Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

HiGR: Industrial-Scale Hierarchical Generative Slate Recommendation Framework in Tencent

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

Pith's one-line read HiGR shows that slate recommendation can be treated as a coarse-to-fine generation problem: first plan the list's preferences, then decode individual items.

desk verdict A novel and coherent industrial generative slate recommender with a real speedup, but the headline offline margin rests on a filtered proprietary test set and Eq. (12) needs a fix. read the letter →

arxiv 2512.24787 v5 pith:5SOADSOG submitted 2025-12-31 cs.IR cs.AI

classification cs.IRcs.AI
keywords SlaterecommendationGenerativeSemanticIDsResidualquantizedVAEHierarchicaldecodingPreferencealignmentORPOIndustrialrecommendersystems
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

HiGR argues that generative slate recommendation can be made fast and accurate enough for industrial use by structuring the item token space and splitting generation into two stages. Its central claim is that a contrastively trained residual quantizer produces semantic IDs whose prefixes are reliable meaning anchors, and that a two-stage decoder can then plan the whole slate at the level of preference embeddings before decoding each item. If this holds, generative recommenders need not trade speed for quality: a 10-item slate can be produced in roughly 10 planning steps plus fast parallel item decodes instead of 30 sequential token steps. The paper reports offline gains of over 10% against state-of-the-art baselines, a 5x inference speedup, and online A/B improvements of 1.22% in watch time and 1.73% in video plays on a large commercial platform.

What carries the argument

The central objects are (a) the Contrastive Residual Quantized VAE (CRQ-VAE), which adds temperature-scaled contrastive losses to the first D-1 layers of a residual quantizer so that similar items cluster under shared prefixes, plus a global quantization loss that prevents residual vanishing; (b) the preference embedding, defined as the sum of an item's SID embeddings, which serves as the unit of planning; (c) the Hierarchical Slate Decoder with a deep coarse-grained planner and a shallow, parameter-shared fine-grained item generator, using cross-attention to a user-context embedding; and (d) the Greedy-Slate Beam-Item inference strategy, which decodes items independently given planner-produ

What would settle it

A direct test of prefix semantic consistency: take a held-out set of items, group them by their first-level code, and ask human raters or an external semantic-similarity benchmark whether items sharing a prefix are more similar than items with different prefixes. A null result would mean the planner's preference embeddings are not semantically grounded. A second test would be to ablate the contrastive loss and show that planner quality and prefix consistency do not degrade, which would contradict the paper's mechanism.

Watch

Extended reading notes

Core claim

The paper introduces HiGR, a hierarchical generative slate recommendation framework. Its first component, CRQ-VAE, applies prefix-level contrastive constraints to residual quantization so that an item's high-level ID codes (prefixes) encode shared semantics, while the final code retains item-level discrimination. Its second component, the Hierarchical Slate Decoder, decouples generation into a coarse-grained slate planner that autoregressively predicts preference embeddings (sums of SID embeddings) for each slot, and a shared fine-grained item generator that independently decodes each predicted preference into an item's SID sequence. Its third component uses ORPO, a reference-model-free pref

Load-bearing premise

The approach assumes that the learned ID prefixes are genuinely semantic anchors, so that the sum of the code embeddings acts as a reliable 'preference embedding' for the planner; if the prefixes are not truly consistent, the two-stage design will propagate and amplify errors.

Editorial extensions

If this is right

  • If the design works as reported, generative slate recommenders can satisfy sub-100ms latency budgets: the paper reports a 5x inference speedup over the strongest autoregressive baseline.
  • Structured, prefix-aligned semantic IDs let the system impose diversity and relevance constraints directly on ID prefixes during decoding, rather than measuring them after the fact.
  • Listwise ORPO alignment with three complementary objectives (ranking fidelity, genuine interest, diversity) improves all offline metrics over no alignment and over DPO/SimPO alternatives.
  • The framework exhibits a power-law scaling relationship between model size and recommendation quality, suggesting that predictable gains come from simply increasing capacity.
  • Online A/B tests on a commercial platform serving hundreds of millions of users show a 1.22% increase in average watch time and a 1.73% increase in average video plays, indicating the approach translates to live business metrics.

