Pith. sign in

REVIEW 4 major objections 6 minor 41 references

This paper argues that the head–tail performance seesaw in sequential recommenders stems from a 'collaborative overwhelming' effect in semantic-ID methods, and that a dual-branch architecture harmonizing semantic and hash IDs resolves it.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 17:09 UTC pith:KAHUGZDT

load-bearing objection A coherent dual-branch SID-HID recommender with plausible gains, but the main alignment mechanism rests on an unverified assumption about code quality, and the abstract oversells an online experiment absent from the body. the 4 major comments →

arxiv 2512.10388 v3 pith:KAHUGZDT submitted 2025-12-11 cs.IR cs.AI

The Best of the Two Worlds: Harmonizing Semantic and Hash IDs for Sequential Recommendation

Ziwei Liu , Yejing Wang , Wanyu Wang , Wang Zejian , Qidong Liu , Zijian Zhang , Chong Chen , Wei Huang
show 1 more author
Xiangyu Zhao
This is my paper
classification cs.IR cs.AI
keywords sequential recommendationsemantic IDshash IDsvector quantizationlong-tailcollaborative overwhelmingdual-branchalignment
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tries to prove that the long-standing trade-off between head and tail item performance in sequential recommendation is not inevitable. It identifies a mechanism it calls 'collaborative overwhelming': when items are represented only by quantized semantic codes, the codes' collisions wash out the unique identity that popular items need, while pure hash-ID embeddings lack the semantics that rare items need. The proposed H2Rec framework runs two branches in parallel—one embedding items by their semantic codes, one by their unique hash IDs—and aligns them with contrastive losses that let tail items borrow collaborative signal from semantically similar head items. On three datasets, the authors report consistent gains over baselines on both head and tail groups.

Core claim

The central claim is that H2Rec achieves the best of both identifier worlds: it keeps the uniqueness of hash IDs for modeling popular items while leveraging the multi-granularity of semantic IDs to improve rare items. The authors diagnose pure-SID methods with the 'collaborative overwhelming' phenomenon—quantization-induced code collisions inflate user-item connections and hurt head items—and pure-HID methods with noisy collaborative sharing and semantic homogeneity. Their remedy is a dual-branch architecture with a multi-granularity fusion network in the SID branch and a multi-granularity cross-attention network in the HID branch, where the HID embedding acts as the query that selectively p

What carries the argument

The central object is the harmonized dual-branch representation. In the HID branch, the multi-granularity cross-attention module uses the unique hash-ID embedding as a query to attend over the code-level semantic sequences (keys/values) and adds the original HID back via a residual connection; this is the mechanism that prevents collaborative overwhelming. In the SID branch, a multi-granularity fusion network weights code levels by user intent. The dual-level alignment defines a positive set P(i) that includes items sharing p semantic code levels or appearing in a local context window, and pulls SID and HID embeddings of those items together—this is how tail items are supposed to 'borrow' he

Load-bearing premise

The method's benefits rest on the assumption that semantic codes from RQ-VAE are a faithful, multi-granular decomposition of item meaning, so that items sharing p code levels are genuinely semantically related and can be trusted as positive training signals; if the codes are unstable, redundant, or poorly utilized (as the paper's Table 8 shows very low utilization at the default setting), the alignment could pull tail items toward semantically unrelated items and the reported

What would settle it

One experiment that could settle the claim: take a well-trained H2Rec model and replace the positive set P_C(i) in the code-guided alignment loss with items that share the same code levels but are known to be semantically unrelated (e.g., by independent textual similarity or human judgment). If tail performance holds or improves, the gains are not coming from semantic alignment; if tail performance collapses, the assumption about code semantics is load-bearing. Alternatively, train H2Rec with the alignment loss removed but all other components intact; the paper's ablation (w/o L_CA) shows a dr

