Pesquisa de Doutorado da Veronica

Vídeo -> https://youtu.be/XKmnAmhEkGs NLIDBs = Natural Language Interfaces para Banco de Dados > Keyword Search over RDF  > Já tem uso do BERT para essas abordagens mas não viu sobre GPT-3 ainda > Codd em 74 já estimulava > Como entender a necessidade de informação do usuário?  Natural Language Understanding (NLU) é IA Hard. Semiótica. Tem um survey no VLDBJ'19 Sistemas Centrados em Dados (SCD)  como os de consulta por palavras-chaves em BD Relacional.  Usam mais regras de mapeamento e menos dependente do banco de dados e mais dependentes das variações de consulta ATHENA (SCD) usa ontologias de domínio para permitir consultas em linguagem natural NaLIR usou o BD do MAG para avaliação das consultas SQL geradas. Quais consultas?   Em python: https://github.com/pr3martins/nalir-sbbd  > Faz mapeamento das palavras em nós do esquema em grafo > Usa Entity Resolution Templar usa o log de consultas para reduzir a interação do usuário (aproveita o conhecimento) Sistemas

Fonte: https://towardsdatascience.com/introduction-to-knowledge-graph-embedding-with-dgl-ke-77ace6fb60ef Amazon recently launched DGL-KE, a software package that simplifies this process with simple command-line scripts. With DGL-KE , users can generate embeddings for very large graphs 2–5x faster than competing techniques. DGL-KE provides users the flexibility to select models used to generate embeddings and optimize performance by configuring hardware, data sampling parameters, and the loss function. To use this package effectively, however, it is important to understand how embeddings work and the optimizations available to compute them. This two-part blog series is designed to provide this information and get you ready to start taking advantage of DGL-KE . Finally, another class of graphs that is especially important for knowledge graphs are multigraphs . These are graphs that can have multiple (directed) edges between the same pair of nodes and can also contain loops. The

Reification is mechanism for adding properties to RDF graph edges, thus making them directly translatable to property graphs. Reification: break down any structured data into triples, without loss of information. N-ary relations Reificação em RDF para suportar atributos nas arestas como por exemplo dados de proveniência. Uso de nós brancos. Reificação: incluir outras propriedades para a tripla rdf:Statement rdf:subject rdf:predicate rdf:object Reificação também se aplica a LPG para relações n-árias assim como em RDF. Reduzir relacionamentos n-ários a n relacionamentos binários (coleção, reificação). The RDF* and SPARQL* Approach to Annotate Statements in RDF: Although this is possible, up to now there has not been one standard, agreed upon way to do this. RDF* is a proposal on how to do this, introduced in 2014, which is getting traction in the RDF world. Relação ternária ============ Professor X ministra a disciplina Y para a turma Z 1 X - é um -> Professor 2 Y - é um -> Disci

Fonte -> https://iamsiva11.github.io/graph-embeddings-2017-part1/ Fonte -> https://iamsiva11.github.io/graph-embeddings-2017-part2/ Network Representation Learning (aka. graph embeddings) 2017: state of art graph embedding techniques (approaches like random walk based , deep learning based , etc) decomposed machine learning into three components: (1) Representation, (2) Evaluation, and (3) Optimization Representation is basically representing the input data (be image, speech, or video for that matter) in a way the learning algorithm can easily process. Representation Learning is using learning algorithms to derive good features or representation automatically, instead of traditional hand-engineered features.  representation learning (aka. feature engineering/learning) deep learning ... Multilayer neural networks can be used to perform feature learning, since they learn a representation of increasing complexity/abstraction of their input at the hidden layer(s) which is sub

Neves, A. B., Leme, L. A. P. P., Izquierdo, Y. T., & Casanova, M. A. (2021). Automatic Construction of Benchmarks for RDF Keyword Search Systems Evaluation. In Proceedings of the 23rd International Conference on Enterprise Information Systems (pp. 126–137). Vídeo ->  https://youtu.be/JZd5DS2hW-c Abstract: Keyword search systems provide users with a friendly alternative to access Resource Description Framework (RDF) datasets. The evaluation of such systems requires adequate benchmarks, consisting of RDF datasets and keyword queries, with their correct answers. However, the sets of correct answers such benchmarks provide for each query are often incomplete, mostly because they are manually built with experts’ help. The central contribution of this paper is an offline method that helps build RDF keyword search benchmarks automatically, leading to more complete sets of correct answers, called solution generators . The paper focuses on computing sets of generators and describes heu