Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

The paper argues that information cocoons in news recommendation can be measured by five indicators spanning individual and group levels, that the cocoon deepens over repeated rounds for most recommenders, and that lightweight interventions

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

2026-08-04 17:01 UTC pith:ZZINU36U

load-bearing objection A useful metric framework for measuring information cocoons, but the central multi-round finding is unverifiable as written because the simulation loop is never specified. the 4 major comments →

arxiv 2509.11139 v1 pith:ZZINU36U submitted 2025-09-14 cs.IR

Understanding the Information Cocoon: A Multidimensional Assessment and Analysis of News Recommendation Systems

Xin Wang , Xiaowen Huang , Jitao Sang This is my paper
classification cs.IR
keywords information cocoonnews recommendationfilter bubbleecho chambertopic diversitygroup polarizationmulti-round evaluationrecommendation mitigation
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.

This paper tries to establish a single, multidimensional way to measure the 'information cocoon' effect in news recommender systems, and to use it to compare recommendation algorithms over multiple rounds of exposure. It argues that the cocoon is not just a matter of individual topic narrowing but also of group-level polarization, captured by five indicators: number of topic categories, category entropy, click repeat rate, network density, and community openness. Applying these metrics to seven recommenders on two real-world news datasets, the paper contends that the cocoon deepens progressively across recommendation rounds and that most models intensify it, with knowledge-graph-aware recommenders showing the mildest effect. It then shows five lightweight mitigation strategies—random exploration, diversity regularization, attention balancing, and two community-aware re-rankers—that improve diversity while keeping accuracy loss near two percent. A sympathetic reader would care because the framework could become a standard evaluation lens for ethical news recommendation.

Core claim

The central claim is that the proposed multidimensional framework—individual homogenization measured by number of topic categories N, category entropy H, and click repeat rate R, and group polarization measured by network density D and community openness O—captures the information cocoon effect in news recommendation. Using multi-round experiments on two real-world news datasets, the paper finds that the cocoon effect deepens progressively: N and H decrease, R and D increase, and O decreases. Most of the seven recommenders exacerbate the effect; the recommenders that integrate external knowledge retain the most diversity. The paper further claims that five mitigation strategies (EGS, CDR, LT

What carries the argument

The central machinery is the five-indicator assessment set (N, H, R, D, O) applied over multi-round recommendation loops. N and H are the average number of distinct topic categories and the category entropy of Top-K lists; R is the proportion of simulated clicks whose categories repeat the user's history; D and O are computed from a user-item bipartite graph with community detection, measuring average internal edge density and the balance of external versus internal edges. This set translates the abstract notion of an information cocoon into measurable quantities at two levels and exposes how each quantity evolves as recommendations are repeated.

Load-bearing premise

The load-bearing premise is that the multi-round simulated clicking faithfully reproduces real user behavior; the paper never specifies how clicks are simulated, how history updates, or how many rounds are run, so if the simulation is unrepresentative, the reported deepening of the cocoon could be an artifact.

What would settle it

A concrete check: re-run the same recommenders under an explicitly specified user-simulation protocol—state how clicks are chosen, how history updates, and how many rounds run—and test whether N and H still fall while R and D rise when clicks are not simply the recommender's top-ranked items. If the trends disappear, the claimed progressive deepening is an artifact of the simulation.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

  • A single multi-indicator report, rather than any one metric, becomes the natural unit for comparing recommenders on cocoon risk.
  • Most current news recommenders would be expected to reduce topical breadth and increase network closure as they are deployed over time; knowledge-graph-aware models would be the most resilient.
  • The first roughly ten rounds are where the cocoon forms fastest, so early evaluation is the most informative.
  • The five mitigation strategies offer a practical toolkit with roughly two percent accuracy cost; community-aware re-ranking gives the largest gains in openness.

Where Pith is reading between the lines

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

  • Inference: The same five indicators could be computed for short-video and social feeds, where topic labels and interaction graphs exist, effectively exporting the framework beyond news.
  • Inference: Because the click simulation is under-specified, a validation study using logged impressions from a deployed recommender—rather than simulated clicks—would test whether the deepening trend holds outside the lab.
  • Inference: The small accuracy cost of the mitigation strategies suggests diversity can be tuned as a knob; combining the strategies with fairness metrics could reveal whether broader exposure also improves source and publisher equity.

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. The paper proposes a multidimensional framework for assessing the information cocoon effect in news recommendation. It defines five metrics: number of topic categories (N), category information entropy (H), click repeat rate (R) at the individual level, and network density (D) and community openness (O) at the group level. Using the MIND and Adressa datasets, the authors report multi-round experiments with seven classic news recommenders, claiming that the cocoon effect deepens progressively over rounds and that most models exacerbate it. They also propose five mitigation strategies and report their effects on the proposed metrics.