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If H2Rec works as claimed, hybrid SID+HID recommenders can replace simple concatenation or contrastive fusion with a dual-branch architecture that yields consistent gains on both head and tail items.
  • The framework being quantization-agnostic and backbone-agnostic implies that existing sequential recommender systems can adopt H2Rec without redesigning their sequence encoders.
  • The analysis of code-matching threshold p shows that deeper semantic matching (p=3) filters noise better than coarse matching (p=1), indicating that code-sharing positives must be carefully selected.
  • Table 8's collision and utilization results imply that codebook size is a trade-off: larger codebooks reduce collisions but waste code capacity, so gains plateau.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The 'collaborative overwhelming' diagnosis likely generalizes beyond recommendation: any task that quantizes entities with heterogeneous frequency (e.g., generative retrieval, product search) may face the same uniqueness-vs-semantics seesaw, and the dual-branch plus query-anchored attention recipe could transfer.
  • A testable extension is to weight the code-guided alignment loss by inverse item popularity, so tail items dominate the positive set; this might amplify the reported tail gains beyond the current uniform formulation.
  • The framework's reliance on RQ-VAE code quality suggests that gains could be further improved by adding a quantization-aware training signal that directly optimizes low collision and high utilization, rather than treating codes as fixed side inputs.
  • The head–tail split at the top 20% may hide behavior at the extreme long tail; re-evaluating with finer popularity deciles (e.g., items with fewer than five interactions) would test whether the borrowing mechanism holds where sparsity is most severe.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. H2Rec is a dual-branch sequential recommendation framework that combines Semantic IDs (SID) generated by RQ-VAE with conventional Hash IDs (HID). The SID branch uses a multi-granularity fusion network, and the HID branch uses multi-granularity cross-attention; a dual-level alignment loss (code-guided item-level alignment and masked sequence granularity user-level alignment) bridges the two spaces. The paper reports consistent improvements over HID-, SID-, and hybrid-embedding baselines on Yelp, Beauty, and Instrument, with relative gains of 1.29%–11.88%, and argues that the method balances head- and tail-item quality.

Significance. If the empirical claims hold, the paper addresses a real limitation of SID-based recommenders—the trade-off between identifier uniqueness for head items and semantic generalization for tail items—and offers a modular, quantization-agnostic, backbone-agnostic enhancement. The ablation study in Table 3 and the generality checks in Tables 6–7 are informative and support the contribution of each component. However, the submission is not reproducible as presented (no code/data), the online-experiment claim in the abstract is not supported anywhere in the body, and the central code-guided alignment mechanism rests on an untested assumption about the semantic precision of RQ-VAE codes. These issues need to be addressed before the results can be taken at face value.

major comments (4)
  1. [Abstract / §4] The abstract promises "online experiments on a large-scale commercial platform," but the body reports only offline experiments on Yelp, Beauty, and Instrument. No online experiment, platform description, or metric appears in §4 or elsewhere. This claim must either be supported by a dedicated section or removed; as written, it is a missing-support discrepancy that affects the summary of the paper's contribution.
  2. [§3.3.1, Eq. (7); Table 8] The positive set in Eq. (7) treats every item sharing p=3 code levels as a trustworthy semantic neighbor. However, Table 8 reports a 22.28% full-SID collision rate and 0.004% codebook utilization at the default 4×128 setting. The paper never measures the semantic precision of P_C(i) and never compares L_CA against a random-positive baseline of matched size; Table 4's p sweep cannot distinguish genuine knowledge transfer from generic smoothing. Tail gains may reflect a large, loosely related positive set rather than multi-granular semantic transfer. Please add a random-positive control and report the size/precision distribution of P_C(i).
  3. [§4.5, Tables 4–5] The only hyperparameter study for p and o is on Yelp, yet p=3 and o=3 are used for all three datasets. No validation protocol or per-dataset sensitivity is shown for Beauty/Instrument, so the cross-dataset gains could be an artifact of Yelp-tuned hyperparameters. Report the tuning procedure or show that the results are stable for nearby values on each dataset.
  4. [§4.1.4, Table 2] All tables report single numbers without standard deviations, despite §4.1.4 saying results are averaged over three seeds. Table 2's "*" is defined by a two-sided t-test with p<0.05, but without variance or per-run values (or a description of paired testing) the reader cannot assess whether the smallest gains (e.g., Head H@10 +1.29% on Beauty) are significant.
minor comments (6)
  1. [Algorithm 1] Step 9 says "Derive the final SID item embedding sequence E_hid by Equation (3)" – it should be E_sid, since Eq. (3) defines E_sid.
  2. [§4.4 vs Figure 3] The text refers to "Figure 4" for hyperparameter results, but the caption is labeled "Figure 3." Please correct the cross-reference.
  3. [Table 2 / §4.2.1] The model is called "SFM-SID" in Table 2 and in §4.2.1, but the baseline list and references use "SPM-SID" [23]. Please standardize the name.
  4. [Eq. (10)] Eq. (10) writes P(v_{j+1}=v^+|S_u) but the recommendation score in Eq. (9) is P(v_{N+1}=v_j|S_u); the summation index j and the v_{j+1} notation are confusing. Rewrite with a clear positive/negative pair.
  5. [§4.6] The group analysis section contains a placeholder "Figure ??" for the detailed results; the figure needs to be included or the reference removed.
  6. [Abstract / §4.1.3] The abstract and implementation details state that the code is available online, but no code or data accompany the submission. Please provide the repository/artifact so the experiments can be reproduced.

