Self-organized inductive learning in a multidimensional graph-like neural network framework
Human cognition heavily relies on inductive learning, a process that the field of machine learning aims to replicate in artificial hardware/software. While connectionist learning methods have yielded great pragmatic results in this area, they still lack a model hierarchy of learning to explain their...
| Autor principal: | Schramm, Ana Carolina Melik |
|---|---|
| Outros Autores: | http://lattes.cnpq.br/6334894528699558 |
| Grau: | Dissertação |
| Idioma: | eng |
| Publicado em: |
Universidade Federal do Amazonas
2022
|
| Assuntos: | |
| Acesso em linha: |
https://tede.ufam.edu.br/handle/tede/9028 |
| id |
oai:https:--tede.ufam.edu.br-handle-:tede-9028 |
|---|---|
| recordtype |
dspace |
| spelling |
oai:https:--tede.ufam.edu.br-handle-:tede-90282022-08-24T05:03:33Z Self-organized inductive learning in a multidimensional graph-like neural network framework Schramm, Ana Carolina Melik Mota, Edjard Souza http://lattes.cnpq.br/6334894528699558 http://lattes.cnpq.br/0757666181169076 Fonseca, Paulo Cesar http://lattes.cnpq.br/3639575844521754 Cristo, Marco Antônio Pinheiro de http://lattes.cnpq.br/6261175351521953 Inteligência artificial Aprendizado de máquina Redes neurais (Computação) CIENCIAS EXATAS E DA TERRA: CIENCIA DA COMPUTACAO Artificial intelligence Neural-symbolic integration Inductive clausal learning Multidimensional graph neural network Human cognition heavily relies on inductive learning, a process that the field of machine learning aims to replicate in artificial hardware/software. While connectionist learning methods have yielded great pragmatic results in this area, they still lack a model hierarchy of learning to explain their results. NeSy computing seeks to develop effective integration between connectionist and symbolic learning. As an effort to achieve this integration, NeMuS is a multi-dimensional graph structure, originally conceived with four spaces of codified elements of first-order logic, that learns patterns of refutation and performs inductive clausal reasoning to induce hypotheses that explain examples non-previously specified in a background knowledge. Recently, there was an experiment in which a connected background knowledge was trained using SOM to generate similarity of concepts according to their attributes, and their respective position within the concepts. In this experiment, induction was performed by human analysis on the organizational map of concepts. In this work, we seek a suitable method to generate neighbourhood patterns to be used for inductive learning and reasoning in order to reduce the search space of hypotheses. Additionally, we define a language bias able to handle predicate invention, to guide the process of generating such hypotheses. Cognição humana depende fortemente de aprendizagem indutiva, um processo que o campo de aprendizagem de máquina busca replicar em hardware/software artificial. Enquanto métodos de aprendizagem coneccionista resultaram em grandes resultados pragmáticos na área, ainda lhes falta uma hierarquia modelo de aprendizagem para explicar seus resultados. Computação NeSy busca desenvolver uma integração efetiva entre aprendizagem coneccionista e simbólica. Como um esforço para alcançar essa integração, NeMuS é uma estrutura de grafo multidimensional, originalmente concebida com quatro espaços de elementos codificados de lógica de primeira ordem, que aprende padrões de refutação e executa raciocínio indutivo clausal para induzir hipóteses que explicam exemplos não previamente especificados em uma base de conhecimento. Recentemente, houve um experimento em que uma base de conhecimento conectada foi treinada usando SOM para gerar similaridade de conceitos de acordo com seus atributos, e suas respectivas posições dentro dos conceitos. Nesse experimento, indução foi feita por análise humana do mapa organizacional de conceitos. Neste trabalho, nós buscamos um método adequado de gerar padrões de vizinhança para serem usados em aprendizagem e raciocínio indutivo para reduzir o espaço de busca de hipóteses. Adicionalmente, nós definimos um viés de linguagem capaz de lidar com invenção de predicados, para guiar o processo de gerar tais hipóteses. FAPEAM - Fundação de Amparo à Pesquisa do Estado do Amazonas CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior 2022-08-23T20:20:07Z 2022-03-21 Dissertação SCHRAMM, Ana Carolina Melik. Self-organized inductive learning in a multidimensional graph-like neural network framework. 