Significance. If the empirical claims are reproducible, the proposed framework would provide a useful unification of individual-level and group-level cocoon indicators and a benchmark for comparing recommender systems on diversity and polarization. The metric definitions (Eqs. 1–5) are explicit and self-consistent, and the use of two real-world datasets is a strength. However, the central empirical contribution is currently unverifiable: the multi-round simulation loop is never specified, the click-repeat-rate metric is partly circular when history is updated from the recommender's own outputs, and no statistical evidence (error bars, seeds, significance tests) is provided for any comparison. The paper also ships no code or detailed protocol, so the headline conclusions cannot be checked.

major comments (4)
  1. [§3.1.3 and §4.2] The central empirical claim—'the information cocoon effect deepens progressively' (Appendix B) and 'most news recommendation models exacerbate the cocoon' (§4.5)—rests on a multi-round simulation that is never specified. The paper does not state the number of rounds, how candidate pools are regenerated, how a user chooses which items to click, how history is updated between rounds, or whether models are retrained. Under the natural protocol (the user clicks top-scored items and the history is extended with those clicks), R and D are partly constructed by the feedback loop. Please provide the full protocol; without it, Figures 5–8 and Tables 2–6 are unreproducible.
  2. [Eq. (3), §3.1.3] The click repeat rate R is defined as the proportion of clicked items whose category is in h_j, the user's historical click categories. If h_j is updated with the user's own clicks each round, any recommender that exploits historical behavior will mechanically increase R, independent of whether a genuine cocoon forms. The reported R values are 0.93–0.99, and these values are used to conclude that recommendation models deepen the cocoon. A non-personalized baseline (e.g., random or popularity-based recommendation) and a precise definition of h_j (initial history only, or cumulative over rounds) are needed to separate metric-induced drift from algorithmic behavior.
  3. [Tables 2–6 and Figures 5–8] The experimental results are reported as single point values with no error bars, standard deviations, number of seeds, or significance tests. Consequently, differences between models, and the claimed 'most models exacerbate' finding, cannot be assessed statistically. Table 4 in particular reports mitigation improvements such as CDR on Adressa D (-15.55%) and LTAO on Adressa O (-19.61%) that are presented as effects but have no variance estimates. Please include multiple runs, error bars, and significance tests for both the benchmark comparisons and the mitigation evaluation.
  4. [Appendix B, Figure 8] The text derives specific temporal conclusions—'dramatic changes occur in the early rounds, especially the first 10' and 'stabilizes around rounds 10–15'—from Figure 8. However, Figure 8 is described as a histogram of a scatter distribution with the x-axis showing category results and the y-axis showing subcategory results; round number is not visibly represented as a temporal axis. Please clarify how the round dimension is encoded in the figure and how the stated early/late-round dynamics follow from the displayed data.
minor comments (6)
  1. [References] Reference [23] is duplicated as [22] in the Introduction; the citation list should be cleaned.
  2. [Eq. (6)] The piecewise expression for the epsilon-greedy strategy is syntactically unclear: 'softmax(s_ui),1-epsilon' should be written as a proper two-branch definition with probabilities summing to one.
  3. [§4.6] The mitigation strategies introduce hyperparameters epsilon, lambda, mu, gamma, and alpha, but their values are never reported. Without these, the mitigation results in Table 4 are not reproducible.
  4. [§4.2, Table 4] The text states that mitigation causes only a 'small fluctuation (2%) in performance metrics including AUC, MRR, NDCG@5, and NDCG@10,' but these accuracy metrics are not reported in any table. Please include them.
  5. [Appendix A.2, Table 5] The table notes say bold and underline indicate the strongest and second-strongest cocoon effects, but these formatting cues are not visible in the rendered table. Please ensure they are displayed or describe the values explicitly.
  6. [§4.1.3] Louvain community detection is applied to a user-item bipartite network, but standard Louvain assumes a unipartite graph. Please state whether a bipartite modularity variant or a projection was used.

Circularity Check

0 steps flagged

No circular derivation: the metrics are operationalizations, and the benchmark conclusion does not reduce to the equations by construction.

full rationale

The paper's central claims are empirical benchmark results, not derived from the metric definitions alone. Equations (1)-(5) define N, H, R, D, and O from recommendation lists, simulated clicks, and the user-item network; no parameter is fitted to the target conclusion and no prior result by these authors is invoked as the justification for the cocoon-deepening finding. The click repeat rate is operationalized as the fraction of simulated clicks whose category already appears in the user's history, which is a standard measurement choice; while a personalized recommender trained on history will tend to produce high R values, the paper's cross-model ranking and the trajectories of N, H, D, and O are not algebraic consequences of Eq. (3) alone. The most serious weakness is that the multi-round click simulation is never specified (Section 3.1.3 simply says 'we simulate users' click behavior' and Section 4.2 refers to 'multiple rounds' without stating the number of rounds, candidate regeneration, click selection, or retraining). This makes the experiments non-reproducible and the magnitude of the R trends difficult to interpret, but it is a correctness/reproducibility concern rather than circularity: there is no equation that reduces to another by construction, and no fitted parameter is renamed as a prediction. Accordingly, no circular step is identified.

