation. Ininternational conference on machine learning. PMLR, 1670– 1679

2016
[27]

Oleksandr Shchur and Stephan Günnemann. 2019. Overlapping community detection with graph neural networks.arXiv preprint arXiv:1909.12201(2019)

2019 arXiv
[28]

Ruslan Salakhutdinov and Andriy Mnih. 2008. Bayesian probabilistic matrix factorization using Markov chain Monte Carlo. InProceedings of the 25th interna- tional conference on Machine learning. 880–887

2008
[29]

Tianao Sun, Ke Fu, Weiming Huang, Kai Zhao, Yongshun Gong, and Meng Chen. 2024. Going where, by whom, and at what time: Next location prediction considering user preference and temporal regularity. InProceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data...

2024
[30]

Wenjie Wang, Fuli Feng, Xiangnan He, Xiang Wang, and Tat-Seng Chua. 2021. Deconfounded recommendation for alleviating bias amplification. InProceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining. 1717– 1725

2021
[31]

Ashudeep Singh and Thorsten Joachims. 2018. Fairness of exposure in rankings. InProceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining. 2219–2228

2018
[32]

Haolun Wu, Chen Ma, Bhaskar Mitra, Fernando Diaz, and Xue Liu. 2022. A Multi-Objective Optimization Framework for Multi-Stakeholder Fairness-Aware Recommendation.ACM Trans. Inf. Syst.41, 2, Article 47 (Dec. 2022), 29 pages. doi:10.1145/3564285

2022 doi
[33]

Yao Wu, Jian Cao, Guandong Xu, and Yudong Tan. 2021. Tfrom: A two-sided fairness-aware recommendation model for both customers and providers. In Proceedings of the 44th international ACM SIGIR conference on research and devel- opment in information retrieval. 1013–1022

2021
[34]

Tianjun Wei, Jianghong Ma, and Tommy WS Chow. 2023. Collaborative residual metric learning. InProceedings of the 46th international ACM SIGIR conference on research and development in information retrieval. 1107–1116

2023
[35]

Dingqi Yang, Daqing Zhang, Vincent W Zheng, and Zhiyong Yu. 2014. Modeling user activity preference by leveraging user spatial temporal characteristics in LBSNs.IEEE Transactions on Systems, Man, and Cybernetics: Systems45, 1 (2014), 129–142

2014
[36]

Song Yang, Jiamou Liu, and Kaiqi Zhao. 2022. Getnext: trajectory flow map enhanced transformer for next poi recommendation. InProceedings of the 45th International ACM SIGIR Conference on research and development in information retrieval. 1144–1153

2022
[37]

Xiaodong Yan, Tengwei Song, Yifeng Jiao, Jianshan He, Jiaotuan Wang, Ruopeng Li, and Wei Chu. 2023. Spatio-temporal hypergraph learning for next POI rec- ommendation. InProceedings of the 46th international ACM SIGIR conference on research and development in information retrie...

2023
[38]

Hongzhi Yin, Bin Cui, Jing Li, Junjie Yao, and Chen Chen. 2012. Challenging the long tail recommendation.arXiv preprint arXiv:1205.6700(2012)

2012 arXiv
[39]

Quan Yuan, Gao Cong, Zongyang Ma, Aixin Sun, and Nadia Magnenat Thalmann
[40]

Yuxuan Yang, Siyuan Zhou, He Weng, Dongjing Wang, Xin Zhang, Dongjin Yu, and Shuiguang Deng. 2024. Siamese learning based on graph differential equation for Next-POI recommendation.Applied Soft Computing150 (2024), 111086

2024
[41]

Yang Zhang, Fuli Feng, Xiangnan He, Tianxin Wei, Chonggang Song, Guohui Ling, and Yongdong Zhang. 2021. Causal intervention for leveraging popularity bias in recommendation. InProceedings of the 44th international ACM SIGIR conference on research and development in information...

2021
[42]

Pengpeng Zhao, Anjing Luo, Yanchi Liu, Jiajie Xu, Zhixu Li, Fuzhen Zhuang, Victor S Sheng, and Xiaofang Zhou. 2020. Where to go next: A spatio-temporal gated network for next poi recommendation.IEEE Transactions on Knowledge and Data Engineering34, 5 (2020), 2512–2524

2020
[43]

Yuhan Zhao, Weixin Chen, Li Chen, and Weike Pan. 2026. The Double-Edged Sword of Knowledge Transfer: Diagnosing and Curing Fairness Pathologies in Cross-Domain Recommendation.arXiv preprint arXiv:2601.21805(2026)

2026
[44]

Anran Zhang, Xingfen Wang, and Yuhan Zhao. 2025. HACD: Harnessing attribute semantics and mesoscopic structure for community detection. InProceedings of the Eighteenth ACM International Conference on Web Search and Data Mining. 616–624

2025
[45]

Lin Zhong, Lingzhi Wang, Xu Yang, and Qing Liao. 2025. Comapoi: A collabo- rative multi-agent framework for next poi prediction bridging the gap between trajectory and language. InProceedings of the 48th International ACM SIGIR Conference on Research and Development in Informa...