Reading between the lines

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

  • An untested consequence is that the preference-embedding space, being decoupled from individual item IDs, may be transferable across domains or time periods; a slate planner trained in one content domain could potentially plan in another with only the item generator retrained.
  • The prefix-contrastive constraint is not specific to recommendation: any discrete tokenization that needs hierarchical semantic structure (e.g., code generation, molecular design, or hierarchical text representation) could benefit from the same contrastive prefix alignment.
  • If prefixes are truly stable semantic anchors, one could precompute a small set of 'preference archetypes' and constrain the planner to emit only those embeddings, enabling explicit user control or explainable slate composition without retraining.
  • The efficiency gain opens a new trade-off: the saved latency budget could be reinvested into larger beam widths or a deeper planner on the same hardware, potentially pushing quality further within fixed response-time limits.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. HiGR proposes a two-stage generative slate recommendation framework for industrial deployment. It first tokenizes items into hierarchical semantic IDs with a Contrastive RQ-VAE (CRQ-VAE), then decouples generation into a coarse-grained slate planner and a shared fine-grained item generator (HSD), and finally applies an ORPO-style listwise preference alignment with three negative-slate objectives. The paper reports offline gains over state-of-the-art baselines of more than 10% on a Tencent industrial dataset, a 5× inference speedup over OneRec, and online A/B improvements of 1.22% in average watch time and 1.73% in average video plays. The same architecture is also evaluated on KuaiRec, where gains over OneRec-25M are smaller (roughly 2.7-6.5% on three of the five reported metrics).

Significance. The framework is a credible industrial contribution: the hierarchical decomposition is well motivated, the online A/B evidence is valuable, and the reported deployment at Tencent scale is a real strength. If the empirical claims withstand scrutiny, the paper demonstrates that coarse-to-fine generative slate planning can be both effective and efficient enough for production. However, the headline quantitative claim is not yet robustly established. The >10% offline margin appears mainly on the proprietary, filtered industrial test set, while the public KuaiRec results are substantially smaller and are reported without error bars or significance tests. In addition, the ORPO loss in Equation (12) is written inconsistently, so the reported post-training gains cannot currently be traced to a reproducible objective. These are addressable issues, but they are load-bearing for the paper's central claim.

major comments (3)
  1. [4.2, Table 1, Appendix A] The claim of over 10% offline improvement is not robustly supported. On KuaiRec, HiGR-25M improves over OneRec-25M by only about 2.7% on Impression hit@5, 3.6% on Effective View hit@5, and 5.2% on NDCG@5; the >10% margin appears on the proprietary industrial metrics. Appendix A also states that all training data consists of filtered high-quality slate-level samples with duration thresholds, and Section 4.1.2 describes a test set constructed from slates with positive feedback. Such filtering can inflate both absolute and relative metrics. The authors should state explicitly whether the same filtering is applied to the test set, report all metrics with standard deviations or confidence intervals, and perform significance tests on both datasets. Without this, the headline margin is not verifiable.
  2. [3.5.2, Eq. (12)] The ORPO post-training objective is dimensionally inconsistent. Eq. (11) defines a token-level log-odds, and the text defines z as a sum over all tokens of a slate, i.e., a slate-level scalar. Eq. (12) then writes z_theta(x, y+_t) and z_theta(x, y-_t), which appear to be token-level quantities, and the first term uses y+_i<t without defining i. As written, the loss cannot be implemented or reproduced. The authors need to rewrite Eq. (12) with consistent notation, specifying whether the ORPO odds ratio is computed over the whole slate or over individual tokens, and how the supervised term is masked. Without a corrected objective, the ORPO gains in Table 4 are not attributable to a well-defined method.
  3. [3.4.1, 4.3.1] The design assumes that the sum of SID embeddings is a reliable preference embedding for list-level planning. This assumption is central to HSD: the planner conditions on those sums, and errors in the planner propagate to the item generator. Table 2 reports internal consistency and concentration metrics for CRQ-VAE, but these are self-defined and are not validated against an external semantic benchmark such as item-item retrieval, taxonomy alignment, or human evaluation. The authors should provide a more direct test of the assumption, such as ablating CRQ-VAE against RQ-VAE within the full HiGR pipeline, or showing that the prefix space supports semantic interpolation or clustering that matches an external label set. Without such evidence, the hierarchical planning step is only weakly grounded.
minor comments (5)
  1. [4.1.2] No error bars, standard deviations, or significance tests are reported for any offline result, including the KuaiRec comparisons. This should be added at least for the main comparison table.
  2. [3.4] In Eq. (10), the generated SID is denoted s-hat^{d+1}_m while the previous text indexes the output SID sequence as s-hat^1_m, ..., s-hat^D_m. This off-by-one notation should be fixed.
  3. [Table 2] The metrics Collision, Concentration, and Consistency are used to compare tokenization methods, but their formal definitions are not given in the main text or appendix. Please define them precisely.
  4. [Appendix C / Figure 4] The manuscript text after Figure 4 contains a large run of garbled '/uni...' glyph sequences. This appears to be a rendering artifact and should be removed before submission.
  5. [4.6] The online A/B test compares HiGR against the incumbent multi-stage system, but the description does not specify the duration of the test, the number of users in each arm, or whether the differences are statistically significant. Adding these details would strengthen the online claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: HiGR's claims are derived from explicit training objectives and evaluated on held-out and external benchmarks, with no prediction reducing to a fitted input.

