Improving preference elicitation in a conversational recommender system with active learning strategies

Iovine, Andrea; Lops, Pasquale; Narducci, Fedelucio; Gemmis, Marco de; Semeraro, Giovanni

doi:10.1145/3412841.3442013

Cited by 4 publications

(5 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We implement our proposed methodology as a food recommendation system where users can evaluate recipes by specifying their preferences over recipe attributes. The proposed methodology is not restricted to food recommendation and could be implemented for several domains like e-commerce, book, movie (Iovine et al, 2022), and fashion recommendation in which recommendation items either contain visual or textual descriptive information that can be evaluated by users to express preferences.…”

Section: Discussionmentioning

confidence: 99%

Explainable Active Learning for Preference Elicitation

Cantürk,

Aydoğan

2024

Preprint

View full text Add to dashboard Cite

Gaining insights into the preferences of new users and subsequently personaliz-ing recommendations necessitate managing user interactions intelligently, namely, posing pertinent questions to elicit valuable information effectively. In this study, our focus is on a specific scenario of the cold-start problem, where the recom-mendation system lacks adequate user presence or access to other users’ data is restricted, obstructing employing user profiling methods utilizing existing data in the system. We employ Active Learning (AL) to solve the addressed problem with the objective of maximizing information acquisition with minimal user effort. AL operates for selecting informative data from a large unlabeled set to inquire an oracle to label them and eventually updating a machine learning (ML) model. We operate AL in an integrated process of unsupervised, semi-supervised, and super-vised ML within an explanatory preference elicitation process. It harvests user feedback (given for the system’s explanations on the presented items) over infor-mative samples to update an underlying ML model estimating user preferences. The designed user interaction facilitates personalizing the system by incorporat-ing user feedback into the ML model and also enhances user trust by refining the system’s explanations on recommendations. We implement the proposed pref-erence elicitation methodology for food recommendation. We conducted human experiments to assess its efficacy in the short term and also experimented with several AL strategies over synthetic user profiles that we created for two food datasets, aiming for long-term performance analysis. The experimental results demonstrate the efficiency of the proposed preference elicitation with limited user-labeled data while also promoting transparency of recommendations through accurate explanations.

show abstract

Section: Discussionmentioning

confidence: 99%

Explainable Active Learning for Preference Elicitation

Cantürk,

Aydoğan

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Rubens et al describe existing AL methods with the viewpoint of the classical new-user problem in RSs that considers ratings by available users to estimate a rating vector of a user over all items (Rubens et al, 2015). Such studies on AL for RSs rely on data of available users in the systems, following the idea of collaborative filtering (Elahi, Ricci, & Rubens, 2014, 2016Iovine et al, 2022). For example, rating variance, popularity, rating entropy, and log-popularity-entropy, which balances popularity and entropy, are the metrics to determine which instances are selected to ask the new user to label them.…”

Section: Related Workmentioning

confidence: 99%

“…These metrics are calculated based on the ratings and/or activity of other users in the system. Alternatively, Iovine et al also use AL with content-based filtering by selecting dissimilar items from pre-existing items in the user profile for rating them by the user (Iovine et al, 2022). Despite AL being relatively more adopted with collaborative filtering, Hernández-Rubio et al propose a hybrid filtering approach with AL, Aspect-based AL (Hernández-Rubio, Bellogín, & Cantador, 2020).…”

Section: Related Workmentioning

confidence: 99%

“…Subsequently, an ML model is updated by including those informative samples into the training dataset. As getting information from users to learn their preferences can be tiresome for them, AL could be a convenient information acquisition method for preference elicitation, considering the cost of obtaining labels for items from users (Iovine, Lops, Narducci, de Gemmis, & Semeraro, 2022;Karimi, Nanopoulos, & Schmidt-Thieme, 2015;Pommeranz, Broekens, Wiggers, Brinkman, & Jonker, 2012;Rubens, Elahi, Sugiyama, & Kaplan, 2015). In addition to AL, we consider the explainability of RS as a second component to fertilize the information acquired through user interactions.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Explainable Active Learning for Preference Elicitation

