Pular para o conteúdo principal

Avaliação de proposta - Exploração de KG

Casos de Uso (Aluno do Schwabe)

The criteria to select the cases were the following:

1. The case should be published as a difficult case in the area. Since the case is published, we infer that the case is a real problematic situation faced by a community of data users, with reasonable complexity.

[Não é inventado pelo pesquisador, não teria viés de construção para validar a proposta mas pode ter viés na seleção]

2. The case is difficult to be solved with operators in the state-of-the-art tools. The rationale for this criterion is that we can use the case studies to compare the expressivity of our model against state-of-the-art tools using the same tasks.

[Permitir comparação com outras ferramentas de exploração]

5.1.Case Study 1: Discovering Technological Trends

The changes in the technological landscape that can be identified by analyzing published patents in different time periods are observed by answering four main questions:
• Which industry fields have increased the level of attention throughout given periods?
• Which industry fields have decreased the level of attention throughout given periods?
• Which industry fields started to be addressed throughout given periods?
• Which industry fields stopped to be addressed throughout given periods?

5.2.Case Study 2: Evaluating a scientific paper

Consider a reviewer evaluating a scientific paper. In order to do so, the user can take the following strategy:
1. Analyze the age of the citations: the reviewer extracts the years of each citation and calculates, for example, the mean year;
2. Check the lack of citations to relevant publications: The reviewer can extract the keywords of the paper and issue a keyword search for related papers; Rank the articles by the number of incoming citations. Keep the first 20 articles; Differentiate the two sets and verify which ones are not in the bibliography of the paper;
3. Analyze the degree of "self-citations": the reviewer analyzes how self-referential is the paper. A self-citation can be either a citation of previous works of one of the authors or citations from authors of the same research group;
4. Evaluate if the paper fits to the scope of a venue: the reviewer might count the number of citations published in the same venue as an indicator of how adequate the paper is to the targeted venue.

The main goal of the reuse in this case is the transference of knowledge not only with regards to the results of the tasks but also concerning the resolutionprocesses. Therefore, new users can draw upon the experience of previous users to aid their task resolution strategies.

[O próprio processo de exploração pode ser um conhecimento a ser compartilhado]

5.3. Case Study 3: Summarizing Gene Clusters

The task consists in crossing the gene identifiers of the cluster with a bibliographical dataset in order to find terms that better describe the genes in the cluster. The strategy employed is the following:
1. A gene can have many identifiers. Therefore, the user tries to obtain all identifiers for each gene in the cluster;
2. Once having a more complete set of identifiers, the user tries to query a bibliographic dataset in order to find all publications that mention the gene identifiers;
3. From publications, find the terms that better describe the genes;
4. Rank the terms using specific ranking criteria designed to extract different information from the cluster

• The framework is useful for formally describing relatively complex exploration tasks of different domains of knowledge;
• The framework leverages analyzes of the task resolution process abstracting interface and interaction details. Therefore, the framework can be used as an epistemic tool for design decisions, where the designers could use it to devise alternative exploration paths and analyze which sequences is mostly indicated given the task execution context and user profile.
• Once the solution strategy employed can be formally represented, they can be shared and reused either in different scenario within the same domain or within different domains. Reuse in different domains will require adaptations of the schema used in the task, which can be achieved by parameterizations. Moreover, by representing exploration strategies formally, it is possible to audit the strategies for results validation purposes, which is of a great value for validating scientific results, for example

[O framework define operações genéricas e os estudos de caso mostraram como aplicar essas operações para atingir um objetivo]

Artigo Exploring KG for Exploratory Search (2014) 

Interface para navegar entre os documentos e o grafo. Duas tarefas de exploração: uma simples e outra complexa. Conjunto de documentos pré-selecionado, resultado do OIE avaliado e corrigido por especialista (mas isso não é possível em tempo de consulta). A tarefa simples é mais semelhante a Q&A e a complexa é um cenário de busca sem resultado exato definido. 

Complexa: 10 minutos e Qual é o posicionamento do Estado da Califórnia sobre pesquisas com células tronco?

Simples:  Tópico Doença de Lyme e algumas perguntas pontuais.

