Submit to Special Issue Submit Abstract to Special Issue Review for Entropy Propose a Special Issue

Journal Menu

Journal Browser

Causal Inference in Recommender Systems

Print Special Issue Flyer
Special Issue Editors
Special Issue Information
Keywords
Benefits of Publishing in a Special Issue
Published Papers

A special issue of Entropy (ISSN 1099-4300). This special issue belongs to the section "Multidisciplinary Applications".

Deadline for manuscript submissions: 15 April 2025 | Viewed by 4694

Share This Special Issue

Special Issue Editors

Prof. Dr. Fuli Feng

E-Mail Website
Guest Editor

School of Cyber Science and Technology, University of Science and Technology of China, Hefei 230052, China
Interests: recommendation; information retrieval; causal inference; large language model; natural language processing

Dr. Xu Chen

E-Mail Website
Guest Editor

Gaoling School of Artificial Intelligence, Renmin University of China, Beijing 100872, China
Interests: recommender system; large language models; causal inference; explainable AI; reinforcement learning

Special Issue Information

Dear Colleagues,

The recommender system serves many users with personalized information filtering across a wide spectrum of online applications such as e-commerce, search engines, and social media. Recent years have witnessed the success of incorporating causal inference theories and techniques into recommender systems to enhance the user experience regarding the accuracy of user preference modeling and estimation, as well as the fairness, unbiasedness, and transparency of recommendations. In addition, these recommender systems also draw upon concepts from entropy and information theory. The connection between these directions indicates opportunities to futher improve the performance of recommender systems. For example, recommender systems can better understand and predict user behavior by considering the entropy of user preferences and the information gain obtained through causal inference models. This Special Issue is aimed at bringing together the most contemporary achievements and breakthroughs in the field of recommender systems that embrace causal inference and information theory. We invite novel contributions on topics including, but not restricted to, the following:

Causal view of recommender system;
Causal user modeling;
Causal effect estimation for recommendation;
Bias and debias in recommender system;
Causal representation learning;
Counterfactual learning for recommendation;
Uncertainty of recommendation;
Information decomposition for user modeling;
Causal discovery in recommender system;
Causal explanation for recommendation;
Causal evaluation of recommender system;
Causal tools and resources of recommender system;
Unmeasured confounder modeling based on information theory;
Debiased recommendation based on information theory.

Technical Committee Member

Name: Haoxuan Li
Email: [email protected]
Affiliation: Center for Data Science, Peking University, Beijing 100091, China
Interests: causal inference; recommendation; selection bias; fairness; large language model
Website: https://pattern.swarma.org/user/62913

Prof. Dr. Fuli Feng
Dr. Xu Chen
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Entropy is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

recommendation
causal inference
inference retrieval
ranking
information theory
information bottleneck
representation learning
counterfactual explanation

Benefits of Publishing in a Special Issue

Ease of navigation: Grouping papers by topic helps scholars navigate broad scope journals more efficiently.
Greater discoverability: Special Issues support the reach and impact of scientific research. Articles in Special Issues are more discoverable and cited more frequently.
Expansion of research network: Special Issues facilitate connections among authors, fostering scientific collaborations.
External promotion: Articles in Special Issues are often promoted through the journal's social media, increasing their visibility.
e-Book format: Special Issues with more than 10 articles can be published as dedicated e-books, ensuring wide and rapid dissemination.

Further information on MDPI's Special Issue polices can be found here.

