Gradient-Boosted Decision Tree for Listwise Context Model in Multimodal Review Helpfulness Prediction

Anh Tuan Luu; Cong-Duy Nguyen; Lidong Bing; Thong Nguyen; Xiaobao Wu; Xinshuai Dong; Zhen Hai

arxiv: 2305.12678 · v3 · submitted 2023-05-22 · 💻 cs.CL

Gradient-Boosted Decision Tree for Listwise Context Model in Multimodal Review Helpfulness Prediction

Thong Nguyen , Xiaobao Wu , Xinshuai Dong , Anh Tuan Luu , Cong-Duy Nguyen , Zhen Hai , Lidong Bing This is my paper

Pith reviewed 2026-05-24 09:27 UTC · model grok-4.3

classification 💻 cs.CL

keywords multimodal review helpfulness predictionlistwise attention networkgradient-boosted decision treeranking contextlistwise optimizatione-commerce reviewsstate-of-the-art performance

0 comments

The pith

A listwise attention network with gradient-boosted decision trees ranks multimodal product reviews by helpfulness more accurately than prior neural methods.

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

The paper argues that fully-connected neural networks split review features inefficiently and that pairwise training objectives miss the full list-ranking goal in multimodal review helpfulness prediction. It introduces a listwise attention network to capture ranking context across all reviews, a listwise optimization objective for better generalization, and gradient-boosted decision trees to partition representations effectively. Experiments on two large-scale benchmarks show these changes deliver state-of-the-art results. A sympathetic reader would care because clearer helpful-review ranking could improve e-commerce customer experience by surfacing useful feedback first.

Core claim

The central discovery is that replacing fully-connected neural networks with gradient-boosted decision trees as the score predictor, while using a listwise attention network and listwise optimization objective instead of pairwise loss, produces state-of-the-art performance and improved generalization on multimodal review helpfulness prediction tasks.

What carries the argument

The listwise attention network that captures ranking context across the entire review list, combined with gradient-boosted decision trees that partition review representations.

If this is right

Review features are partitioned more effectively, allowing clearer separation of helpful from unhelpful reviews.
Model generalization improves during testing because the objective matches the full list ranking goal.
State-of-the-art results are achieved on two large-scale MRHP benchmark datasets.
The approach applies to e-commerce by presenting customers with more useful reviews.

Where Pith is reading between the lines

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

Listwise methods may generalize to other ranking tasks in multimodal settings beyond reviews.
Gradient-boosted trees could replace neural predictors in similar feature-heavy ranking problems where splitting efficiency matters.
Testing on additional datasets with varying review list sizes would further validate the listwise objective's advantage.

Load-bearing premise

That fully-connected neural networks perform inefficient splitting for review features and that pairwise objectives fail to capture the full list ranking goal.

What would settle it

An experiment showing that a fully-connected network with pairwise loss matches or exceeds the proposed method on the same benchmarks would falsify the claim.

Figures

Figures reproduced from arXiv: 2305.12678 by Anh Tuan Luu, Cong-Duy Nguyen, Lidong Bing, Thong Nguyen, Xiaobao Wu, Xinshuai Dong, Zhen Hai.

**Figure 2.** Figure 2: Illustration of our Multimodal Review Helpfulness Prediction model. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 4.** Figure 4: Generalization error curves per training epoch [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

**Figure 3.** Figure 3: Generalization error curves per training epoch [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

**Figure 5.** Figure 5: Generalization error curves per training epoch on the Clothing category in Amazon-MRHP and Lazada [PITH_FULL_IMAGE:figures/full_fig_p016_5.png] view at source ↗

**Figure 6.** Figure 6: Generalization error curves per training epoch on the Home category in Amazon-MRHP and Lazada [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗

**Figure 7.** Figure 7: Mean µi routing probabilities at the proposed GBDT’s leaves for 1-rating and 2-rating reviews in AmazonHome dataset. 0 5 10 15 20 25 30 Leaf nodes 0.00 0.02 0.04 0.06 0.08 µi Mean µi values on 3-rating reviews of Amazon-Home dataset 0 5 10 15 20 25 30 Leaf nodes 0.00 0.02 0.04 0.06 0.08 0.10 µi Mean µi values on 4-rating reviews of Amazon-Home dataset [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗

**Figure 8.** Figure 8: Mean µi routing probabilities at the proposed GBDT’s leaves for 3-rating and 4-rating reviews in AmazonHome dataset [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗

**Figure 9.** Figure 9: Mean µi routing probabilities at the proposed GBDT’s leaves for 0-rating and 1-rating reviews in LazadaClothing dataset. 1 2 3 4 5 6 7 8 Leaf nodes 0.00 0.05 0.10 0.15 0.20 0.25 0.30 µi Mean µi values on 2-rating reviews of Lazada-Clothing dataset 1 2 3 4 5 6 7 8 Leaf nodes 0.00 0.05 0.10 0.15 0.20 0.25 µi Mean µi values on 3-rating reviews of Lazada-Clothing dataset [PITH_FULL_IMAGE:figures/full_fig_p01… view at source ↗

**Figure 10.** Figure 10: Mean µi routing probabilities at the proposed GBDT’s leaves for 2-rating and 3-rating reviews in Lazada-Clothing dataset. 1 2 3 4 5 6 7 8 Leaf nodes 0.00 0.05 0.10 0.15 0.20 0.25 µi Mean µi values on 4-rating reviews of Lazada-Clothing dataset [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗

**Figure 11.** Figure 11: Mean µi routing probabilities at the proposed GBDT’s leaves for 4-rating reviews in Lazada-Clothing dataset. E Examples of Product and Review Samples We articulate product and review samples in [PITH_FULL_IMAGE:figures/full_fig_p018_11.png] view at source ↗

read the original abstract

Multimodal Review Helpfulness Prediction (MRHP) aims to rank product reviews based on predicted helpfulness scores and has been widely applied in e-commerce via presenting customers with useful reviews. Previous studies commonly employ fully-connected neural networks (FCNNs) as the final score predictor and pairwise loss as the training objective. However, FCNNs have been shown to perform inefficient splitting for review features, making the model difficult to clearly differentiate helpful from unhelpful reviews. Furthermore, pairwise objective, which works on review pairs, may not completely capture the MRHP goal to produce the ranking for the entire review list, and possibly induces low generalization during testing. To address these issues, we propose a listwise attention network that clearly captures the MRHP ranking context and a listwise optimization objective that enhances model generalization. We further propose gradient-boosted decision tree as the score predictor to efficaciously partition product reviews' representations. Extensive experiments demonstrate that our method achieves state-of-the-art results and polished generalization performance on two large-scale MRHP benchmark datasets.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper targets FCNN and pairwise limitations in MRHP with listwise attention plus GBDT, but the SOTA claim has zero supporting numbers or details visible.

read the letter

The main thing to know is that the authors identify two concrete problems with prior MRHP work—FCNNs making poor feature splits and pairwise losses failing to optimize full-list ranking—and respond with a listwise attention network, a listwise objective, and GBDT as the final predictor. That combination is the actual new element; each piece exists elsewhere, but the targeted application to multimodal review ranking is not already in the cited literature they reference. The motivation section is straightforward and the proposed fixes follow logically from the stated shortcomings. Credit is due for framing the task as listwise from the start rather than bolting on listwise components later. The soft spot is obvious and central: the abstract asserts extensive experiments, SOTA results, and polished generalization on two large benchmark datasets, yet supplies no metrics, no baseline descriptions, no ablation tables, and no error bars. Without those, the performance claims cannot be evaluated at all. The assumptions about FCNN inefficiency and pairwise incompleteness are plausible but remain untested in the provided text. This work is aimed at researchers building review-ranking systems for e-commerce platforms. A reader already working on multimodal recommendation might extract the architectural idea and test it themselves, but the paper as presented offers no evidence that the changes produce measurable gains. I would not send it to peer review until the experimental section is present and the quantitative results can be inspected; if those turn out solid, the targeted fix could be worth a look in the subfield.

Referee Report

2 major / 0 minor

Summary. The paper proposes replacing FCNN predictors and pairwise losses in multimodal review helpfulness prediction (MRHP) with a listwise attention network, a listwise optimization objective, and a gradient-boosted decision tree (GBDT) score predictor. It claims these changes address inefficient feature splitting and incomplete list ranking, yielding state-of-the-art results and improved generalization on two large-scale MRHP benchmark datasets.

