Short Paper SBBD 2022 - Feedback dos Revisores

===== Review (1) =====

*** Relevance to SBBD  = Accept (4)
*** Originality = Accept (4)
*** Technical quality = Accept (4)
*** Impact of ideas and results  = Weak Accept (3)
*** Presentation = Weak Accept (3)
*** Overall Evaluation = Weak Accept (minor revisions) (3)

*** Main contributions (up to one paragraph) (): This paper presents an approach to create and explore Contextual Knowledge Graphs. The proposed approach is divided into two phases: KG Engineering and KG Exploratory Search. The first phase extends a hyper relational KG H into a Contextual Hyper KG H by explicitly representing context dimensions and semantic interpretation. The second phase retrieves from H possible answers for a query Q. Besides, the paper proposes a contextual hyper KG definition as an extension of hyper-relational KG.

[Não são respostas possíveis, é a melhor resposta dependendo da completude do KG e da query]

*** Positive aspects (please number then P1, P2,...) (): P1. The paper investigates an interesting and applicable topic. P2. The paper is well written, well-structured, with surprisingly few syntactic inaccuracies and the ideas are presented in a reasonably clear way. P3. It discusses the related issues in a precise, comprehensive, and focused way. P4. Definitions are presented in a formal way.

*** Negative aspects (Please number then N1, N2,...) (): N1. No experiments or case studies were presented. N2. At the end of the abstract an introduction, if we ask ourselves about the contribution of
this paper, this remains unclear.

[Um estudo de caso mais detalhado seria possível mas o espaço de Short Paper limita as escolhas, ou apresento a formalização ou o caso. Experimentos estão previstos mais não projetados uma vez que ainda não consegui definir se vai ser qualitativo, quantitativo ou ambos, poderia ter sinalizado em trabalhos futuros na última seção. Na última seção estão elencadas as duas contribuições: a extensão do KG e o conceito de melhor resposta.]

*** Detailed comments (please number them) (): In my opinion, the paper needs to improve on some aspects: -  The abstract/introduction does not clearly introduce the objectives of the work. - The title of some sections are not insightful (Main Concepts, Proposal).

Clarity issues:

Inaccurate information. - log searches. What are log searches?

Avoid repeating the same word in the same sentence or in close sentences. - An information need arises from the perception of a KNOWLEDGE gap when an agent, human or machine, has insufficient KNOWLEDGE - Therefore, it aims to understand the necessary and sufficient conditions for information to be considered KNOWLEDGE. It depends on the beliefs, feelings and prior KNOWLEDGE - Motivated by an INFORMATION need, a human triggers a conscious effort to acquire INFORMATION through

Avoid repeating the same phase. - Information is potential knowledge. It becomes knowledge when used to solve a problem. - Motivated by an information need, a human triggers a conscious effort to acquire information through exploratory search.

Minor problems. - Use Figure 1 instead of figure 1 - Use Table 1 instead of table 1 - for exploratory search and KG Summarization helps to - Our contribution IS a step forward

[Minor Ajustado]

===== Review (2) =====

*** Relevance to SBBD = Weak Accept (3)
*** Originality = Accept (4)
*** Technical quality = Weak Reject (2)
*** Impact of ideas and results  = Weak Accept (3)
*** Presentation  = Weak Reject (2)
*** Overall Evaluation = Weak Reject (major revisions) (2)

*** Main contributions (up to one paragraph) (): Initial ideas on a method enabling the contextual querying of knowledge graph KG). The querying is based on a KG engineering phase, in which context is added. The querying phase based on the contextually enriched KG enables exploratory search.

*** Positive aspects (please number then P1, P2,...) (): - KG is an important trend, and context-aware querying is an up-to-date and important problem - related work is relevant - the steps of a two-phased
method are outlined, with some formalization and definitions

*** Negative aspects (Please number then N1, N2,...) (): N1 the goal and the contributions are not clearly defined. N2, although there is a related work section and a comparison, it is not clear to understand the main differences between the proposed approach and existing works, and how it advances the state of the art N3 although there is some formalization, it is not enough to understand how sound the approach is. Obviously, this is also affected by space limitations, but concepts
should be clear. N4 it is not clear if it is a vision paper or a work-in-progress paper. The conclusions do not provide future work.