Published Papers (4 papers)

Download All Papers

Order results

Result details

Show export options Show export options

Select all

Export citation of selected articles as:

Research

20 pages, 932 KiB

Open AccessArticle

Gradient-Based Multiple Robust Learning Calibration on Data Missing-Not-at-Random via Bi-Level Optimization

by Shuxia Gong and Chen Ma

Entropy 2025, 27(2), 196; https://doi.org/10.3390/e27020196 - 13 Feb 2025

Abstract

Recommendation systems (RS) have become integral to numerous digital platforms and applications, ranging from e-commerce to content streaming field. A critical problem in RS is that the ratings are missing not at random (MNAR), which is due to the users always giving feedback on items with self-selection. The biased selection of rating data results in inaccurate rating prediction for all user-item pairs. Doubly robust (DR) learning has been studied in many tasks in RS, which is unbiased when either a single imputation or a single propensity model is accurate. In addition, multiple robust (MR) has been proposed with multiple imputation models and propensity models, and is unbiased when there exists a linear combination of these imputation models and propensity models is correct. However, we claim that the imputed errors and propensity scores are miscalibrated in the MR method. In this paper, we propose a gradient-based calibrated multiple robust learning method to enhance the debiasing performance and reliability of the rating prediction model. Specifically, we propose to use bi-level optimization to solve the weights and model coefficients of each propensity and imputation model in MR framework. Moreover, we adopt the differentiable expected calibration error as part of the objective to optimize the model calibration quality directly. Experiments on three real-world datasets show that our method outperforms the state-of-the-art baselines. Full article

(This article belongs to the Special Issue Causal Inference in Recommender Systems)

► Show Figures

Figure 1

16 pages, 2114 KiB

Open AccessArticle

SPIRIT: Structural Entropy Guided Prefix Tuning for Hierarchical Text Classification

by He Zhu, Jinxiang Xia, Ruomei Liu and Bowen Deng

Entropy 2025, 27(2), 128; https://doi.org/10.3390/e27020128 - 26 Jan 2025

Viewed by 172

Abstract

Hierarchical text classification (HTC) is a challenging task that requires classifiers to solve a series of multi-label subtasks considering hierarchical dependencies among labels. Recent studies have introduced prompt tuning to create closer connections between the language model (LM) and the complex label hierarchy. However, we find that the model’s attention to the prompt gradually decreases as the prompt moves from the input to the output layer, revealing the limitations of previous prompt tuning methods for HTC. Given the success of prefix tuning-based studies in natural language understanding tasks, we introduce Structural entroPy guIded pRefIx Tuning (SPIRIT). Specifically, we extract the essential structure of the label hierarchy via structural entropy minimization and decode the abstractive structural information as the prefix to prompt all intermediate layers in the LM. Additionally, a depth-wise reparameterization strategy is developed to enhance optimization and propagate the prefix throughout the LM layers. Extensive evaluation on four popular datasets demonstrates that SPIRIT achieves a state-of-the-art performance. Full article

(This article belongs to the Special Issue Causal Inference in Recommender Systems)

► Show Figures

Figure 1

Figure 1
In HPT [<a href="#B6-entropy-27-00128" class="html-bibr">6</a>], the attention scores from text tokens to prompts significantly drop in the deeper LM layers. Different colored lines represent different samples. Full article ">Figure 2
An example of SPIRIT with <math display="inline"><semantics> <mrow> <mi>K</mi> <mo>=</mo> <mn>3</mn> </mrow> </semantics></math>. Given a pretrained LM, we first feed input prompt (<a href="#sec3dot1-entropy-27-00128" class="html-sec">Section 3.1</a>) to the model, then we utilize the structural entropy guided prefix network (<a href="#sec3dot2-entropy-27-00128" class="html-sec">Section 3.2</a>) to generate hierarchy-aware prefix tokens. The LM is fine-tuned with ZLPR loss and MLM loss (<a href="#sec3dot3-entropy-27-00128" class="html-sec">Section 3.3</a>). After observation of the prompt and prefix tokens, the model is constrained to predict labels according to the hierarchy (<a href="#sec3dot3-entropy-27-00128" class="html-sec">Section 3.3</a>). Full article ">Figure 3
Test performance of SPIRIT with <math display="inline"><semantics> <mrow> <mi>K</mi> <mo>∈</mo> <mo>{</mo> <mn>2</mn> <mo>,</mo> <mn>3</mn> <mo>,</mo> <mn>4</mn> <mo>,</mo> <mn>5</mn> <mo>,</mo> <mn>6</mn> <mo>}</mo> </mrow> </semantics></math> on WOS and NYTimes. Full article ">Figure 4
Evaluation on P metrics and C metrics. Full article ">Figure 5
LM attention scores from the text tokens to prompt to the prompt contents. Full article ">Figure 6
A heat map on attention scores to the prefixes in LM layers. Case: WOS # 619, Ground Truth: [‘Medical’, ‘Menopause’]. Model Prediction: [‘Medical’, ‘Menopause’]. Full article ">Figure A1
An example of coding tree with height <math display="inline"><semantics> <mrow> <mi>K</mi> <mo>=</mo> <mn>3</mn> </mrow> </semantics></math> for synthesized simple graph. Full article ">