Significance. If the empirical claims are substantiated by properly documented experiments, the approach could offer a practical alternative for listwise ranking tasks in e-commerce by combining attention-based context modeling with tree-based partitioning, potentially improving both accuracy and generalization over standard neural baselines.

major comments (2)

Abstract: the central claim that 'extensive experiments demonstrate that our method achieves state-of-the-art results and polished generalization performance' is unsupported by any quantitative results, baseline descriptions, metrics, ablation tables, dataset statistics, or error bars in the supplied manuscript text, rendering the primary empirical assertion unverifiable.
Motivation paragraph: the assertions that 'FCNNs have been shown to perform inefficient splitting for review features' and that 'pairwise objective ... may not completely capture the MRHP goal' are presented without accompanying citations, prior empirical evidence, or quantitative motivation within the provided text, leaving the load-bearing motivation ungrounded.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their review and comments. We address each major comment below and commit to revisions that strengthen the manuscript's clarity and verifiability.

read point-by-point responses

Referee: Abstract: the central claim that 'extensive experiments demonstrate that our method achieves state-of-the-art results and polished generalization performance' is unsupported by any quantitative results, baseline descriptions, metrics, ablation tables, dataset statistics, or error bars in the supplied manuscript text, rendering the primary empirical assertion unverifiable.

Authors: The full manuscript contains a dedicated Experiments section (Section 4) with quantitative results, baseline comparisons (including prior FCNN and pairwise methods), standard metrics, ablation tables, dataset statistics for the two MRHP benchmarks, and error bars from multiple runs. However, we acknowledge that the abstract itself does not embed these specifics. We will revise the abstract to include key quantitative highlights (e.g., relative improvements and dataset names) so the central claim becomes directly verifiable from the abstract text. revision: yes
Referee: Motivation paragraph: the assertions that 'FCNNs have been shown to perform inefficient splitting for review features' and that 'pairwise objective ... may not completely capture the MRHP goal' are presented without accompanying citations, prior empirical evidence, or quantitative motivation within the provided text, leaving the load-bearing motivation ungrounded.

Authors: We agree these statements require explicit grounding. The claims draw from established literature on decision-tree advantages for feature partitioning in tabular/review data and listwise ranking objectives outperforming pairwise ones for full-list evaluation. We will add relevant citations to prior work on these topics and include a short quantitative motivation (e.g., referencing performance gaps reported in related ranking papers) to make the motivation self-contained and evidence-based. revision: yes

Circularity Check

0 steps flagged

Empirical model proposal with no derivation chain or self-referential predictions

full rationale

The paper proposes architectural components (listwise attention network, listwise optimization objective, GBDT score predictor) motivated by stated limitations of FCNNs and pairwise losses, then asserts SOTA performance via experiments on benchmark datasets. No equations, derivations, or first-principles results appear in the provided text. The central claim is an empirical outcome rather than a mathematical reduction; no fitted parameters are renamed as predictions, no self-citations bear load for uniqueness theorems, and no ansatz or renaming of known results occurs. The derivation chain is therefore self-contained as a standard empirical ML contribution.

Axiom & Free-Parameter Ledger

1 free parameters · 0 axioms · 0 invented entities

The proposal rests on standard machine-learning assumptions that attention can capture list context and that GBDT can partition representations more effectively than FCNNs; no new physical or mathematical axioms are introduced.

free parameters (1)

GBDT and neural-network hyperparameters
Typical ML training requires tuning of learning rates, tree depth, number of trees, attention dimensions, and loss weights; these are fitted or chosen on validation data.

pith-pipeline@v0.9.0 · 5733 in / 1178 out tokens · 25362 ms · 2026-05-24T09:27:08.692187+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We propose a listwise attention network … and a listwise optimization objective … gradient-boosted decision tree as the score predictor … Llist = −∑ y′ij log(f′ij)
IndisputableMonolith/Foundation/ArithmeticFromLogic.lean embed_injective unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Theorem 4 … E(f listD) ≤ E(f pairD)

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

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

DemaFormer: Damped Exponential Moving Average Transformer with Energy-Based Modeling for Temporal Language Grounding
cs.CV 2023-12 unverdicted novelty 4.0

DemaFormer pairs energy-based modeling with a damped-EMA Transformer to localize video moments matching language queries and reports gains over baselines on four datasets.