["This ongoing research aims to prepare and manipulate Knowledge Graphs (KG) to support exploratory search." .... foi a única "dica" que acabei dando para o estágio da pesquisa. Poderia ter incluído os próximos passos]

*** Detailed comments (please number them) (): C1 - make clear the paper's goal, the contributions, and how it must evolve. C2 - formalizations should add to the paper, not make it more confusing C3 -
comparison with related work should enable to understand how the proposal advances the state-of-the-art C4 - make clear how the work will evolve with additional/future work

[C2 é justamente o Ponto positivo 4 do revisor anterior.]

===== Review (3) =====

*** Relevance to SBBD = Accept (4)
*** Originality = Weak Reject (2)
*** Technical quality = Weak Reject (2)
*** Impact of ideas and results = Weak Accept (3)
*** Presentation  = Weak Accept (3)
*** Overall Evaluation  = Weak Reject (major revisions) (2)

*** Main contributions (up to one paragraph) (): The main contribution seems to be the knowledge engineering phase, which was not the focus of the paper.

[Não. As contribuições estão elencadas na última seção.]

*** Positive aspects (please number then P1, P2,...) (): - P1: Querying knowledge bases and querying with constraints are important problems. - P2: The paper is well written, in general.

*** Negative aspects (Please number then N1, N2,...) (): - N1: The proposal is not exploratory - N2: No prototype, proof of concept, or formal demonstrations - N3: The introduction has a lot of subjective
discussions that are unnecessary and imprecise. Information need is an objective concept for the database community. The authors should just build from there and avoid philosophical discussions that do not contribute to the objective.

[A definição de necessidade de informação vem de Ciência da Informação e não de Banco de Dados.]

*** Detailed comments (please number them) (): The authors have an interesting topic, but the focus needs to be adjusted. The emphasis in the title and the text is on the exploratory aspect of the proposal, but there is no exploratory component in the query workflow. I might be missing something, but none of the descriptions, figures, and examples suggest any exploratory procedure.

Being exploratory assumes that there is a human in the loop of the knowledge discovery process. What the authors present is a system that produces an answer based on a query. 

[Sim, o ser humano está nas duas etapas do processo. O usuário é que o busca. O engenheiro é o que direciona a construção do KG e a extensão contextual do mesmo. Mas o foco não é interface de sistemas para modelar a interação do usuário.]

There is no suggestion presented to the user, no opportunity to build upon previous results. 

[Pensei em incluir algo quando houver a Não Resposta, uma sugestão para a próxima consulta do usuário. Talvez com embeddings de grafo ou flexibilizando outros elementos da consulta e mantendo o contexto .... ]

In the related work section, the authors highlight the differences to other proposals, which are precisely the absence of the exploratory components that the others include, such as suggestions for query expansion and hierarchical summarization. This is equivalent to saying that a bicycle is a car but better because it does not have extra wheels or a roof. They cannot be compared because one lacks the elements that would make them comparable.

The only part of the proposal that seems to have an exploratory component is in the knowledge engineering phase, which was not detailed in the paper and is not a querying task.

Therefore, the authors should either add exploratory components to the proposal or present the solution as general querying under constraints and/or fuzzy querying.

It would be also important to have a prototype or provide details about how such system would be implemented.

The text is in general well written, but the first paragraphs are very subjective and imprecise. The authors draw conclusions that are not supported by the premises. For example:

- "Motivated by an information need, a human triggers a conscious effort to acquire information through exploratory search." No, there are many ways to acquire knowledge. This is just one path.

- "Therefore, it aims to understand the necessary and sufficient conditions for information to be considered knowledge." No, this is just one aspect of Epistemology (which does not need to be mentioned in the paper anyway).

In general, the authors are dealing with an interesting problem. But to have a real contribution, it is important to adjust the focus of the research.

============================================================

*** Relevance to SBBD           = 2 x Accept (4) + 1 x Weak Accept (3) = 11/15
*** Originality                 = 2 x Accept (4) + 1 x Weak Reject (2) = 10/15
*** Technical quality           = 1 x Accept (4) + 2 x Weak Reject (2) = 8/15
*** Impact of ideas and results = 3 x Weak Accept (3)                  = 9/15
*** Presentation                = 2 x Weak Accept (3) + 1 x Weak Reject (2) = 8/15
*** Overall Evaluation          = 1 x Weak Accept (minor revisions) (3) + 2 x Weak Reject (major revisions) (3) = 7/15