2025
[46]

Ziwei Zhu, Yun He, Xing Zhao, Yin Zhang, Jianling Wang, and James Caverlee
[47]

Ziwei Zhu, Jingu Kim, Trung Nguyen, Aish Fenton, and James Caverlee. 2021. Fairness among new items in cold start recommender systems. InProceedings of the 44th international ACM SIGIR conference on research and development in information retrieval. 767–776

2021
[48]

Yong Zheng. 2017. Multi-stakeholder recommendation: Applications and chal- lenges.arXiv preprint arXiv:1707.08913(2017)

2017 arXiv
[53]

Jiankai Zuo and Yaying Zhang. 2024. Diff-DGMN: A diffusion-based dual graph multiattention network for POI recommendation.IEEE Internet of Things Journal 11, 23 (2024), 38393–38409

2024
[2013]

InProceedings of the 36th international ACM SIGIR conference on Research and development in information retrieval

Time-aware point-of-interest recommendation. InProceedings of the 36th international ACM SIGIR conference on Research and development in information retrieval. 363–372
[2021]

InProceedings of the RecSys ’26, September 27-October 02, 2026, Minneapolis, MN, USA Zhang et al

Popularity-opportunity bias in collaborative filtering. InProceedings of the RecSys ’26, September 27-October 02, 2026, Minneapolis, MN, USA Zhang et al. 14th ACM international conference on web search and data mining. 85–93

2026
[2024]

InProceedings of the 18th ACM Conference on Recom- mender Systems(Bari, Italy)(RecSys ’24)

CAPRI-FAIR: Integration of Multi-sided Fairness in Contextual POI Rec- ommendation Framework. InProceedings of the 18th ACM Conference on Recom- mender Systems(Bari, Italy)(RecSys ’24). Association for Computing Machinery, New York, NY, USA, 924–928. doi:10.1145/3640457.3688170
[2025]

InProceed- ings of the 48th International ACM SIGIR conference on research and development in Information Retrieval

Enhancing new-item fairness in dynamic recommender systems. InProceed- ings of the 48th International ACM SIGIR conference on research and development in Information Retrieval. 1707–1716

Pith tools

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

[1] [1]

Robin Burke, Nasim Sonboli, and Aldo Ordonez-Gauger. 2018. Balanced neigh- borhoods for multi-sided fairness in recommendation. InConference on fairness, accountability and transparency. PMLR, 202–214

work page 2018

[2] [2]

Jiazun Chen, Yikuan Xia, Jun Gao, Zhao Li, and Hongyang Chen. 2025. Com- munitydf: A guided denoising diffusion approach for community search. In2025 IEEE 41st International Conference on Data Engineering (ICDE). IEEE, 460–473

work page 2025

[3] [3]

Weixin Chen, Yuhan Zhao, Li Chen, and Weike Pan. 2025. Leave no one behind: Fairness-aware cross-domain recommender systems for non-overlapping users. InProceedings of the Nineteenth ACM Conference on Recommender Systems. 226– 236

work page 2025

[4] [4]

Salim, Yonchanok Khaokaew, and Jeffrey Chan

Francis Zac Dela Cruz, Flora D. Salim, Yonchanok Khaokaew, and Jeffrey Chan

work page

[5] [5]

Marco Cuturi. 2013. Sinkhorn distances: Lightspeed computation of optimal transport.Advances in neural information processing systems26 (2013)

2013

[6] [6]

Jinfei Gao, Xiao Wang, Tian Gan, Jianhua Yin, Chuanchen Luo, and Liqiang Nie. 2025. Social Context-Aware Community-Level Propagation Prediction. In Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval. 1995–2005

work page 2025

[7] [7]

Huizhong Guo, Zhu Sun, Dongxia Wang, Tianjun Wei, Jinfeng Li, and Jie Zhang

[8] [8]

Huizhong Guo, Dongxia Wang, Zhu Sun, Haonan Zhang, Jinfeng Li, and Jie Zhang. 2024. Configurable fairness for new item recommendation considering entry time of items. InProceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. 437–447

work page 2024