Axiom & Free-Parameter Ledger

6 free parameters · 4 axioms · 0 invented entities

The framework introduces no new entities or forces. Its free parameters are mostly mitigation hyperparameters that are not reported, plus the unspecified number of simulation rounds. The key domain assumptions concern the click simulation, category reliability, and the interpretation of Louvain communities as polarization groups.

free parameters (6)
  • epsilon (EGS) = not reported
    Exploration probability in Eq. 6; likely chosen by hand to maximize mitigation gains; no sensitivity analysis shown.
  • lambda (CDR) = not reported
    Regularization strength in Eq. 7; not specified, yet CDR results depend on it.
  • mu (LTAO) = not reported
    KL-divergence regularization weight in Eq. 8; not specified, yet LTAO results depend on it.
  • gamma (CCR) = not reported
    Re-ranking adjustment strength in Eq. 9; not specified, yet CCR results depend on it.
  • alpha (CPF) = not reported
    Penalty weight in Eq. 10; not specified, yet CPF results depend on it.
  • number of recommendation rounds = not stated
    Appendix B refers to rounds 10, 20, and 30, but the total number of rounds and the update schedule are never explicitly defined.
axioms (4)
  • domain assumption Users' clicks in multi-round simulation can be generated from recommender scores or some implicit policy.
    Section 3.1.3 states 'we simulate users' click behavior' without defining the policy; the entire click repeat rate and multi-round dynamics depend on this unspecified assumption.
  • domain assumption Topic categories and subcategories are reliable and complete signals of information cocoon severity.
    Section 3.1.1 and 3.1.2 define cocoon through category counts and entropy; if categories are noisy or incomplete, the metric picture changes.
  • domain assumption Louvain communities on the user-news bipartite graph correspond to meaningful echo chambers or polarized groups.
    Section 4.1.3 adopts Louvain and interprets communities as clusters of similar users and news; no validation that these communities map to opinion groups or echo chambers.
  • domain assumption Top-K recommendation lists and simulated clicks are sufficient proxies for long-term evolution of user information exposure.
    The paper infers cocoon deepening across rounds solely from these proxies; no comparison to actual user behavior or external validation is presented.

pith-pipeline@v1.3.0-alltime-deepseek · 18522 in / 11971 out tokens · 129435 ms · 2026-08-04T17:01:48.740071+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Understanding the Information Cocoon: A Multidimensional Assessment and Analysis of News Recommendation Systems." pith.science (2026). https://pith.science/paper/ZZINU36U

@misc{pith2026250911139,
  author       = {Pith},
  title        = {Pith review of: Understanding the Information Cocoon: A Multidimensional Assessment and Analysis of News Recommendation Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZZINU36U}},
  note         = {Machine review of arXiv:2509.11139}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Personalized news recommendation systems inadvertently create information cocoons--homogeneous information bubbles that reinforce user biases and amplify societal polarization. To address the lack of comprehensive assessment frameworks in prior research, we propose a multidimensional analysis that evaluates cocoons through dual perspectives: (1) Individual homogenization via topic diversity (including the number of topic categories and category information entropy) and click repetition; (2) Group polarization via network density and community openness. Through multi-round experiments on real-world datasets, we benchmark seven algorithms and reveal critical insights. Furthermore, we design five lightweight mitigation strategies. This work establishes the first unified metric framework for information cocoons and delivers deployable solutions for ethical recommendation systems.

Figures

Figures reproduced from arXiv: 2509.11139 by Jitao Sang, Xiaowen Huang, Xin Wang.