full rationale

HiGR's derivation chain is self-contained in the relevant sense. CRQ-VAE is defined as a weighted combination of reconstruction loss, global quantization loss, and an InfoNCE contrastive term with stated coefficients (Eqs. 2-5). The HSD planner and item generator are trained with the cross-entropy loss in Eq. (10), and the ORPO-based post-training loss is explicitly written in Eq. (12) (though with a dimensional inconsistency that is a reproducibility/correctness issue, not circularity). Offline metrics are reported on a held-out test period and on the public KuaiRec dataset, and online A/B tests compare against an incumbent system, so the headline improvements are not obtained by re-predicting a fitted parameter. The paper contains no load-bearing self-citations: the cited generative baselines (TIGER, OneRec, HSTU, MTGR, etc.) are external work, not prior publications of the present authors. The filtered positive-feedback test set noted in Appendix A is an evaluation-validity concern, not a circularity concern, because the model is not fitted to those test labels and then reported as predicting them. The internal consistency metric in Table 2 would warrant suspicion if it merely re-measured the InfoNCE objective, but the paper does not define the metric precisely enough to exhibit that reduction, so under the hard-evidence rule it cannot be flagged as a circular step. Missing code and the undefined consistency metric are verification limitations, not evidence that the derivation is equivalent to its own inputs.

Assumptions & free parameters 10 free parameters · 4 assumptions · 2 invented entities

The framework depends on a dozen hyperparameters tuned on the industrial dataset, none with publicly reported search ranges. The contrastive prefix hypothesis and the preference-embedding definition are load-bearing design choices not independently validated. No external physical entities are introduced; the internal latent constructs are listed above.