15 pages, 2003 KiB

Open AccessArticle

Invariant Representation Learning in Multimedia Recommendation with Modality Alignment and Model Fusion

by Xinghang Hu and Haiteng Zhang

Entropy 2025, 27(1), 56; https://doi.org/10.3390/e27010056 - 10 Jan 2025

Viewed by 460

Abstract

Multimedia recommendation systems aim to accurately predict user preferences from multimodal data. However, existing methods may learn a recommendation model from spurious features, i.e., appearing to be related to an outcome but actually having no causal relationship with the outcome, leading to poor generalization ability. While previous approaches have adopted invariant learning to address this issue, they simply concatenate multimodal data without proper alignment, resulting in information loss or redundancy. To overcome these challenges, we propose a framework called M³-InvRL, designed to enhance recommendation system performance through common and modality-specific representation learning, invariant learning, and model merging. Specifically, our approach begins by learning modality-specific representations along with a common representation for each modality. To achieve this, we introduce a novel contrastive loss that aligns representations and imposes mutual information constraints to extract modality-specific features, thereby preventing generalization issues within the same representation space. Next, we generate invariant masks based on the identification of heterogeneous environments to learn invariant representations. Finally, we integrate both invariant-specific and shared invariant representations for each modality to train models and fuse them in the output space, reducing uncertainty and enhancing generalization performance. Experiments on real-world datasets demonstrate the effectiveness of our approach. Full article

(This article belongs to the Special Issue Causal Inference in Recommender Systems)

► Show Figures

Figure 1

14 pages, 634 KiB

Open AccessArticle

Debiasing the Conversion Rate Prediction Model in the Presence of Delayed Implicit Feedback

by Taojun Hu and Xiao-Hua Zhou

Entropy 2024, 26(9), 792; https://doi.org/10.3390/e26090792 - 15 Sep 2024

Viewed by 1375

Abstract

The recommender system (RS) has been widely adopted in many applications, including online advertisements. Predicting the conversion rate (CVR) can help in evaluating the effects of advertisements on users and capturing users’ features, playing an important role in RS. In real-world scenarios, implicit rather than explicit feedback data are more abundant. Thus, directly training the RS with collected data may lead to suboptimal performance due to selection bias inherited from the nature of implicit feedback. Methods such as reweighting have been proposed to tackle selection bias; however, these methods omit delayed feedback, which often occurs due to limited observation times. We propose a novel likelihood approach combining the assumed parametric model for delayed feedback and the reweighting method to address selection bias. Specifically, the proposed methods minimize the likelihood-based loss using the multi-task learning method. The proposed methods are evaluated on the real-world Coat and Yahoo datasets. The proposed methods improve the AUC by 5.7% on Coat and 3.7% on Yahoo compared with the best baseline models. The proposed methods successfully debias the CVR prediction model in the presence of delayed implicit feedback. Full article

(This article belongs to the Special Issue Causal Inference in Recommender Systems)

► Show Figures

Journal Menu

Journal Browser

Causal Inference in Recommender Systems

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Benefits of Publishing in a Special Issue

Published Papers (4 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI