A Soft Computing Approach for Learning to Aggregate Rankings

Muñoz, Javier Alvaro Vargas; Torres, Ricardo da Silva; Gonçalves, Marcos André

doi:10.1145/2806416.2806478

Cited by 13 publications

(7 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Some of the most widespread score-based methods are: CombMAX, CombMIN, CombSUM, Com-bANZ, CombMNZ). In turn, BordaCount, Median Rank Aggregation (MRA) and Reciprocal Rank Fusion (RRF) are popular order-based methods [Vargas Muñoz et al 2015]. However, although these algorithms have been used in many applications, they do not consider diversification explicitly.…”

Section: Related Workmentioning

confidence: 99%

A Novel Graph-based Diversity-aware Rank Fusion Method Applied to Image Metasearch

Figuerêdo,

Maia,

Calumby

2023

Anais Do XXXVIII Simpósio Brasileiro De Banco De Dados (SBBD 2023)

View full text Add to dashboard Cite

While search result diversification is used to handle ambiguous or underspecified queries, rank aggregation is a widely used approach in metasearch. However, current aggregation methods assume that the input rankings are built only according to the relevance of the items, disregarding the inter-relationship between images in each ranking. Hence, these methods tend to be inadequate for diversity-oriented retrieval. In this work, we introduce a diversity-aware rank fusion method that is validated in the context of diverse image metasearch. The experimental findings indicate that the proposed method significantly improves the overall diversity of metasearch results, in comparison to the state-of-the-art positional and score-based fusion methods.

show abstract

Section: Related Workmentioning

confidence: 99%

A Novel Graph-based Diversity-aware Rank Fusion Method Applied to Image Metasearch

Figuerêdo,

Maia,

Calumby

2023

Anais Do XXXVIII Simpósio Brasileiro De Banco De Dados (SBBD 2023)

View full text Add to dashboard Cite

show abstract

“…No primeiro grupo, a função de agregação leva como entrada as informações de pontuação dos objetos de cada lista. Na segunda, apenas a ordem relativa entre os documentos é considerada (Muñoz et al , 2015). Dentre as técnicas baseadas em escores, pode-se destacar a família Comb* (Shaw & Fox, 1994) (e.g., CombMIN, CombMAX, CombSUM, CombMED, CombMNZ e CombANZ).…”

Section: Metodologiaunclassified

Fusão De Rankings Para Recuperação De Imagens

Silva¹

2021

SEMIC

View full text Add to dashboard Cite

Dada a complexidade da consulta ou a pouca informação disponível para a configuraçãode um sistema de busca de imagens com realimentação de relevância, o conjunto inicialde resultados, anterior à primeira interação do usuário, pode não apresentar informaçõesrelevantes suficientes. Isso pode acarretar em um feedback com pouca informação econsequente dificultar a etapa de aprendizado. Por ser uma abordagem interativa, ainfluência do primeiro resultado propaga-se para as demais iterações, dado que oresultado inicial ruim limita a troca de informação entre o usuário e o sistema (Calumbyet al , 2017) e consequentemente os modelos de aprendizado da intenção do usuário.Visando atenuar este problema, é imprescindível que o resultado da primeira interaçãoseja o melhor possível.

show abstract

“…In the recent years, different methods have been proposed with the intent of fusing results from distinct visual features [10]- [13]. Most of them still require labels or user intervention to achieve effective results, and only few works have addressed aspects of selection without labeled data [14]- [18].…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Selective Rank Fusion on Content-Based Image Retrieval

Valem

Pedronette

2019

Anais Estendidos Do XXXII Conference on Graphics, Patterns and Images (SIBRAPI Estendido 2019)

View full text Add to dashboard Cite

Mainly due to the evolution of technologies to store and share images, the growth of image collections have been remarkable for years. Therefore, developing effective methods to index and retrieve such extensive available visual information is indispensable. The CBIR (Content-Based Image Retrieval) systems are one of the main solutions for image retrieval tasks. These systems are mainly supported by the use of different visual descriptors and machine learning methods. Despite the relevant advances in the area, mainly driven by deep learning technologies, accurately computing the similarity between images remains a complex task in various scenarios due to the well known semantic gap problem. As distinct features produce complementary ranking results with different effectiveness performance, a promising solution consists in combining them. However, how to decide which visual features to combine is a very challenging task. This work proposes three novel methods for selecting and combining ranked lists by estimating their effectiveness in an unsupervised way. The approaches were evaluated in five different image collections and several descriptors, achieving results comparable or superior to the state-of-the-art in most of the evaluated scenarios.

show abstract

A Soft Computing Approach for Learning to Aggregate Rankings

Cited by 13 publications

References 24 publications

A Novel Graph-based Diversity-aware Rank Fusion Method Applied to Image Metasearch

A Novel Graph-based Diversity-aware Rank Fusion Method Applied to Image Metasearch

Fusão De Rankings Para Recuperação De Imagens

Unsupervised Selective Rank Fusion on Content-Based Image Retrieval

Contact Info

Product

Resources

About