free parameters (10)
  • contrastive loss weight λ2 = 0.01
    Coefficient of prefix-level InfoNCE loss in Eq. (5); tuned on industrial data.
  • global quant loss weight λ1 = 0.1
    Coefficient of global quantization loss in Eq. (5).
  • layer-wise quant trade-off η = 0.1
    Trade-off between the two terms in Eq. (2).
  • layer-wise contrast weights w_d = w1=1, w2=0.1, w3=0.01
    Weights applied to InfoNCE at each of the first D−1 layers in Eq. (4).
  • ORPO coefficient α = 0.1
    Scales the odds-ratio preference term in Eq. (12).
  • cosine similarity threshold for positives = 0.8
    Threshold determining semantic-neighbor positives vs negatives in CRQ-VAE training.
  • contrastive temperature τ = not reported
    Temperature in InfoNCE Eq. (4); missing from Appendix A, impeding reproduction.
  • codebook size K / depth D = 1024 / 3
    Semantic ID codebook size and number of residual levels; fixed without ablations in the paper.
  • planner/item generator layers (l_slate / l_item) = 14 / 2
    Model capacities; chosen heuristically.
  • beam width B and slate size M = 5 / 5
    Decoding beam width and slate length used in experiments.
assumptions (4)
  • domain assumption Prefix-level contrastive constraints yield semantically consistent SID prefixes that support controllable generation
    Central hypothesis of CRQ-VAE; supported only by internally defined consistency/concentration metrics, not an external semantic benchmark.
  • ad hoc to paper Sum of SID embeddings is an adequate preference embedding for list-level planning
    Defines ground-truth preference embedding i_m in §3.4.1; no principled justification, only an ablation comparing pooling strategies.
  • domain assumption ORPO listwise alignment with three hand-crafted negative types approximates real user slate preferences
    Assumes shuffled, negative-feedback, and semantically-dissimilar slates cover suboptimality relevant to user satisfaction; not validated against human judgments.
  • standard math RQ-VAE and InfoNCE losses behave as described in prior work
    Background techniques from TIGER [27] and InfoNCE [22].
invented entities (2)
  • Semantic ID prefix space
    purpose: A structured discrete vocabulary where high-level prefixes encode shared item semantics, used for controllable generation and diversity control
    Postulated by CRQ-VAE; the paper's own metrics (collision, concentration, consistency) are internal and not externally validated.
  • Slate preference embedding
    purpose: Sum of SID embeddings serving as a coarse-grained representation connecting the list planner to the item generator
    An internal latent construct with no outside-the-paper falsifiable prediction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HiGR: Industrial-Scale Hierarchical Generative Slate Recommendation Framework in Tencent." pith.science (2026). https://pith.science/paper/5SOADSOG

@misc{pith2026251224787,
  author       = {Pith},
  title        = {Pith review of: HiGR: Industrial-Scale Hierarchical Generative Slate Recommendation Framework in Tencent},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5SOADSOG}},
  note         = {Machine review of arXiv:2512.24787}
}
abstract

Slate recommendation, which presents users with a ranked item list in a single display, is ubiquitous across mainstream online platforms. While recent generative recommendation methods have shown strong potential in modeling item sequences with semantic IDs, directly applying them to industrial-scale slate recommendation faces a fundamental disconnect: entangled SID spaces confound high-level list planning, fine-grained autoregressive decoding over long sequences limits semantic planning efficiency, and token-level objectives misalign with holistic slate quality. In this paper, we propose HiGR, an industrial-scale hierarchical generative framework for slate recommendation that bridges this disconnect through a co-designed pipeline. First, HiGR learns structured SIDs via a Prefix-Contrastive Residual Quantized VAE (PCRQ-VAE). By enforcing high-level prefixes to capture shared semantics, PCRQ-VAE creates a controllable discrete space that acts as a prerequisite for efficient planning. Leveraging this structured space, our Hierarchical Slate Decoder (HSD) shifts autoregressive modeling from entangled token-level decoding to coarse-grained preference embeddings. This design significantly reduces inference latency while allowing explicit global slate structure planning. Finally, this stable planning space enables an ORPO-based listwise alignment mechanism to optimize triple-objective implicit feedback-ranking fidelity, genuine user interest, and diversity. Extensive offline experiments show that HiGR outperforms state-of-the-art baselines by over 10% in offline recommendation quality while achieving a $5\times$ inference speedup. Online A/B tests on Tencent platforms further improve watch time by 1.22% and video plays by 1.73%. HiGR has been deployed on multiple Tencent platform surfaces, serving hundreds of millions of users and proving its industrial-scale applicability.

Figures

Figures reproduced from arXiv: 2512.24787 by the authors.