2022. 59 f. Dissertação (Mestrado em Informática) - Universidade Federal do Amazonas, Manaus (AM), 2022. https://tede.ufam.edu.br/handle/tede/9028 eng Acesso Aberto application/pdf Universidade Federal do Amazonas Instituto de Computação Brasil UFAM Programa de Pós-graduação em Informática |
| institution |
TEDE - Universidade Federal do Amazonas |
| collection |
TEDE-UFAM |
| language |
eng |
| topic |
Inteligência artificial Aprendizado de máquina Redes neurais (Computação) CIENCIAS EXATAS E DA TERRA: CIENCIA DA COMPUTACAO Artificial intelligence Neural-symbolic integration Inductive clausal learning Multidimensional graph neural network |
| spellingShingle |
Inteligência artificial Aprendizado de máquina Redes neurais (Computação) CIENCIAS EXATAS E DA TERRA: CIENCIA DA COMPUTACAO Artificial intelligence Neural-symbolic integration Inductive clausal learning Multidimensional graph neural network Schramm, Ana Carolina Melik Self-organized inductive learning in a multidimensional graph-like neural network framework |
| topic_facet |
Inteligência artificial Aprendizado de máquina Redes neurais (Computação) CIENCIAS EXATAS E DA TERRA: CIENCIA DA COMPUTACAO Artificial intelligence Neural-symbolic integration Inductive clausal learning Multidimensional graph neural network |
| description |
Human cognition heavily relies on inductive learning, a process that the field of machine learning aims to replicate in artificial hardware/software. While connectionist learning methods have yielded great pragmatic results in this area, they still lack a model hierarchy of learning to explain their results. NeSy computing seeks to develop effective integration between connectionist and symbolic learning. As an effort to achieve this integration, NeMuS is a multi-dimensional graph structure, originally conceived with four spaces of codified elements of first-order logic, that learns patterns of refutation and performs inductive clausal reasoning to induce hypotheses that explain examples non-previously specified in a background knowledge. Recently, there was an experiment in which a connected background knowledge was trained using SOM to generate similarity of concepts according to their attributes, and their respective position within the concepts. In this experiment, induction was performed by human analysis on the organizational map of concepts. In this work, we seek a suitable method to generate neighbourhood patterns to be used for inductive learning and reasoning in order to reduce the search space of hypotheses. Additionally, we define a language bias able to handle predicate invention, to guide the process of generating such hypotheses. |
| author_additional |
Mota, Edjard Souza |
| author_additionalStr |
Mota, Edjard Souza |
| format |
Dissertação |
| author |
Schramm, Ana Carolina Melik |
| author2 |
http://lattes.cnpq.br/6334894528699558 |
| author2Str |
http://lattes.cnpq.br/6334894528699558 |
| title |
Self-organized inductive learning in a multidimensional graph-like neural network framework |
| title_short |
Self-organized inductive learning in a multidimensional graph-like neural network framework |
| title_full |
Self-organized inductive learning in a multidimensional graph-like neural network framework |
| title_fullStr |
Self-organized inductive learning in a multidimensional graph-like neural network framework |
| title_full_unstemmed |
Self-organized inductive learning in a multidimensional graph-like neural network framework |
| title_sort |
self-organized inductive learning in a multidimensional graph-like neural network framework |
| publisher |
Universidade Federal do Amazonas |
| publishDate |
2022 |
| url |
https://tede.ufam.edu.br/handle/tede/9028 |
| _version_ |
1781302724147544064 |
| score |
11.653393 |