Reference graph

Works this paper leans on

36 extracted references · 36 canonical work pages · cited by 1 Pith paper · 3 internal anchors

[1]

Mahdi Abavisani, Liwei Wu, Shengli Hu, Joel Tetreault, and Alejandro Jaimes. 2020. Multimodal categorization of crisis events in social media. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14679--14689

work page 2020
[2]

Ali Akbari, Muhammad Awais, Manijeh Bashar, and Josef Kittler. 2021. How does loss function affect generalization performance of deep learning? application to human age estimation. In International Conference on Machine Learning, pages 141--151. PMLR

work page 2021
[3]

Alex Beutel, Paul Covington, Sagar Jain, Can Xu, Jia Li, Vince Gatto, and Ed H Chi. 2018. Latent cross: Making use of context in recurrent recommender systems. In Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining, pages 46--54

work page 2018
[4]

Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2017. Enriching word vectors with subword information. Transactions of the association for computational linguistics, 5:135--146

work page 2017
[5]

Cen Chen, Yinfei Yang, Jun Zhou, Xiaolong Li, and Forrest Bao. 2018. Cross-domain review helpfulness prediction based on convolutional neural networks with auxiliary domain discriminators. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Paper...

work page 2018
[6]

Zhuyun Dai, Chenyan Xiong, Jamie Callan, and Zhiyuan Liu. 2018. Convolutional neural networks for soft-matching n-grams in ad-hoc search. In Proceedings of the eleventh ACM international conference on web search and data mining, pages 126--134

work page 2018
[7]

Miao Fan, Chao Feng, Lin Guo, Mingming Sun, and Ping Li. 2019. Product-aware helpfulness prediction of online reviews. In The World Wide Web Conference, pages 2715--2721

work page 2019
[8]

Wei Han, Hui Chen, Zhen Hai, Soujanya Poria, and Lidong Bing. 2022. Sancl: Multimodal review helpfulness prediction with selective attention and natural contrastive learning. In Proceedings of the 29th International Conference on Computational Linguistics, pages 5666--5677

work page 2022
[9]

Soo-Min Kim, Patrick Pantel, Timothy Chklovski, and Marco Pennacchiotti. 2006. Automatically assessing review helpfulness. In Proceedings of the 2006 Conference on empirical methods in natural language processing, pages 423--430

work page 2006
[10]

Srikumar Krishnamoorthy. 2015. Linguistic features for review helpfulness prediction. Expert Systems with Applications, 42(7):3751--3759

work page 2015
[11]

Jean-Samuel Leboeuf, Fr \'e d \'e ric LeBlanc, and Mario Marchand. 2020. Decision trees as partitioning machines to characterize their generalization properties. Advances in Neural Information Processing Systems, 33:18135--18145

work page 2020
[12]

Haijing Liu, Yang Gao, Pin Lv, Mengxue Li, Shiqiang Geng, Minglan Li, and Hao Wang. 2017. Using argument-based features to predict and analyse review helpfulness. arXiv preprint arXiv:1707.07279

work page internal anchor Pith review Pith/arXiv arXiv 2017
[13]

Junhao Liu, Zhen Hai, Min Yang, and Lidong Bing. 2021. Multi-perspective coherent reasoning for helpfulness prediction of multimodal reviews. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 5927--5936

work page 2021
[14]

Anh Tuan Luu, Jung-jae Kim, and See Kiong Ng. 2015. Incorporating trustiness and collective synonym/contrastive evidence into taxonomy construction. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1013--1022

work page 2015
[15]

Anh Tuan Luu, Yi Tay, Siu Cheung Hui, and See Kiong Ng. 2016. Learning term embeddings for taxonomic relation identification using dynamic weighting neural network. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 403--413

work page 2016
[16]

Jiaqi Ma, Xinyang Yi, Weijing Tang, Zhe Zhao, Lichan Hong, Ed Chi, and Qiaozhu Mei. 2021. Learning-to-rank with partitioned preference: fast estimation for the plackett-luce model. In International Conference on Artificial Intelligence and Statistics, pages 928--936. PMLR

work page 2021
[17]

Thong Nguyen and Anh Tuan Luu. 2021. Contrastive learning for neural topic model. Advances in Neural Information Processing Systems, 34:11974--11986

work page 2021
[18]

Thong Nguyen, Anh Tuan Luu, Truc Lu, and Tho Quan. 2021. Enriching and controlling global semantics for text summarization. arXiv preprint arXiv:2109.10616

work page internal anchor Pith review Pith/arXiv arXiv 2021
[19]