Figure 1
Figure 1. The overall framework of HiGR, which contains (a) CRQ-VAE for semantic tokenization, (b) Hierarchical Slate Decoder [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The overall architecture of HSD. loss Lglobal_quan: Lglobal_quan = ∥𝑧ˆ − sg(𝑧) ∥2 2 + 𝜂∥𝑧 − sg(𝑧ˆ) ∥2 2 (3) which directly optimizes the global quantization error at the latent level. This explicitly counteracts residual collapse, ensuring deeper codebooks preserve meaningful semantics rather than decaying into noise, thus stabilizing the hierarchical learning process. 3.3.2 Prefix-level contrastive constraints to a… view at source ↗
Figure 3
Figure 3. Scaling trends of HiGR on convergence loss and [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Efficiency comparison of HiGR and OneRec. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Effects of different decoding lengths [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Bridging the Structural Gap: Adapting Autoregressive Generation for Recommendation

    cs.IR 2026-07 conditional novelty 6.0 of 10

    BARGE improves generative sequential recommendation by restoring item boundaries in the encoder and suppressing hierarchical semantic drift in decoding, outperforming prior generative baselines on public and industria...

  2. TriAlignGR: Triangular Multitask Alignment with Multimodal Deep Interest Mining for Generative Recommendation

    cs.IR 2026-05 unverdicted novelty 6.0 of 10

    TriAlignGR proposes a triangular multitask alignment framework with cross-modal semantic alignment, deep interest mining via chain-of-thought, and joint training on eight tasks to address content degradation and seman...

Reference graph

Works this paper leans on

45 extracted references · 13 linked inside Pith · cited by 2 Pith papers

  1. [1]

    Tesfaye Fenta Boka, Zhendong Niu, and Rama Bastola Neupane. 2024. A sur- vey of sequential recommendation systems: Techniques, evaluation, and future directions.Information Systems125 (2024), 102427

  2. [2]

    Gaode Chen, Ruina Sun, Yuezihan Jiang, Jiangxia Cao, Qi Zhang, Jingjian Lin, Han Li, Kun Gai, and Xinghua Zhang. 2024. A Multi-modal Modeling Framework for Cold-start Short-video Recommendation. InProceedings of the 18th ACM Conference on Recommender Systems. 391–400

  3. [3]

    Jianlv Chen, Shitao Xiao, Peitian Zhang, Kun Luo, Defu Lian, and Zheng Liu

  4. [4]

    Runjin Chen, Mingxuan Ju, Ngoc Bui, Dimosthenis Antypas, Stanley Cai, Xi- aopeng Wu, Leonardo Neves, Zhangyang Wang, Neil Shah, and Tong Zhao

  5. [5]

    Heng-Tze Cheng, Levent Koc, Jeremiah Harmsen, Tal Shaked, Tushar Chandra, Hrishi Aradhye, Glen Anderson, Greg Corrado, Wei Chai, Mustafa Ispir, et al

  6. [6]

    Enhancing item tokenization for generative recommendation through self-improvement.arXiv preprint arXiv:2412.17171(2024)

  7. [7]

    Sunhao Dai, Jiakai Tang, Jiahua Wu, Kun Wang, Yuxuan Zhu, Bingjun Chen, Bangyang Hong, Yu Zhao, Cong Fu, Kangle Wu, Yabo Ni, Anxiang Zeng, Wen- jie Wang, Xu Chen, Jun Xu, and See-Kiong Ng. 2025. OnePiece: Bringing Context Engineering and Reasoning to Industrial Cascade Ranking System. arXiv:2509.18091 [cs.IR] https://arxiv.org/abs/2509.18091

  8. [8]

    Romain Deffayet, Thibaut Thonet, Jean-Michel Renders, and Maarten De Rijke

Show all 45 references
  1. [9]

    Paul Covington, Jay Adams, and Emre Sargin. 2016. Deep neural networks for youtube recommendations. InProceedings of the 10th ACM conference on recommender systems. 191–198

  2. [10]

    Chongming Gao, Shijun Li, Wenqiang Lei, Jiawei Chen, Biao Li, Peng Jiang, Xiangnan He, Jiaxin Mao, and Tat-Seng Chua. 2022. KuaiRec: A Fully-observed Dataset and Insights for Evaluating Recommender Systems. InProceedings of the 31st ACM International Conference on Information ...

  3. [11]

    Huifeng Guo, Ruiming Tang, Yunming Ye, Zhenguo Li, and Xiuqiang He. 2017. DeepFM: a factorization-machine based neural network for CTR prediction.arXiv preprint arXiv:1703.04247(2017)

  4. [12]

    Ruidong Han, Bin Yin, Shangyu Chen, He Jiang, Fei Jiang, Xiang Li, Chi Ma, Mincong Huang, Xiaoguang Li, Chunzhen Jing, et al . 2025. Mtgr: Industrial- scale generative recommendation framework in meituan. InProceedings of the 34th ACM International Conference on Information an...

  5. [13]

    Jiaxin Deng, Shiyao Wang, Kuo Cai, Lejian Ren, Qigen Hu, Weifeng Ding, Qiang Luo, and Guorui Zhou. 2025. Onerec: Unifying retrieve and rank with generative recommender and iterative preference alignment.arXiv preprint arXiv:2502.18965 (2025)

  6. [14]

    Yupeng Hou, Jiacheng Li, Ashley Shin, Jinsung Jeon, Abhishek Santhanam, Wei Shao, Kaveh Hassani, Ning Yao, and Julian McAuley. 2025. Generating long semantic ids in parallel for recommendation. InProceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Min...

  7. [15]

    Peiyu Hu, Wayne Lu, and Jia Wang. 2025. From ids to semantics: A generative framework for cross-domain recommendation with adaptive semantic tokeniza- tion.arXiv preprint arXiv:2511.08006(2025)

  8. [16]

    Ray Jiang, Sven Gowal, Timothy A Mann, and Danilo J Rezende. 2018. Beyond greedy ranking: Slate optimization via list-CVAE.arXiv preprint arXiv:1803.01682 (2018)

  9. [17]

    Jiwoo Hong, Noah Lee, and James Thorne. 2024. Orpo: Monolithic preference optimization without reference model.arXiv preprint arXiv:2403.07691(2024)

  10. [18]

    Yu Lei, Wenjie Li, Ziyu Lu, and Miao Zhao. 2017. Alternating pointwise-pairwise learning for personalized item ranking. InProceedings of the 2017 ACM on Con- ference on Information and Knowledge Management. 2155–2158

  11. [19]

    Shuchang Liu, Qingpeng Cai, Zhankui He, Bowen Sun, Julian McAuley, Dong Zheng, Peng Jiang, and Kun Gai. 2023. Generative flow network for listwise rec- ommendation. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 1524–1534

  12. [20]

    Zhanyu Liu, Shiyao Wang, Xingmei Wang, Rongzhou Zhang, Jiaxin Deng, Honghui Bao, Jinghao Zhang, Wuchao Li, Pengfei Zheng, Xiangyu Wu, Yifei Hu, Qigen Hu, Xinchen Luo, Lejian Ren, Zixing Zhang, Qianqian Wang, Kuo Cai, Yunfan Wu, Hongtao Cheng, Zexuan Cheng, Lu Ren, Huanjie Wang...

  13. [21]

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

  14. [22]

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. 2018. Representation learning with contrastive predictive coding.arXiv preprint arXiv:1807.03748(2018)

  15. [23]

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. 2022. Training language models to follow instructions with human feedback.Advances in neural information processing systems35 (...

  16. [24]

    Junyan Qiu, Ze Wang, Fan Zhang, Zuowu Zheng, Jile Zhu, Jiangke Fan, Teng Zhang, Haitao Wang, and Xingxing Wang. 2025. One Model to Rank Them All: Unifying Online Advertising with End-to-End Learning.arXiv preprint arXiv:2505.19755(2025)

  17. [25]

    Yu Meng, Mengzhou Xia, and Danqi Chen. 2024. Simpo: Simple preference optimization with a reference-free reward.Advances in Neural Information Processing Systems37 (2024), 124198–124235

  18. [26]

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. 2023. Direct preference optimization: Your language model is secretly a reward model.Advances in neural information processing systems36 (2023), 53728–53741

  19. [27]

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

  20. [28]

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

  21. [29]

    Manning, and Chelsea Finn

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Stefano Ermon, Christopher D. Manning, and Chelsea Finn. 2024. Direct Preference Optimization: Your Language Model is Secretly a Reward Model. arXiv:2305.18290 [cs.LG] https://arxiv.org/ abs/2305.18290

  22. [30]

    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)

  23. [31]

    Wenjie Wang, Honghui Bao, Xinyu Lin, Jizhi Zhang, Yongqi Li, Fuli Feng, See- Kiong Ng, and Tat-Seng Chua. 2024. Learnable item tokenization for generative recommendation. InProceedings of the 33rd ACM International Conference on Information and Knowledge Management. 2400–2409

  24. [32]

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

  25. [33]

    Wanqi Xue, Qingpeng Cai, Zhenghai Xue, Shuo Sun, Shuchang Liu, Dong Zheng, Peng Jiang, Kun Gai, and Bo An. 2023. Prefrec: Recommender systems with human preferences for reinforcing long-term user engagement. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discover...

  26. [34]

    Jiaqi Zhai, Lucy Liao, Xing Liu, Yueming Wang, Rui Li, Xuan Cao, Leon Gao, Zhao- jie Gong, Fangda Gu, Michael He, et al. 2024. Actions speak louder than words: Trillion-parameter sequential transducers for generative recommendations.arXiv preprint arXiv:2402.17152(2024)

  27. [35]

    Federico Tomasi, Francesco Fabbri, Mounia Lalmas, and Zhenwen Dai. 2024. Diffusion Model for Slate Recommendation.arXiv preprint arXiv:2408.06883 (2024)

  28. [36]

    Junjie Zhang, Beichen Zhang, Wenqi Sun, Hongyu Lu, Wayne Xin Zhao, Yu Chen, and Ji-Rong Wen. 2025. Slow Thinking for Sequential Recommendation.arXiv preprint arXiv:2504.09627(2025)

  29. [37]

    Zhaoqi Zhang, Haolei Pei, Jun Guo, Tianyu Wang, Yufei Feng, Hui Sun, Shaowei Liu, and Aixin Sun. 2025. OneTrans: Unified Feature Interaction and Sequence Modeling with One Transformer in Industrial Recommender.arXiv preprint arXiv:2510.26104(2025)

  30. [38]

    Chuhan Wu, Fangzhao Wu, Tao Qi, Qi Liu, Xuan Tian, Jie Li, Wei He, Yongfeng Huang, and Xing Xie. 2022. Feedrec: News feed recommendation with various user feedbacks. InProceedings of the ACM Web Conference 2022. 2088–2097

  31. [39]

    Guorui Zhou, Hengrui Hu, Hongtao Cheng, Huanjie Wang, Jiaxin Deng, Jinghao Zhang, Kuo Cai, Lejian Ren, Lu Ren, Liao Yu, et al. 2025. Onerec-v2 technical report.arXiv preprint arXiv:2508.20900(2025). Conference acronym ’XX, June 03–05, 2018, Woodstock, NY Y. Pang et al. A Imple...

  32. [41]

    Jun Zhang, Yi Li, Yue Liu, Changping Wang, Yuan Wang, Yuling Xiong, Xun Liu, Haiyang Wu, Qian Li, Enming Zhang, et al. 2025. GPR: Towards a Generative Pre-trained One-Model Paradigm for Large-Scale Advertising Recommendation. arXiv preprint arXiv:2511.10138(2025)

  33. [44]

    Qi Zhao, Yi Zhang, Daniel Friedman, and Fangfang Tan. 2015. E-commerce recommendation with personalized promotion. InProceedings of the 9th ACM Conference on Recommender Systems. 219–226

  34. [2016]

    InProceedings of the 1st workshop on deep learning for recommender systems

    Wide & deep learning for recommender systems. InProceedings of the 1st workshop on deep learning for recommender systems. 7–10

  35. [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

  36. [2023]

    InPro- ceedings of the Sixteenth ACM International Conference on Web Search and Data Mining

    Generative slate recommendation with reinforcement learning. InPro- ceedings of the Sixteenth ACM International Conference on Web Search and Data Mining. 580–588

  37. [2024]

    arXiv:2402.03216 [cs.CL]

    BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation. arXiv:2402.03216 [cs.CL]

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.