Figure 1
Figure 1. Figure 1: Sentiment analysis of users’ clicked news. The higher means and the lower variance imply greater emo￾tional polarization risk. Existing research primarily focuses on detecting and mit￾igating information cocoons [2]. Some works explore the emergence and development mechanisms of information cocoons [36, 40]. Several works have attempted to study the information cocoon effect at the individual user level [1… view at source ↗
Figure 2
Figure 2. Figure 2: Information cocoon effect in news recommenda￾tions. From an individual perspective, it appears as reduced diversity in recommendation lists. From a group perspective, it manifests as network clustering and closure. caused by multiple influencing factors, which is the primary focus of this study. We aim to analyze and assess the factors that influence the information cocoon from both individual and group pe… view at source ↗
Figure 3
Figure 3. Figure 3: The Sankey diagram compares the same five users’ original history with the Top-6 list topic categories after multiple recommendation rounds. The results of the NRMS model on the MIND dataset are used here as an example. (a) Before Recommendation (b) After Recommendation [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The User-Item networks are based on click history and its evolution across multiple recommendation rounds. Using the Louvain algorithm, nodes are grouped into com￾munities for different colors. It uses the results of the NRMS model and includes 6000 randomly selected users. lists after multiple rounds of recommendation, compared to their historical click lists, which is evident for both cate￾gory and subca… view at source ↗
Figure 5
Figure 5. Figure 5: Number of topic categories and category information entropy of the Top-20 list, and click repeat rate after multiple recommendation rounds for each model under the individual perspective. and NAML utilize attention mechanisms, with NAML addi￾tionally treating the news title, content, and topic categories as distinct views. This precise modeling approach makes the recommender highly focused on the user’s kn… view at source ↗
Figure 6
Figure 6. Figure 6: Network density after multiple recommendation rounds for each model under the group perspective. (a) MIND (b) Adressa [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Community openness after multiple recommenda￾tion rounds for each model under the group perspective. 𝑃 (𝑖 | 𝑢) = ( 1 | C𝑢 | , 𝜀 softmax(𝑠𝑢𝑖), 1 − 𝜀 (6) where C𝑢 denotes the candidate item set for user 𝑢, 𝑠𝑢𝑖 is the predicted relevance score of item 𝑖 for user 𝑢, 𝜀 ∈ (0, 1) is the exploration probability, 𝑃 (𝑖 | 𝑢) is the final recommendation probability for item 𝑖. Content Diversity Regularization (CDR) is… view at source ↗
Figure 8
Figure 8. Figure 8: The histogram of the scatter distribution across different ranges of the rounds in MIND, where the x-axis shows the NRMS model’s results of the indicators under the category and the y-axis shows the subcategory [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The results of all metrics on subcategory in MIND after multiple recommendation rounds for each model [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗

discussion (0)

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

Forward citations

Cited by 1 Pith paper

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

  1. Do Generative Recommenders Deepen the Information Cocoon? A Closed-Loop Simulation with LLM-powered User Simulators

    cs.IR 2026-06 unverdicted novelty 6.0

    Closed-loop LLM simulations find generative recommenders form fewer exposure-level information cocoons than traditional sequential baselines on Amazon data, though tokenization strategy and model scale affect concentr...

Reference graph

Works this paper leans on

71 extracted references · 1 canonical work pages · cited by 1 Pith paper

  1. [1]

    Mingxiao An, Fangzhao Wu, Chuhan Wu, Kun Zhang, Zheng Liu, and Xing Xie. 2019. Neural News Recommendation with Long- and Short- term User Representations. InProceedings of the 57th Annual Meeting of the Association for Computational Linguistics, Anna Korhonen, David Traum, and Lluís Màrquez (Eds.). Florence, Italy, 336–345

    2019

  2. [2]

    Md Sanzeed Anwar, Grant Schoenebeck, and Paramveer S. Dhillon

  3. [3]

    Chen Avin, Hadassa Daltrophe, and Zvi Lotker. 2024. On the impos- sibility of breaking the echo chamber effect in social media using regulation.Scientific Reports14, 1 (11 Jan 2024), 1107

    2024

  4. [4]

    Eytan Bakshy, Solomon Messing, and Lada A. Adamic. 2015. Exposure to ideologically diverse news and opinion on Facebook.Science348, 6239 (2015), 1130–1132

    2015

  5. [5]

    Alessandro Bellina, Claudio Castellano, Paul Pineau, Giulio Iannelli, and Giordano De Marzo. 2023. Effect of collaborative-filtering-based recommendation algorithms on opinion polarization.Physical Review E108, 5 (2023), 054304

    2023

  6. [6]

    Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Eti- enne Lefebvre

    Vincent D. Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Eti- enne Lefebvre. 2008. Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory and Experiment2008 (2008), P10008

    2008

  7. [7]

    Wanyu Chen, Pengjie Ren, Fei Cai, Fei Sun, and Maarten de Rijke

  8. [8]

    Federico Cinus, Marco Minici, Corrado Monti, and Francesco Bonchi

  9. [9]

    Tim Donkers and Jürgen Ziegler. 2021. The Dual Echo Chamber: Modeling Social Media Polarization for Interventional Recommending. InProceedings of the 15th ACM Conference on Recommender Systems (Amsterdam, Netherlands)(RecSys ’21). New York, NY, USA, 12–22

    2021

  10. [10]

    Wenqi Fan, Yao Ma, Qing Li, Yuan He, Eric Zhao, Jiliang Tang, and Dawei Yin. 2019. Graph Neural Networks for Social Recommendation. InThe World Wide Web Conference(San Francisco, CA, USA)(WWW ’19). New York, NY, USA, 417–426

    2019

  11. [11]

    Chongming Gao, Shiqi Wang, Shijun Li, Jiawei Chen, Xiangnan He, Wenqiang Lei, Biao Li, Yuan Zhang, and Peng Jiang. 2023. CIRS: Bursting Filter Bubbles by Counterfactual Interactive Recommender System.ACM Trans. Inf. Syst.42, 1, Article 14 (Aug. 2023), 27 pages

    2023

  12. [12]

    Jie Gao, Xin Xin, Junshuai Liu, Rui Wang, Jing Lu, Biao Li, Xin Fan, and Ping Guo. 2018. Fine-Grained Deep Knowledge-Aware Network for News Recommendation with Self-Attention. In2018 IEEE/WIC/ACM International Conference on Web Intelligence (WI). 81–88

    2018

  13. [13]

    Granovetter

    Mark S. Granovetter. 1973. The Strength of Weak Ties.Amer. J. Sociology78, 6 (1973), 1360–1380

    1973

  14. [14]

    Quentin Grossetti, Cédric du Mouza, and Nicolas Travers. 2019. Community-Based Recommendations on Twitter: Avoiding the Filter Bubble. InWeb Information Systems Engineering – WISE 2019, Reynold Cheng, Nikos Mamoulis, Yizhou Sun, and Xin Huang (Eds.). Springer International Publishing, Cham, 212–227

    2019

  15. [15]

    Ming Gu, Tian-Fang Zhao, Liang Yang, Xiao-Kun Wu, and Wei-Neng Chen. 2024. Modeling information cocoons in networked populations: Insights from backgrounds and preferences.IEEE Transactions on Computational Social Systems11, 3 (2024), 4497–4510

    2024

  16. [16]

    Jon Atle Gulla, Lemei Zhang, Peng Liu, Özlem Özgöbek, and Xiaomeng Su. 2017. The adressa dataset for news recommendation. InProceedings of the international conference on web intelligence. 1042–1048

    2017

  17. [17]

    Filter Bubble

    Han Han, Can Wang, Yunwei Zhao, Min Shu, Wenlei Wang, and Yong Min. 2022. SSLE: A framework for evaluating the “Filter Bubble” effect on the news aggregator and recommenders.World Wide Web25, 3 (May 2022), 1169–1195

    2022

  18. [18]

    Chen Hao, Xie Runfeng, Cui Xiangyang, Yan Zhou, Wang Xin, Xuan Zhanwei, and Zhang Kai. 2024. LKPNR: Large Language Models and Knowledge Graph for Personalized News Recommendation Frame- work.Computers, Materials & Continua79, 3 (2024), 4283–4296

    2024

  19. [19]

    Yuxuan Hu, Shiqing Wu, Chenting Jiang, Weihua Li, Quan Bai, and Erin Roehrer. 2022. AI Facilitated Isolations? The Impact of Recommendation-based Influence Diffusion in Human Society.. In IJCAI. 5080–5086

    2022

  20. [20]

    Dimitris Kalimeris, Smriti Bhagat, Shankar Kalyanaraman, and Udi Weinsberg. 2021. Preference Amplification in Recommender Systems. InProceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining (KDD ’21). 805–815

    2021

  21. [21]

    Dongho Lee, Byungkook Oh, Seungmin Seo, and Kyong-Ho Lee. 2020. News Recommendation with Topic-Enriched Knowledge Graphs. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management(Virtual Event, Ireland)(CIKM ’20). New York, NY, USA, 695–704

    2020

  22. [22]

    Nian Li, Chen Gao, Jinghua Piao, Xin Huang, Aizhen Yue, Liang Zhou, Qingmin Liao, and Yong Li. 2022. An Exploratory Study of Information Cocoon on Short-form Video Platform. InProceedings of the 31st ACM International Conference on Information & Knowledge Management Understanding the Information Cocoon: A Multidimensional Assessment and Analysis of News R...

    2022

  23. [23]

    Nian Li, Chen Gao, Jinghua Piao, Xin Huang, Aizhen Yue, Liang Zhou, Qingmin Liao, and Yong Li. 2022. An Exploratory Study of Information Cocoon on Short-form Video Platform. InProceedings of the 31st ACM International Conference on Information & Knowledge Management (Atlanta, GA, USA)(CIKM ’22). New York, NY, USA, 4178–4182

    2022

  24. [24]

    Yijiang Li and Jun Wu. 2024. Improving Prompt-based News Rec- ommendation with Individual Template and Customized Answer. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management(Boise, ID, USA)(CIKM ’24). New York, NY, USA, 3887–3891

    2024

  25. [25]

    Zhenyang Li, Yancheng Dong, Chen Gao, Yizhou Zhao, Dong Li, Jianye Hao, Kai Zhang, Yong Li, and Zhi Wang. 2023. Breaking filter bubble: A reinforcement learning framework of controllable recommender system. InProceedings of the ACM web conference 2023. 4041–4049

    2023

  26. [26]

    Yile Liang, Tieyun Qian, Qing Li, and Hongzhi Yin. 2021. Enhancing Domain-Level and User-Level Adaptivity in Diversified Recommen- dation. InProceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval(Virtual Event, Canada)(SIGIR ’21). New York, NY, USA, 747–756

    2021

  27. [27]

    Zihan Lin, Hui Wang, Jingshu Mao, Wayne Xin Zhao, Cheng Wang, Peng Jiang, and Ji-Rong Wen. 2022. Feature-aware Diversified Re- ranking with Disentangled Representations for Relevant Recommenda- tion. InProceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD ’22). 3327–3335