Circularity Check

0 steps flagged

No circularity: the alignment losses are regularizers, the reported metrics come from held-out ranking, and the self-citations are not load-bearing.

full rationale

H2Rec is an empirical architecture paper, not a derivation from a fitted quantity. The code-guided alignment loss (Eq. 7) and masked sequence granularity loss (Eq. 8) are auxiliary training objectives; the reported head/tail metrics are computed from the ranking loss (Eq. 10) and the dot-product score (Eq. 9) on held-out next-item prediction. No fitted parameter or codebook setting is renamed as a prediction, and no equation reduces to its own input by construction. RQ-VAE is adopted from external work [14,22] rather than from the authors' prior results; the SID/HID baselines, including LLM-ESR [18] and MME-SID [32], are comparisons, not premises that force the outcome. The self-citations to the authors' prior LLM-ESR/SIGMA/LLMemb work appear as baselines and data-processing conventions and do not carry the central argument. Non-circular credibility concerns should be noted separately: the Abstract claims 'online experiments on a large-scale commercial platform', but Section 4.1.1 reports only three public datasets ('Yelp, Amazon Beauty, and Amazon Instrument'); and Table 8 shows 22.28% collision and 0.004% utilization at the default 4x128 SID setting, so the semantic-precision assumption behind P_C(i) in Eq. (7) needs stronger validation. These concerns affect evidence quality, not the circularity of the derivation.

Axiom & Free-Parameter Ledger

6 free parameters · 4 axioms · 0 invented entities

The method rests on standard neural-network learnable parameters plus user-chosen hyperparameters (β, γ, p, o, code levels). The largest unpurchased assumption is that RQ-VAE code overlap marks semantic similarity suitable for transferring collaborative signals; the paper's own Table 8 shows meaningful collision and low codebook utilization at the default setting.

free parameters (6)
  • β (alignment loss weight) = 0.5 (Yelp)
    Searched from {0.1,0.3,0.5,0.7,0.9}; optimum on Yelp and then used for all datasets.
  • γ (masked sequence granularity loss weight) = 0.3 (Yelp)
    Searched from the same set; optimum on Yelp.
  • code-matching threshold p = 3
    Number of shared RQ-VAE code levels required for a positive in Eq. (7); chosen on Yelp in Table 4.
  • context window size o = 3
    Local co-occurrence window for P_H(i); chosen on Yelp in Table 5.
  • code levels L and codebook size = 4×128
    Default SID configuration; chosen as a trade-off between collision (22.28%) and utilization in Table 8.
  • b_prior residual bias = learned vector
    Learnable vector in Eq. (2) enforcing a coarse-granularity prior; its form is an ad hoc modeling choice.
axioms (4)
  • domain assumption RQ-VAE codes preserve multi-granular item semantics useful for recommendation
    Central enabler of the SID branch (§2.2, §3.2.1). If codes are noisy or too colliding, the whole framework loses its advantage.
  • domain assumption Items sharing p code levels are semantically similar enough to share collaborative signals
    Defines the positive set P_C(i) in Eq. (7). Table 4 shows p=1 hurts, so the assumption is sensitive to the threshold.
  • domain assumption Items in a local context window of size o are relevant positives
    P_H(i) in Eq. (7) reintroduces co-occurrence signals that the paper elsewhere argues can be noisy (§1 vs. §3.3.1).
  • domain assumption Cosine similarity is a valid measure for aligning semantic and collaborative spaces
    Used in Eqs. (6)-(8) to define alignment and masked-sequence objectives; no independent evidence is given that cosine distance is the right metric for this transfer.

pith-pipeline@v1.3.0-alltime-deepseek · 16381 in / 10075 out tokens · 101013 ms · 2026-08-03T17:09:35.896805+00:00 · methodology