Cantürk,

Aydoğan

2023

Preprint

View full text Add to dashboard Cite

Gaining insights into the preferences of new users and subsequently personalizing recommendations necessitate managing user interactions intelligently, namely, posing pertinent questions to elicit valuable information effectively. In this study, our focus is on a specific scenario of the cold-start problem, where the recommendation system lacks adequate user presence or access to other users' data is restricted, obstructing employing user profiling methods utilizing existing data in the system. We employ Active Learning (AL) to solve the addressed problem with the objective of maximizing information acquisition with minimal user effort. AL operates for selecting informative data from a large unlabeled set to inquire an oracle to label them and eventually updating a machine learning (ML) model. We operate AL in an integrated process of unsupervised, semi-supervised, and supervised ML within an explanatory preference elicitation process. It harvests user feedback (given for the system's explanations on the presented items) over informative samples to update an underlying ML model estimating user preferences. The designed user interaction facilitates personalizing the system by incorporating user feedback into the ML model and also enhances user trust by refining the system's explanations on recommendations. We implement the proposed preference elicitation methodology for food recommendation. We conducted human experiments to assess its efficacy in the short term and also experimented with several AL strategies over synthetic user profiles that we created for two food datasets, aiming for long-term performance analysis. The experimental results demonstrate the efficiency of the proposed preference elicitation with limited user-labeled data while also enhancing user trust through accurate explanations.

show abstract

“…This challenge hinders recommendation systems to make e ective and accurate recommendations. To better model users' interest evolution and improve recommendation performance, recent research has focused on Conversational Recommendation System (CRS) [10,15,19,21]. CRS aims to infer user interests by asking a few questions regarding user preference and then produces relevant recommendations by utilizing the responses [29,35,55].…”

Section: Introductionmentioning

confidence: 99%

Enhancing Conversational Recommendation Systems with Representation Fusion

et al. 2023

View full text Add to dashboard Cite

Conversational Recommendation Systems (CRSs) aim to improve recommendation performance by utilizing information from a conversation session. A CRS first constructs questions and then asks users for their feedback in each conversation session in order to refine better recommendation lists to users. The key design of CRS is to construct proper questions and obtain users’ feedback in response to these questions so as to effectively capture user preferences. Many CRS works have been proposed; however, they suffer from defects when constructing questions for users to answer: (1) employing a dialogue policy agent for constructing questions is one of the most common choices in CRS, but it needs to be trained with huge corpus; (2) it is not appropriate that constructing questions from a single policy (e.g., A CRS only selects attributes that the user has interacted with) for all users with different preferences. To address these defects, we propose a novel CRS model, namely, a R epresentation F usion-based C onversational R ecommendation model (RFCR), where the whole conversation session is divided into two subsessions (i.e., Local Question Search subsession and Global Question Search subsession) and two different question search methods are proposed to construct questions in the corresponding subsessions without employing policy agents. In particular, in the Local Question Search subsession we adopt a novel graph mining method to find questions, where the paths in the graph between users and attributes can eliminate irrelevant attributes; in the Global Question Search subsession we propose to initialize user preference on items with the user and all item historical rating records and construct questions based on user’s preference. Then, we update the embeddings independently over the two subsessions according to user’s feedback and fuse the final embeddings from the two subsessions for the recommendation. Experiments on three real-world recommendation datasets demonstrate that our proposed method outperforms five state-of-the-art baselines.

show abstract

Improving preference elicitation in a conversational recommender system with active learning strategies

Cited by 4 publications

References 18 publications

Explainable Active Learning for Preference Elicitation

Explainable Active Learning for Preference Elicitation

Explainable Active Learning for Preference Elicitation

Enhancing Conversational Recommendation Systems with Representation Fusion

Contact Info

Product

Resources

About