    2022

  28. [28]

    Jian Liu, Shaohua Zhang, Siying Yan, Xianchao Zhang, and Wei Ren

  29. [29]

    Qijiong Liu, Nuo Chen, Tetsuya Sakai, and Xiao-Ming Wu. 2023. A First Look at LLM-Powered Generative News Recommendation.CoRR abs/2305.06566 (2023). arXiv:2305.06566

    Pith/arXiv arXiv 2023

  30. [30]

    Zheng Liu, Yu Xing, Fangzhao Wu, Mingxiao An, and Xing Xie. 2019. Hi-Fi Ark: deep user representation via high-fidelity archive network. InProceedings of the 28th International Joint Conference on Artificial Intelligence(Macao, China)(IJCAI’19). AAAI Press, 3059–3065

    2019

  31. [31]

    Saravanapriya Manoharan, Radha Senthilkumar, and José Alfredo Hernández-Pérez. 2020. An Intelligent Fuzzy Rule-Based Personalized News Recommendation Using Social Media Mining.Intell. Neuroscience 2020 (Jan. 2020), 10 pages

    2020

  32. [32]

    In2024 IEEE 11th International Conference on Cyber Security and Cloud Computing (CSCloud)

    Formal Analysis on Interaction Flow and Information Cocoon Based on Probabilistic Graph. In2024 IEEE 11th International Conference on Cyber Security and Cloud Computing (CSCloud). IEEE, 96–100

  33. [33]

    Lien Michiels, Jorre Vannieuwenhuyze, Jens Leysen, Robin Verachtert, Annelien Smets, and Bart Goethals. 2023. How Should We Measure Filter Bubbles? A Regression Model and Evidence for Online News. InProceedings of the 17th ACM Conference on Recommender Systems (Singapore, Singapore)(RecSys ’23). New York, NY, USA, 640–651

    2023

  34. [34]

    Thi Nguyen

    C. Thi Nguyen. 2020. ECHO CHAMBERS AND EPISTEMIC BUBBLES. Episteme17, 2 (2020), 141–161

    2020

  35. [35]

    Nguyen, Pik-Mai Hui, F

    Tien T. Nguyen, Pik-Mai Hui, F. Maxwell Harper, Loren Terveen, and Joseph A. Konstan. 2014. Exploring the filter bubble: the effect of using recommender systems on content diversity. InProceedings of the 23rd International Conference on World Wide Web(Seoul, Korea)(WWW ’14). New York, NY, USA, 677–686

    2014

  36. [36]

    Dana McKay, Kaipin Owyong, Stephann Makri, and Marisela Gutier- rez Lopez. 2022. Turn and face the strange: Investigating filter bubble bursting information interactions. InProceedings of the 2022 Conference on Human Information Interaction and Retrieval. 233–242

    2022

  37. [37]

    Tao Qi, Fangzhao Wu, Chuhan Wu, Yongfeng Huang, and Xing Xie

  38. [38]

    Shaoyun Shi, Weizhi Ma, Zhen Wang, Min Zhang, Kun Fang, Jingfang Xu, Yiqun Liu, and Shaoping Ma. 2021. WG4Rec: Modeling Textual Content with Word Graph for News Recommendation. InProceedings of the 30th ACM International Conference on Information & Knowledge Management(Virtual Event, Queensland, Australia)(CIKM ’21). New York, NY, USA, 1651–1660

    2021

  39. [39]

    Dusan Stamenkovic, Alexandros Karatzoglou, Ioannis Arapakis, Xin Xin, and Kleomenis Katevas. 2022. Choosing the best of both worlds: Diverse and novel recommendations through multi-objective rein- forcement learning. InProceedings of the fifteenth acm international conference on web search and data mining. 957–965

    2022

  40. [40]

    Jinghua Piao, Jiazhen Liu, Fang Zhang, Jun Su, and Yong Li. 2023. Human–AI adaptive dynamics drives the emergence of information cocoons.Nature Machine Intelligence5, 11 (01 Nov 2023), 1214–1224

    2023

  41. [41]

    Sunstein

    Cass R. Sunstein. 2001.Republic.com. Princeton University Press, Princeton, NJ

    2001

  42. [42]

    In Findings of the Association for Computational Linguistics: EMNLP 2020, Trevor Cohn, Yulan He, and Yang Liu (Eds.)