0 comments
read the original abstract

Conventional Sequential Recommender Systems (SRS) typically assign unique hash IDs (HID) to construct item embeddings, which mainly capture collaborative signals from historical user-item interactions. However, such embeddings are vulnerable in long-tail scenarios where most items are rarely consumed. Recent methods that incorporate auxiliary information often face noisy collaborative sharing from co-occurrence signals or semantic homogeneity caused by flat dense embeddings. In contrast, Semantic IDs (SID), with their support for code sharing and multi-granular semantic modeling, offer a promising alternative. Nevertheless, SID-based methods are hindered by a collaborative overwhelming phenomenon: commonly adopted quantization mechanisms compromise the identifier uniqueness needed to model head items, resulting in a performance trade-off between head and tail items. To address this challenge, we propose H2Rec, a novel framework that harmonizes SID and HID. We design a dual-branch modeling architecture that simultaneously captures the multi-granular semantics of SID while preserving the unique collaborative identity provided by HID. Moreover, we introduce a dual-level alignment strategy to bridge the two representations, enabling effective knowledge transfer and robust preference modeling. Extensive offline experiments on three public benchmarks and online experiments on a large-scale commercial platform demonstrate that H2Rec achieves a better balance between head and tail recommendation quality and consistently outperforms existing baselines.

Figures

Figures reproduced from arXiv: 2512.10388 by Chong Chen, Qidong Liu, Wang Zejian, Wanyu Wang, Wei Huang, Xiangyu Zhao, Yejing Wang, Zijian Zhang, Ziwei Liu.

Figure 1
Figure 1. Figure 1: Performance comparison with SASRec on Yelp (item [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The framework overview for H2Rec. In this figure, we set 𝐿 = 3. Specifically, for a semantic embedding 𝒆𝐿𝐿𝑀 derived from a lan￾guage model encoder, we leverage RQ-VAE to quantize it into 𝐿 semantic codes, where 𝐿 denotes the number of code layers. Specifi￾cally, the item𝑣𝑖 can be represented as a tuple𝐶𝑖 = {𝑐 1 𝑖 , . . . , 𝑐𝑙 𝑖 , . . . , 𝑐𝐿 𝑖 }, where 𝑐 𝑙 𝑖 is a single code whose embedding 𝒆 𝑙 𝑐𝑖 is the co… view at source ↗
Figure 3
Figure 3. Figure 3: Hyper-parameter Results on Yelp Dataset. Note that [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Detailed results in different item groups on the Yelp [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

41 extracted references · 10 linked inside Pith

  1. [1]

    Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, et al. 2023. Qwen technical report.arXiv preprint arXiv:2309.16609(2023)

    Pith/arXiv arXiv 2023

  2. [2]

    Qiwei Chen, Huan Zhao, Wei Li, Pipei Huang, and Wenwu Ou. 2019. Behavior sequence transformer for e-commerce recommendation in alibaba. InProceedings of the 1st international workshop on deep learning practice for high-dimensional sparse data. 1–4

    2019

  3. [3]

    Hui Fang, Danning Zhang, Yiheng Shu, and Guibing Guo. 2020. Deep learning for sequential recommendation: Algorithms, influential factors, and evaluations. ACM Transactions on Information Systems (TOIS)39, 1 (2020), 1–42

    2020

  4. [4]

    Robert Gray. 1984. Vector quantization.IEEE Assp Magazine1, 2 (1984), 4–29

    1984

  5. [5]

    Lei Guo, Hongzhi Yin, Qinyong Wang, Tong Chen, Alexander Zhou, and Nguyen Quoc Viet Hung. 2019. Streaming session-based recommendation. InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. 1569–1577

    2019

  6. [6]

    Yupeng Hou, Zhankui He, Julian McAuley, and Wayne Xin Zhao. 2023. Learning vector-quantized item representation for transferable sequential recommenders. InProceedings of the ACM Web Conference 2023. 1162–1171

    2023

  7. [7]

    Yupeng Hou, An Zhang, Leheng Sheng, Zhengyi Yang, Xiang Wang, Tat-Seng Chua, and Julian McAuley. 2025. Generative Recommendation Models: Progress and Directions. InCompanion Proceedings of the ACM on Web Conference 2025. 13–16

    2025

  8. [8]

    Jun Hu, Wenwen Xia, Xiaolu Zhang, Chilin Fu, Weichang Wu, Zhaoxin Huan, Ang Li, Zuoli Tang, and Jun Zhou. 2024. Enhancing sequential recommendation via llm-based semantic embedding learning. InCompanion Proceedings of the ACM Web Conference 2024. 103–111

    2024

  9. [9]

    Seongwon Jang, Hoyeop Lee, Hyunsouk Cho, and Sehee Chung. 2020. Cities: Contextual inference of tail-item embeddings for sequential recommendation. In 2020 IEEE International Conference on Data Mining (ICDM). IEEE, 202–211

    2020

  10. [10]

    Dietmar Jannach and Malte Ludewig. 2017. When recurrent neural networks meet the neighborhood for session-based recommendation. InProceedings of the eleventh ACM conference on recommender systems. 306–310

    2017

  11. [11]

    Herve Jegou, Matthijs Douze, and Cordelia Schmid. 2010. Product quantization for nearest neighbor search.IEEE transactions on pattern analysis and machine intelligence33, 1 (2010), 117–128

    2010

  12. [12]

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

    2018

  13. [13]

    Kibum Kim, Dongmin Hyun, Sukwon Yun, and Chanyoung Park. 2023. Melt: Mutual enhancement of long-tailed user and item for sequential recommendation. InProceedings of the 46th international ACM SIGIR conference on Research and development in information retrieval. 68–77

    2023

  14. [14]

    Doyup Lee, Chiheon Kim, Saehoon Kim, Minsu Cho, and Wook-Shin Han. 2022. Autoregressive image generation using residual quantization. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 11523–11532

    2022

  15. [15]

    Guanyu Lin, Zhigang Hua, Tao Feng, Shuang Yang, Bo Long, and Jiaxuan You

  16. [16]

    Enze Liu, Bowen Zheng, Wayne Xin Zhao, and Ji-Rong Wen. 2025. Bridging Textual-Collaborative Gap through Semantic Codes for Sequential Recommenda- tion. InProceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V. 2. 1788–1798

    2025

  17. [17]

    Qidong Liu, Xian Wu, Wanyu Wang, Yejing Wang, Yuanshao Zhu, Xiangyu Zhao, Feng Tian, and Yefeng Zheng. 2025. Llmemb: Large language model can be a good embedding generator for sequential recommendation. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 39. 12183–12191

    2025

  18. [18]

    Qidong Liu, Xian Wu, Yejing Wang, Zijian Zhang, Feng Tian, Yefeng Zheng, and Xiangyu Zhao. 2024. Llm-esr: Large language models enhancement for long- tailed sequential recommendation.Advances in Neural Information Processing Systems37 (2024), 26701–26727

    2024

  19. [19]

    Qidong Liu, Xiangyu Zhao, Yuhao Wang, Yejing Wang, Zijian Zhang, Yuqi Sun, Xi- ang Li, Maolin Wang, Pengyue Jia, Chong Chen, et al. 2024. Large Language Model Enhanced Recommender Systems: A Survey.arXiv preprint arXiv:2412.13432 (2024)

    Pith/arXiv arXiv 2024

  20. [20]

    Siyi Liu and Yujia Zheng. 2020. Long-tail session-based recommendation. In Proceedings of the 14th ACM conference on recommender systems. 509–514

    2020

  21. [21]

    Ziwei Liu, Qidong Liu, Yejing Wang, Wanyu Wang, Pengyue Jia, Maolin Wang, Zitao Liu, Yi Chang, and Xiangyu Zhao. 2025. SIGMA: Selective Gated Mamba for Sequential Recommendation.Proceedings of the AAAI Conference on Artificial Intelligence39, 12 (Apr. 2025), 12264–12272. https://doi.org/10.1609/aaai.v39i12. 33336

    doi:10.1609/aaai.v39i12 2025

  22. [22]

    Shashank Rajput, Nikhil Mehta, Anima Singh, Raghunandan Hulikal Keshavan, Trung Vu, Lukasz Heldt, Lichan Hong, Yi Tay, Vinh Tran, Jonah Samost, et al

  23. [23]

    Anima Singh, Trung Vu, Nikhil Mehta, Raghunandan Keshavan, Maheswaran Sathiamoorthy, Yilin Zheng, Lichan Hong, Lukasz Heldt, Li Wei, Devansh Tandon, et al. 2024. Better generalization with semantic ids: A case study in ranking for recommendations. InProceedings of the 18th ACM Conference on Recommender Systems. 1039–1044

    2024

  24. [24]