Thong Nguyen, Xiaobao Wu, Anh-Tuan Luu, Cong-Duy Nguyen, Zhen Hai, and Lidong Bing. 2022. Adaptive contrastive learning on multimodal transformer for review helpfulness predictions. arXiv preprint arXiv:2211.03524

work page internal anchor Pith review Pith/arXiv arXiv 2022
[20]

Thong Thanh Nguyen and Anh Tuan Luu. 2022. Improving neural cross-lingual abstractive summarization via employing optimal transport distance for knowledge distillation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 11103--11111

work page 2022
[21]

Rama Kumar Pasumarthi, Xuanhui Wang, Michael Bendersky, and Marc Najork. 2019. Self-attentive document interaction networks for permutation equivariant ranking. arXiv preprint arXiv:1910.09676

work page arXiv 2019
[22]

Jeffrey Pennington, Richard Socher, and Christopher D Manning. 2014. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pages 1532--1543

work page 2014
[23]

Przemys aw Pobrotyn and Rados aw Bia obrzeski. 2021. Neuralndcg: Direct optimisation of a ranking metric via differentiable relaxation of sorting. arXiv preprint arXiv:2102.07831

work page arXiv 2021
[24]

Zhen Qin, Le Yan, Honglei Zhuang, Yi Tay, Rama Kumar Pasumarthi, Xuanhui Wang, Mike Bendersky, and Marc Najork. 2021. Are neural rankers still outperformed by gradient boosted decision trees?

work page 2021
[25]

Xiaoru Qu, Zhao Li, Jialin Wang, Zhipeng Zhang, Pengcheng Zou, Junxiao Jiang, Jiaming Huang, Rong Xiao, Ji Zhang, and Jun Gao. 2020. Category-aware graph neural networks for improving e-commerce review helpfulness prediction. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management, pages 2693--2700

work page 2020
[26]

Luu Anh Tuan, Siu Cheung Hui, and See Kiong Ng. 2016. Utilizing temporal information for taxonomy construction. Transactions of the Association for Computational Linguistics, 4:551--564

work page 2016
[27]

Zhiguo Wang, Wael Hamza, and Radu Florian. 2017. Bilateral multi-perspective matching for natural language sentences. In Proceedings of the 26th International Joint Conference on Artificial Intelligence, pages 4144--4150

work page 2017
[28]

Xiaobao Wu, Xinshuai Dong, Thong Nguyen, Chaoqun Liu, Liangming Pan, and Anh Tuan Luu. 2023 a . Infoctm: A mutual information maximization perspective of cross-lingual topic modeling. arXiv preprint arXiv:2304.03544

work page arXiv 2023
[29]

Xiaobao Wu, Xinshuai Dong, Thong Thanh Nguyen, and Anh Tuan Luu. 2023 b . Effective neural topic modeling with embedding clustering regularization. In International Conference on Machine Learning. PMLR

work page 2023
[30]

Xiaobao Wu, Chunping Li, Yan Zhu, and Yishu Miao. 2020. Short text topic modeling with topic distribution quantization and negative sampling decoder. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1772--1782, Online. Association for Computational Linguistics

work page 2020
[31]

Xiaobao Wu, Anh Tuan Luu, and Xinshuai Dong. 2022. Mitigating data sparsity for short text topic modeling by topic-semantic contrastive learning. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 2748--2760, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics

work page 2022
[32]

Fen Xia, Tie-Yan Liu, Jue Wang, Wensheng Zhang, and Hang Li. 2008. Listwise approach to learning to rank: theory and algorithm. In Proceedings of the 25th international conference on Machine learning, pages 1192--1199

work page 2008
[33]

Nan Xu, Zhixiong Zeng, and Wenji Mao. 2020. Reasoning with multimodal sarcastic tweets via modeling cross-modality contrast and semantic association. In Proceedings of the 58th annual meeting of the association for computational linguistics, pages 3777--3786

work page 2020
[34]

Yinfei Yang, Yaowei Yan, Minghui Qiu, and Forrest Bao. 2015. Semantic analysis and helpfulness prediction of text for online product reviews. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2: Short Papers), pages 38--44

work page 2015
[35]

URL: " 'urlintro :=

ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year eprint doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRINGS urlintro eprinturl eprintpr...

work page
[36]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

work page

[1] [1]

Mahdi Abavisani, Liwei Wu, Shengli Hu, Joel Tetreault, and Alejandro Jaimes. 2020. Multimodal categorization of crisis events in social media. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14679--14689

work page 2020

[2] [2]