Questionários sobre o conhecimento prévio do assunto e o que aprenderam com as buscas. E registro do log de navegação na interface para avaliar os padrões de acesso ao grafo versus documentos.

Será que consigo: 

  1. (Pesquisador) Selecionar um benchmark de perguntas e respostas que use KG hiper-relacional;
  2. (Pesquisador) Agrupar perguntas deste benchmark sobre o mesmo assunto ou entidade;
  3. (Pesquisador) Analisar o esquema/resumo KG hiper-relacional para identificar dimensões contextuais de interesse e possíveis regras a serem aplicadas;
  4. (Pesquisador) Formular uma necessidade de informação que envolva aprender ou investigar um assunto ou entidade (ou mais de uma relacionada) encontrada no benchmark e também considerar as dimensões de contexto disponíveis;
  5. (Pesquisador) Apresentar tarefas de busca exploratória e hiper KG contextualizado aos usuários;
  6. (Usuário) Realizar a exploração do KG e registrar as consultas (com o respectivo tipo de respostas);
  7. (Pesquisador) Elaborar um questionário pós-tarefa;
  8. (Usuário) Responder ao questionário para testar os conhecimentos adquiridos;
  9. (Pesquisador) Analisar padrões de consulta;
  10. (Pesquisador) Analisar as respostas dos usuários à pesquisa.




Comentários

Postagens mais visitadas deste blog

Connected Papers: Uma abordagem alternativa para revisão da literatura

Durante um projeto de pesquisa podemos encontrar um artigo que nos identificamos em termos de problema de pesquisa e também de solução. Então surge a vontade de saber como essa área de pesquisa se desenvolveu até chegar a esse ponto ou quais desdobramentos ocorreram a partir dessa solução proposta para identificar o estado da arte nesse tema. Podemos seguir duas abordagens:  realizar uma revisão sistemática usando palavras chaves que melhor caracterizam o tema em bibliotecas digitais de referência para encontrar artigos relacionados ou realizar snowballing ancorado nesse artigo que identificamos previamente, explorando os artigos citados (backward) ou os artigos que o citam (forward)  Mas a ferramenta Connected Papers propõe uma abordagem alternativa para essa busca. O problema inicial é dado um artigo de interesse, precisamos encontrar outros artigos relacionados de "certa forma". Find different methods and approaches to the same subject Track down the state of the art rese...

Knowledge Graph Embedding with Triple Context - Leitura de Abstract

  Jun Shi, Huan Gao, Guilin Qi, and Zhangquan Zhou. 2017. Knowledge Graph Embedding with Triple Context. In Proceedings of the 2017 ACM on Conference on Information and Knowledge Management (CIKM '17). Association for Computing Machinery, New York, NY, USA, 2299–2302. https://doi.org/10.1145/3132847.3133119 ABSTRACT Knowledge graph embedding, which aims to represent entities and relations in vector spaces, has shown outstanding performance on a few knowledge graph completion tasks. Most existing methods are based on the assumption that a knowledge graph is a set of separate triples, ignoring rich graph features, i.e., structural information in the graph. In this paper, we take advantages of structures in knowledge graphs, especially local structures around a triple, which we refer to as triple context. We then propose a Triple-Context-based knowledge Embedding model (TCE). For each triple, two kinds of structure information are considered as its context in the graph; one is the out...

KnOD 2021

Beyond Facts: Online Discourse and Knowledge Graphs A preface to the proceedings of the 1st International Workshop on Knowledge Graphs for Online Discourse Analysis (KnOD 2021, co-located with TheWebConf’21) https://ceur-ws.org/Vol-2877/preface.pdf https://knod2021.wordpress.com/   ABSTRACT Expressing opinions and interacting with others on the Web has led to the production of an abundance of online discourse data, such as claims and viewpoints on controversial topics, their sources and contexts . This data constitutes a valuable source of insights for studies into misinformation spread, bias reinforcement, echo chambers or political agenda setting. While knowledge graphs promise to provide the key to a Web of structured information, they are mainly focused on facts without keeping track of the diversity, connection or temporal evolution of online discourse data. As opposed to facts, claims are inherently more complex. Their interpretation strongly depends on the context and a vari...