    Fei Sun, Jun Liu, Jian Wu, Changhua Pei, Xiao Lin, Wenwu Ou, and Peng Jiang

  25. [25]

    Bin Tan, Wangyao Ge, Yidi Wang, Xin Liu, Jeff Burtoft, Hao Fan, and Hui Wang

  26. [26]

    Jiaxi Tang and Ke Wang. 2018. Personalized top-n sequential recommenda- tion via convolutional sequence embedding. InProceedings of the eleventh ACM international conference on web search and data mining. 565–573

    2018

  27. [27]

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023. Llama: Open and efficient foundation language models.arXiv preprint arXiv:2302.13971(2023)

    Pith/arXiv arXiv 2023

  28. [28]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need.Advances in neural information processing systems30 (2017)

    2017

  29. [29]

    PCR-CA: Parallel Codebook Representations with Contrastive Alignment for Multiple-Category App Recommendation.arXiv preprint arXiv:2508.18166 (2025)

    arXiv 2025

  30. [30]

    Shijia Wang, Tianpei Ouyang, Qiang Xiao, Dongjing Wang, Yintao Ren, Songpei Xu, Da Guo, and Chuanjiang Luo. 2025. Progressive Semantic Residual Quantiza- tion for Multimodal-Joint Interest Modeling in Music Recommendation.arXiv preprint arXiv:2508.20359(2025)

    Pith/arXiv arXiv 2025

  31. [31]

    Wenjie Wang, Xinyu Lin, Fuli Feng, Xiangnan He, and Tat-Seng Chua. 2023. Generative recommendation: Towards next-generation recommender paradigm. arXiv preprint arXiv:2304.03516(2023)

    Pith/arXiv arXiv 2023

  32. [32]

    Yuhao Wang, Junwei Pan, Xinhang Li, Maolin Wang, Yuan Wang, Yue Liu, Dapeng Liu, Jie Jiang, and Xiangyu Zhao. 2025. Empowering Large Language Model for Sequential Recommendation via Multimodal Embeddings and Semantic IDs. arXiv preprint arXiv:2509.02017(2025)

    Pith/arXiv arXiv 2025

  33. [33]

    Shoujin Wang, Liang Hu, Yan Wang, Longbing Cao, Quan Z Sheng, and Mehmet Orgun. 2019. Sequential recommender systems: challenges, progress and prospects.arXiv preprint arXiv:2001.04830(2019)

    Pith/arXiv arXiv 2019

  34. [34]

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

    Pith/arXiv arXiv 2012

  35. [35]

    An Zhang, Yuxin Chen, Leheng Sheng, Xiang Wang, and Tat-Seng Chua. 2024. On generative agents in recommendation. InProceedings of the 47th international ACM SIGIR conference on research and development in Information Retrieval. 1807– 1817

    2024

  36. [36]

    Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, et al. 2023. A survey of large language models.arXiv preprint arXiv:2303.182231, 2 (2023)

    Pith/arXiv arXiv 2023

  37. [37]

    Yi Xu, Moyu Zhang, Chaofan Fan, Jinxin Hu, Xiaochen Li, Yu Zhang, Xiaoyi Zeng, and Jing Zhang. 2025. MMQ-v2: Align, Denoise, and Amplify: Adaptive Behavior Mining for Semantic IDs Learning in Recommendation.arXiv preprint arXiv:2510.25622(2025)

    arXiv 2025

  38. [41]

    Carolina Zheng, Minhui Huang, Dmitrii Pedchenko, Kaushik Rangadurai, Siyu Wang, Fan Xia, Gaby Nahum, Jie Lei, Yang Yang, Tao Liu, et al. 2025. Enhancing embedding representation stability in recommendation systems with semantic id. InProceedings of the Nineteenth ACM Conference on Recommender Systems. 954–957

    2025

  39. [2019]

    InProceedings of the 28th ACM international conference on information and knowledge management

    BERT4Rec: Sequential recommendation with bidirectional encoder rep- resentations from transformer. InProceedings of the 28th ACM international conference on information and knowledge management. 1441–1450

  40. [2023]

    Recommender systems with generative retrieval.Advances in Neural Information Processing Systems36 (2023), 10299–10315

    2023

  41. [2025]

    arXiv preprint arXiv:2502.16474(2025)

    Unified semantic and ID representation learning for deep recommenders. arXiv preprint arXiv:2502.16474(2025)

    Pith/arXiv arXiv 2025