    Privacy-Preserving News Recommendation Model Learning. In Findings of the Association for Computational Linguistics: EMNLP 2020, Trevor Cohn, Yulan He, and Yang Liu (Eds.). Online, 1423–1432

    2020

  43. [43]

    Hongwei Wang, Fuzheng Zhang, Min Hou, Xing Xie, Minyi Guo, and Qi Liu. 2018. SHINE: Signed Heterogeneous Information Network Em- bedding for Sentiment Link Prediction. InProceedings of the Eleventh ACM International Conference on Web Search and Data Mining(Marina Del Rey, CA, USA)(WSDM ’18). New York, NY, USA, 592–600

    2018

  44. [44]

    Hongwei Wang, Fuzheng Zhang, Jialin Wang, Miao Zhao, Wenjie Li, Xing Xie, and Minyi Guo. 2018. RippleNet: Propagating User Prefer- ences on the Knowledge Graph for Recommender Systems. InPro- ceedings of the 27th ACM International Conference on Information and Knowledge Management(Torino, Italy)(CIKM ’18). New York, NY, USA, 417–426

    2018

  45. [45]

    Nicholas Sukiennik, Chen Gao, and Nian Li. 2024. Uncovering the deep filter bubble: narrow exposure in short-video recommendation. InProceedings of the ACM Web Conference 2024. 4727–4735

    2024

  46. [46]

    Likun Wang and Kyungyee Kim. 2024. Analyzing group polarization through text emotion measurement and time series prediction: A com- parative study across three online platforms.Measurement: Sensors33 (2024), 101216

    2024

  47. [47]

    Antonela Tommasel, Juan Manuel Rodriguez, and Daniela Godoy. 2021. I Want to Break Free! Recommending Friends from Outside the Echo Chamber. InProceedings of the 15th ACM Conference on Recommender Systems(Amsterdam, Netherlands)(RecSys ’21). New York, NY, USA, 23–33

    2021

  48. [48]

    Chuhan Wu, Fangzhao Wu, Mingxiao An, Jianqiang Huang, Yongfeng Huang, and Xing Xie. 2019. Neural News Recommendation with Attentive Multi-View Learning. InInternational Joint Conference on Artificial Intelligence

    2019

  49. [49]

    Chuhan Wu, Fangzhao Wu, Mingxiao An, Jianqiang Huang, Yongfeng Huang, and Xing Xie. 2019. NPA: Neural News Recommendation with Personalized Attention. InProceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining(An- chorage, AK, USA)(KDD ’19). New York, NY, USA, 2576–2584

    2019

  50. [50]

    Hongwei Wang, Fuzheng Zhang, Xing Xie, and Minyi Guo. 2018. DKN: Deep Knowledge-Aware Network for News Recommendation. InPro- ceedings of the 2018 World Wide Web Conference(Lyon, France)(WWW ’18). International World Wide Web Conferences Steering Committee, Republic and Canton of Geneva, CHE, 1835–1844

    2018

  51. [51]

    Chuhan Wu, Fangzhao Wu, Suyu Ge, Tao Qi, Yongfeng Huang, and Xing Xie. 2019. Neural News Recommendation with Multi-Head Self- Attention. InProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference AAAI’26, Jan.20-27, Singapore Xin Wang, Xiaowen Huang, and Jitao Sang on Natural Languag...

    2019

  52. [52]

    Wenjie Wang, Fuli Feng, Liqiang Nie, and Tat-Seng Chua. 2022. User- controllable Recommendation Against Filter Bubbles. InProceedings of the 45th International ACM SIGIR Conference on Research and Devel- opment in Information Retrieval(Madrid, Spain)(SIGIR ’22). New York, NY, USA, 1251–1261

    2022

  53. [53]

    Chuhan Wu, Fangzhao Wu, Tao Qi, and Yongfeng Huang. 2021. User modeling with click preference and reading satisfaction for news rec- ommendation. InProceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence(Yokohama, Yokohama, Japan)(IJ- CAI’20). Article 418, 7 pages

    2021

  54. [54]

    Chuhan Wu, Fangzhao Wu, Tao Qi, and Yongfeng Huang. 2022. Two Birds with One Stone: Unified Model Learning for Both Recall and Ranking in News Recommendation. InFindings of the Association for Computational Linguistics: ACL 2022, Smaranda Muresan, Preslav Nakov, and Aline Villavicencio (Eds.). Dublin, Ireland, 3474–3480

    2022

  55. [55]

    Chuhan Wu, Fangzhao Wu, Mingxiao An, Yongfeng Huang, and Xing Xie. 2019. Neural News Recommendation with Topic-Aware News Representation. InProceedings of the 57th Annual Meeting of the Asso- ciation for Computational Linguistics, Anna Korhonen, David Traum, and Lluís Màrquez (Eds.). Florence, Italy, 1154–1159

    2019

  56. [56]

    Fangzhao Wu, Ying Qiao, Jiun-Hung Chen, Chuhan Wu, Tao Qi, Jianxun Lian, Danyang Liu, Xing Xie, Jianfeng Gao, Winnie Wu, and Ming Zhou. 2020. MIND: A Large-scale Dataset for News Recommen- dation. InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics, Dan Jurafsky, Joyce Chai, Natalie Schluter, and Joel Tetreault (Eds.)...

