Generating recommendations for entity-oriented exploratory search
- URL: http://arxiv.org/abs/2204.00743v1
- Date: Sat, 2 Apr 2022 02:19:47 GMT
- Title: Generating recommendations for entity-oriented exploratory search
- Authors: David Wadden, Nikita Gupta, Kenton Lee, Kristina Toutanova
- Abstract summary: We introduce the task of recommendation set generation for entity-oriented exploratory search.
We build a text-to-text model capable of generating a collection of recommendations directly.
We train the model to generate recommendation sets which optimize a cost function designed to encourage comprehensiveness, interestingness, and non-redundancy.
- Score: 23.62142250257486
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: We introduce the task of recommendation set generation for entity-oriented
exploratory search. Given an input search query which is open-ended or
under-specified, the task is to present the user with an easily-understandable
collection of query recommendations, with the goal of facilitating domain
exploration or clarifying user intent. Traditional query recommendation systems
select recommendations by identifying salient keywords in retrieved documents,
or by querying an existing taxonomy or knowledge base for related concepts. In
this work, we build a text-to-text model capable of generating a collection of
recommendations directly, using the language model as a "soft" knowledge base
capable of proposing new concepts not found in an existing taxonomy or set of
retrieved documents. We train the model to generate recommendation sets which
optimize a cost function designed to encourage comprehensiveness,
interestingness, and non-redundancy. In thorough evaluations performed by crowd
workers, we confirm the generalizability of our approach and the high quality
of the generated recommendations.
Related papers
- Why Not Together? A Multiple-Round Recommender System for Queries and Items [37.709748983831034]
A fundamental technique of recommender systems involves modeling user preferences, where queries and items are widely used as symbolic representations of user interests.
We propose a novel approach named Multiple-round Auto Guess-and-Update System (MAGUS) that capitalizes on the synergies between both types.
arXiv Detail & Related papers (2024-12-14T10:49:00Z) - Preference Discerning with LLM-Enhanced Generative Retrieval [28.309905847867178]
We propose a new paradigm, which we term preference discerning.
In preference dscerning, we explicitly condition a generative sequential recommendation system on user preferences within its context.
We generate user preferences using Large Language Models (LLMs) based on user reviews and item-specific data.
arXiv Detail & Related papers (2024-12-11T18:26:55Z) - A Survey of Generative Search and Recommendation in the Era of Large Language Models [125.26354486027408]
generative search (retrieval) and recommendation aims to address the matching problem in a generative manner.
Superintelligent generative large language models have sparked a new paradigm in search and recommendation.
arXiv Detail & Related papers (2024-04-25T17:58:17Z) - Improving Retrieval in Theme-specific Applications using a Corpus
Topical Taxonomy [52.426623750562335]
We introduce ToTER (Topical taxonomy Enhanced Retrieval) framework.
ToTER identifies the central topics of queries and documents with the guidance of the taxonomy, and exploits their topical relatedness to supplement missing contexts.
As a plug-and-play framework, ToTER can be flexibly employed to enhance various PLM-based retrievers.
arXiv Detail & Related papers (2024-03-07T02:34:54Z) - InteraRec: Screenshot Based Recommendations Using Multimodal Large Language Models [0.6926105253992517]
We introduce a sophisticated and interactive recommendation framework denoted as InteraRec.
InteraRec captures high-frequency screenshots of web pages as users navigate through a website.
We demonstrate the effectiveness of InteraRec in providing users with valuable and personalized offerings.
arXiv Detail & Related papers (2024-02-26T17:47:57Z) - Evaluating Generative Ad Hoc Information Retrieval [58.800799175084286]
generative retrieval systems often directly return a grounded generated text as a response to a query.
Quantifying the utility of the textual responses is essential for appropriately evaluating such generative ad hoc retrieval.
arXiv Detail & Related papers (2023-11-08T14:05:00Z) - Impression-Aware Recommender Systems [53.48892326556546]
We present a systematic literature review on recommender systems using impressions.
We define a theoretical framework to delimit recommender systems using impressions and a novel paradigm for personalized recommendations, called impression-aware recommender systems.
arXiv Detail & Related papers (2023-08-15T16:16:02Z) - Recommender Systems with Generative Retrieval [58.454606442670034]
We propose a novel generative retrieval approach, where the retrieval model autoregressively decodes the identifiers of the target candidates.
To that end, we create semantically meaningful of codewords to serve as a Semantic ID for each item.
We show that recommender systems trained with the proposed paradigm significantly outperform the current SOTA models on various datasets.
arXiv Detail & Related papers (2023-05-08T21:48:17Z) - A Survey on Neural Recommendation: From Collaborative Filtering to
Content and Context Enriched Recommendation [70.69134448863483]
Research in recommendation has shifted to inventing new recommender models based on neural networks.
In recent years, we have witnessed significant progress in developing neural recommender models.
arXiv Detail & Related papers (2021-04-27T08:03:52Z) - Exploration-Exploitation Motivated Variational Auto-Encoder for
Recommender Systems [1.52292571922932]
We introduce an exploitation-exploration motivated variational auto-encoder (XploVAE) to collaborative filtering.
To facilitate personalized recommendations, we construct user-specific subgraphs, which contain the first-order proximity capturing observed user-item interactions.
A hierarchical latent space model is utilized to learn the personalized item embedding for a given user, along with the population distribution of all user subgraphs.
arXiv Detail & Related papers (2020-06-05T17:37:46Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.