Ali Akbari, Muhammad Awais, Manijeh Bashar, and Josef Kittler. 2021. How does loss function affect generalization performance of deep learning? application to human age estimation. In International Conference on Machine Learning, pages 141--151. PMLR

work page 2021

[3] [3]

Alex Beutel, Paul Covington, Sagar Jain, Can Xu, Jia Li, Vince Gatto, and Ed H Chi. 2018. Latent cross: Making use of context in recurrent recommender systems. In Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining, pages 46--54

work page 2018

[4] [4]

Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2017. Enriching word vectors with subword information. Transactions of the association for computational linguistics, 5:135--146

work page 2017

[5] [5]

Cen Chen, Yinfei Yang, Jun Zhou, Xiaolong Li, and Forrest Bao. 2018. Cross-domain review helpfulness prediction based on convolutional neural networks with auxiliary domain discriminators. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Paper...

work page 2018

[6] [6]

Zhuyun Dai, Chenyan Xiong, Jamie Callan, and Zhiyuan Liu. 2018. Convolutional neural networks for soft-matching n-grams in ad-hoc search. In Proceedings of the eleventh ACM international conference on web search and data mining, pages 126--134

work page 2018

[7] [7]

Miao Fan, Chao Feng, Lin Guo, Mingming Sun, and Ping Li. 2019. Product-aware helpfulness prediction of online reviews. In The World Wide Web Conference, pages 2715--2721

work page 2019

[8] [8]

Wei Han, Hui Chen, Zhen Hai, Soujanya Poria, and Lidong Bing. 2022. Sancl: Multimodal review helpfulness prediction with selective attention and natural contrastive learning. In Proceedings of the 29th International Conference on Computational Linguistics, pages 5666--5677

work page 2022

[9] [9]

Soo-Min Kim, Patrick Pantel, Timothy Chklovski, and Marco Pennacchiotti. 2006. Automatically assessing review helpfulness. In Proceedings of the 2006 Conference on empirical methods in natural language processing, pages 423--430

work page 2006

[10] [10]

Srikumar Krishnamoorthy. 2015. Linguistic features for review helpfulness prediction. Expert Systems with Applications, 42(7):3751--3759

work page 2015

[11] [11]

Jean-Samuel Leboeuf, Fr \'e d \'e ric LeBlanc, and Mario Marchand. 2020. Decision trees as partitioning machines to characterize their generalization properties. Advances in Neural Information Processing Systems, 33:18135--18145

work page 2020

[12] [12]

Haijing Liu, Yang Gao, Pin Lv, Mengxue Li, Shiqiang Geng, Minglan Li, and Hao Wang. 2017. Using argument-based features to predict and analyse review helpfulness. arXiv preprint arXiv:1707.07279

work page internal anchor Pith review Pith/arXiv arXiv 2017

[13] [13]

Junhao Liu, Zhen Hai, Min Yang, and Lidong Bing. 2021. Multi-perspective coherent reasoning for helpfulness prediction of multimodal reviews. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 5927--5936

work page 2021

[14] [14]

Anh Tuan Luu, Jung-jae Kim, and See Kiong Ng. 2015. Incorporating trustiness and collective synonym/contrastive evidence into taxonomy construction. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1013--1022

work page 2015

[15] [15]

Anh Tuan Luu, Yi Tay, Siu Cheung Hui, and See Kiong Ng. 2016. Learning term embeddings for taxonomic relation identification using dynamic weighting neural network. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 403--413

work page 2016

[16] [16]

Jiaqi Ma, Xinyang Yi, Weijing Tang, Zhe Zhao, Lichan Hong, Ed Chi, and Qiaozhu Mei. 2021. Learning-to-rank with partitioned preference: fast estimation for the plackett-luce model. In International Conference on Artificial Intelligence and Statistics, pages 928--936. PMLR

work page 2021

[17] [17]

Thong Nguyen and Anh Tuan Luu. 2021. Contrastive learning for neural topic model. Advances in Neural Information Processing Systems, 34:11974--11986

work page 2021

[18] [18]

Thong Nguyen, Anh Tuan Luu, Truc Lu, and Tho Quan. 2021. Enriching and controlling global semantics for text summarization. arXiv preprint arXiv:2109.10616

work page internal anchor Pith review Pith/arXiv arXiv 2021

[19] [19]

Thong Nguyen, Xiaobao Wu, Anh-Tuan Luu, Cong-Duy Nguyen, Zhen Hai, and Lidong Bing. 2022. Adaptive contrastive learning on multimodal transformer for review helpfulness predictions. arXiv preprint arXiv:2211.03524

work page internal anchor Pith review Pith/arXiv arXiv 2022

[20] [20]

Thong Thanh Nguyen and Anh Tuan Luu. 2022. Improving neural cross-lingual abstractive summarization via employing optimal transport distance for knowledge distillation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 11103--11111

work page 2022

[21] [21]

Rama Kumar Pasumarthi, Xuanhui Wang, Michael Bendersky, and Marc Najork. 2019. Self-attentive document interaction networks for permutation equivariant ranking. arXiv preprint arXiv:1910.09676

work page arXiv 2019

[22] [22]

Jeffrey Pennington, Richard Socher, and Christopher D Manning. 2014. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pages 1532--1543

work page 2014

[23] [23]

Przemys aw Pobrotyn and Rados aw Bia obrzeski. 2021. Neuralndcg: Direct optimisation of a ranking metric via differentiable relaxation of sorting. arXiv preprint arXiv:2102.07831

work page arXiv 2021

[24] [24]

Zhen Qin, Le Yan, Honglei Zhuang, Yi Tay, Rama Kumar Pasumarthi, Xuanhui Wang, Mike Bendersky, and Marc Najork. 2021. Are neural rankers still outperformed by gradient boosted decision trees?

work page 2021

[25] [25]

Xiaoru Qu, Zhao Li, Jialin Wang, Zhipeng Zhang, Pengcheng Zou, Junxiao Jiang, Jiaming Huang, Rong Xiao, Ji Zhang, and Jun Gao. 2020. Category-aware graph neural networks for improving e-commerce review helpfulness prediction. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management, pages 2693--2700

work page 2020

[26] [26]

Luu Anh Tuan, Siu Cheung Hui, and See Kiong Ng. 2016. Utilizing temporal information for taxonomy construction. Transactions of the Association for Computational Linguistics, 4:551--564

work page 2016

[27] [27]

Zhiguo Wang, Wael Hamza, and Radu Florian. 2017. Bilateral multi-perspective matching for natural language sentences. In Proceedings of the 26th International Joint Conference on Artificial Intelligence, pages 4144--4150

work page 2017

[28] [28]

Xiaobao Wu, Xinshuai Dong, Thong Nguyen, Chaoqun Liu, Liangming Pan, and Anh Tuan Luu. 2023 a . Infoctm: A mutual information maximization perspective of cross-lingual topic modeling. arXiv preprint arXiv:2304.03544

work page arXiv 2023

[29] [29]

Xiaobao Wu, Xinshuai Dong, Thong Thanh Nguyen, and Anh Tuan Luu. 2023 b . Effective neural topic modeling with embedding clustering regularization. In International Conference on Machine Learning. PMLR

work page 2023

[30] [30]

Xiaobao Wu, Chunping Li, Yan Zhu, and Yishu Miao. 2020. Short text topic modeling with topic distribution quantization and negative sampling decoder. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1772--1782, Online. Association for Computational Linguistics

work page 2020

[31] [31]

Xiaobao Wu, Anh Tuan Luu, and Xinshuai Dong. 2022. Mitigating data sparsity for short text topic modeling by topic-semantic contrastive learning. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 2748--2760, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics

work page 2022

[32] [32]

Fen Xia, Tie-Yan Liu, Jue Wang, Wensheng Zhang, and Hang Li. 2008. Listwise approach to learning to rank: theory and algorithm. In Proceedings of the 25th international conference on Machine learning, pages 1192--1199

work page 2008

[33] [33]

Nan Xu, Zhixiong Zeng, and Wenji Mao. 2020. Reasoning with multimodal sarcastic tweets via modeling cross-modality contrast and semantic association. In Proceedings of the 58th annual meeting of the association for computational linguistics, pages 3777--3786

work page 2020

[34] [34]

Yinfei Yang, Yaowei Yan, Minghui Qiu, and Forrest Bao. 2015. Semantic analysis and helpfulness prediction of text for online product reviews. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2: Short Papers), pages 38--44

work page 2015

[35] [35]

URL: " 'urlintro :=

ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year eprint doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRINGS urlintro eprinturl eprintpr...

work page

[36] [36]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

work page