    2020

  57. [57]

    Chuhan Wu, Fangzhao Wu, Yongfeng Huang, and Xing Xie. 2021. User-as-Graph: User Modeling with Heterogeneous Graph Pooling for News Recommendation. InProceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21, Zhi-Hua Zhou (Ed.). International Joint Conferences on Artificial Intelligence Organization, 1624–1630. doi:10...

    work page doi:10.24963/ijcai.2021/224main 2021

  58. [58]

    Yuanqing Yu, Chongming Gao, Jiawei Chen, Heng Tang, Yuefeng Sun, Qian Chen, Weizhi Ma, and Min Zhang. 2024. EasyRL4Rec: An Easy-to-use Library for Reinforcement Learning Based Recommender Systems. InProceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval(Washington DC, USA)(SIGIR ’24). New York, NY, ...

    2024

  59. [59]

    Zhaodian Zeng, Yining Wang, Yanbin Zhao, and Wenxuan Shi. 2024. A Survey of Music Recommendation Systems. InProceedings of the 5th International Conference on Computer Information and Big Data Appli- cations(Wuhan, China)(CIBDA ’24). New York, NY, USA, 507–519

    2024

  60. [60]

    Chuhan Wu, Fangzhao Wu, Tao Qi, Chenliang Li, and Yongfeng Huang

  61. [61]

    InProceedings of the 45th International ACM SIGIR Conference on Re- search and Development in Information Retrieval(Madrid, Spain)(SIGIR ’22)

    Is News Recommendation a Sequential Recommendation Task?. InProceedings of the 45th International ACM SIGIR Conference on Re- search and Development in Information Retrieval(Madrid, Spain)(SIGIR ’22). New York, NY, USA, 2382–2386

  62. [62]

    Qi Zhang, Jingjie Li, Qinglin Jia, Chuyuan Wang, Jieming Zhu, Zhaowei Wang, and Xiuqiang He. 2021. UNBERT: User-News Matching BERT for News Recommendation. InProceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21, Zhi-Hua Zhou (Ed.). International Joint Conferences on Artificial Intelligence Organization, 3356–336...

    2021

  63. [63]

    Yu, and Taiqing Wang

    Liangwei Yang, Shengjie Wang, Yunzhe Tao, Jiankai Sun, Xiaolong Liu, Philip S. Yu, and Taiqing Wang. 2023. DGRec: Graph Neural Network for Recommendation with Diversified Embedding Generation. InProceedings of the Sixteenth ACM International Conference on Web Search and Data Mining(Singapore, Singapore)(WSDM ’23). New York, NY, USA, 661–669

    2023

  64. [64]

    Yongsen Zheng, Ziliang Chen, Jinghui Qin, and Liang Lin. 2024. FacetCRS: multi-faceted preference learning for pricking filter bubbles in conversational recommender system. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 38. 9405–9413

    2024

  65. [65]

    Yu Zheng, Chen Gao, Liang Chen, Depeng Jin, and Yong Li. 2021. DGCN: Diversified Recommendation with Graph Convolutional Net- works. InProceedings of the Web Conference 2021(Ljubljana, Slovenia) (WWW ’21). New York, NY, USA, 401–412. A EXPERIMENTAL DETAILS A.1 News Recommenders In this part, we present detailed information about the seven typical news rec...

    Pith/arXiv arXiv 2021

  66. [66]

    Han Zhang, Ziwei Zhu, and James Caverlee. 2023. Evolution of fil- ter bubbles and polarization in news recommendation. InEuropean Conference on Information Retrieval. Springer, 685–693

    2023

  67. [67]

    Lemei Zhang, Peng Liu, and Jon Atle Gulla. 2019. Dynamic attention- integrated neural network for session-based news recommendation. Mach. Learn.108, 10 (Oct. 2019), 1851–1875

    2019

  68. [69]

    Tao Zhang, Luwei Yang, Zhibo Xiao, Wen Jiang, and Wei Ning. 2024. On Practical Diversified Recommendation with Controllable Category Diversity Framework. InCompanion Proceedings of the ACM Web Con- ference 2024(Singapore, Singapore)(WWW ’24). New York, NY, USA, 255–263

    2024

  69. [2020]

    InProceedings of the 29th ACM International Conference on Information & Knowledge Management(Virtual Event, Ireland)(CIKM ’20)

    Improving End-to-End Sequential Recommendations with Intent- aware Diversification. InProceedings of the 29th ACM International Conference on Information & Knowledge Management(Virtual Event, Ireland)(CIKM ’20). New York, NY, USA, 175–184

  70. [2022]

    InProceedings of the International AAAI Conference on Web and Social Media, Vol

    The effect of people recommenders on echo chambers and polarization. InProceedings of the International AAAI Conference on Web and Social Media, Vol. 16. 90–101

  71. [2024]

    In Proceedings of the ACM Web Conference 2024(Singapore, Singapore) (WWW ’24)

    Filter Bubble or Homogenization? Disentangling the Long- Term Effects of Recommendations on User Consumption Patterns. In Proceedings of the ACM Web Conference 2024(Singapore, Singapore) (WWW ’24). New York, NY, USA, 123–134

    2024