| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 9.23 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A importância das florestas a nível mundial está bem estabelecida na sociedade, sabendo-se que a desflorestação tem sérias implicações na biodiversidade, nas sociedades rurais dependentes das florestas para comida e sustento, e nas emissões de gases de efeito de estufa que responsáveis pelas alterações climáticas. Mecanismos como o REDD+, desenvolvidos pela UNFCCC, ajudam países em desenvolvimento através de recompensas por evitarem a desflorestação e degradação florestal enquanto promovem a conservação das florestas, desenvolvimento sustentável e melhorias no sequestro de carbono nas florestas. No entanto, para estes mecanismos serem eficazes, é necessário monitorizar florestas de forma eficiente. Este trabalho contribui para os domínios científicos de ciências da computação e de ciências ambientais ao aplicar aprendizagem automática evolutiva neste tópicos, lidando com dois aspetos importantes da monitorização florestal: o mapeamento de tipos de coberto e a deteção de degradação florestal. Este trabalho também contribui para a área de aprendizagem automática através do estudo de métodos para engenharia de atributos que criem modelos robustos e interpretáveis que possam ser lidos, percebidos e corrigidos pelos especialistas das respetivas áreas científicas. Várias aplicações nossas conseguiram aumentar a robustez e transferibilidade de modelos de aprendizagem automática, melhorando o grau de confiança dos modelos de monitorização florestal. Estes modelos podem ser validados estudando se os atributos usados fazem sentido no contexto do problema, e também reduzindo a dimensionalidade dos conjuntos de dados ao ponto de se conseguir visualmente detetar problemas nos dados. Também estudamos a relação entre a complexidade de modelos com a sua interpretabilidade. Interpretabilidade e explicabilidade de modelos têm vindo a tornar-se tópicos populares nos últimos anos. Isto motiva trabalho futuro no uso de aprendizagem automática evolucionaria para forçar a evolução de modelos que sejam mais facilmente interpretáveis através de funções de fitness específicas.
The importance of forests worldwide is well established throughout society, and it is well known that deforestation has serious implications for biodiversity, for rural communities dependent on forests for food and income, and for greenhouse gas emissions driving the global climate. Mechanisms such as REDD+, developed by the UNFCCC, help developing countries by rewarding them for avoiding deforestation and forest degradation while promoting forest conservation, sustainable management and enhancement of forest carbon stocking. However, for those mechanisms to be effective, they require the ability to monitor forest development efficiently. This work contributes to the scientific fields of computer science and environmental science by applying evolutionary machine learning to tackle this issue, addressing two important objectives in forest monitoring: mapping land cover changes and detecting forest degradation. It also contributes to the machine learning field on topics such as automatic feature construction methods for creating robust and interpretable models that can be read, understood and corrected by the experts of the application fields. Several of our applications successfully increase the robustness and transferability of machine learning models, improving the reliability of forest monitoring models. The induced models can be validated by identifying whether the features being used make sense in the context of the problem, and also by reducing the dimensionality of the datasets and visually identifying issues with the data. We also study the relationship between the complexity of feature engineering models and their interpretability. Model interpretability and explainability have become hot topics over the last few years. This motivates future studies on using evolutionary machine learning to force the evolution of more easily interpretable models via specific fitness functions.
The importance of forests worldwide is well established throughout society, and it is well known that deforestation has serious implications for biodiversity, for rural communities dependent on forests for food and income, and for greenhouse gas emissions driving the global climate. Mechanisms such as REDD+, developed by the UNFCCC, help developing countries by rewarding them for avoiding deforestation and forest degradation while promoting forest conservation, sustainable management and enhancement of forest carbon stocking. However, for those mechanisms to be effective, they require the ability to monitor forest development efficiently. This work contributes to the scientific fields of computer science and environmental science by applying evolutionary machine learning to tackle this issue, addressing two important objectives in forest monitoring: mapping land cover changes and detecting forest degradation. It also contributes to the machine learning field on topics such as automatic feature construction methods for creating robust and interpretable models that can be read, understood and corrected by the experts of the application fields. Several of our applications successfully increase the robustness and transferability of machine learning models, improving the reliability of forest monitoring models. The induced models can be validated by identifying whether the features being used make sense in the context of the problem, and also by reducing the dimensionality of the datasets and visually identifying issues with the data. We also study the relationship between the complexity of feature engineering models and their interpretability. Model interpretability and explainability have become hot topics over the last few years. This motivates future studies on using evolutionary machine learning to force the evolution of more easily interpretable models via specific fitness functions.
Description
Keywords
Computação Evolutiva Aprendizagem Automática Interpretabilidade Ciências Ambientais Deteção Remota Evolutionary Computation Machine Learning Interpretability Environmental Science Remote Sensing