GeoAI: A Model-Agnostic Meta-Ensemble Zero-Shot Learning Method for Hyperspectral Image Analysis and Classification

Abstract

Deep learning architectures are the most effective methods for analyzing and classifying Ultra-Spectral Images (USI). However, effective training of a Deep Learning (DL) gradient classifier aiming to achieve high classification accuracy, is extremely costly and time-consuming. It requires huge datasets with hundreds or thousands of labeled specimens from expert scientists. This research exploits the MAML++ algorithm in order to introduce the Model-Agnostic Meta-Ensemble Zero-shot Learning (MAME-ZsL) approach. The MAME-ZsL overcomes the above difficulties, and it can be used as a powerful model to perform Hyperspectral Image Analysis (HIA). It is a novel optimization-based Meta-Ensemble Learning architecture, following a Zero-shot Learning (ZsL) prototype. To the best of our knowledge it is introduced to the literature for the first time. It facilitates learning of specialized techniques for the extraction of user-mediated representations, in complex Deep Learning architectures. Moreover, it leverages the use of first and second-order derivatives as pre-training methods. It enhances learning of features which do not cause issues of exploding or diminishing gradients; thus, it avoids potential overfitting. Moreover, it significantly reduces computational cost and training time, and it offers an improved training stability, high generalization performance and remarkable classification accuracy. Algorithms 2020, 13, 61 2 of 25 Algorithms 2020, 13, 61 3 of 25In contrast to the learning process of the traditional neural networks, it is not sufficient for the network to learn good representations of the training classes, as the testing classes are distinct and they are not presented in training. However, it is desirable to learn features which distinguish the existing classes.The evaluation process consists of two distinct stages of the following format [4]:Step 1: Given k examples (value of k-shot), if k = 1, then the process is called one-shot; if k = 5, five-shot, and so on. The parameter k represents the number of labeled samples given to the algorithm by each class. By considering these samples, which comprise the support set, the network is required to classify and eventually adapt to existing classes.Step 2: Unknown examples of the labeled classes are presented randomly, unlike the ones presented in the previous step, which the network is called to correctly classify. The set of examples in this stage is known as the query set.The above procedure (steps) is repeated many times using random classes and examples which are sampled from the testing-evaluation set.As it is immediately apparent from the description of the evaluation process, as the number of classes increases, the task becomes more difficult, because the network has to decide between several alternatives. This means that Zero-shot Learning [5] is clearly more difficult than the one-shot, which is more difficult than the five-shot, and so on. Although humans have the ability to cope with this process, traditional ANN require many more examples to genera...

Attribute-Modulated Generative Meta Learning for Zero-Shot Classification

“…Recently, meta-model is proposed to further eliminate the bias towards seen classes in ZSL [5,27,28,37,39]. For example, TAFE-NET [45] uses a meta learner for task-aware feature embedding.…”

“…For example, TAFE-NET [45] uses a meta learner for task-aware feature embedding. ZSML [39] and GeoAI [5] are closely related to our work. They both incorporate MAML [9] into generative model for ZSL.…”

“…Some recent work uses generative methods to synthesize visual features conditioned on semantic side information and learn a conventional supervised classifier from generated samples to overcome the above issue [3,8,11,52,59]. Given that generative models learned from seen classes exhibits inherent biases when generalizing to unseen classes, meta generative approaches for ZSL emerge as a new trend to mitigate the biases [5,27,28,37,39]. Meta generative approaches incorporate meta-learning models, e.g., Model-Agnostic Meta-Learning (MAML) [9], into generative models.…”

Attribute-Modulated Generative Meta Learning for Zero-Shot Learning

“…To address this issue, some work (Qin et al 2019;Pal and Balasubramanian 2019;Soh, Cho, and Cho 2020;Demertzis and Iliadis 2020;Nooralahzadeh et al 2020;Verma, Brahma, and Rai 2020) introduces meta-learning, e.g., model-agnostic meta-learning framework (Finn, Abbeel, and Levine 2017), into ZSL, namely meta-ZSL. Meta-ZSL splits existing training classes into two disjoint sets, namely support and query sets, to mimic seen and unseen classes.…”

“…However, current meta-ZSL approaches (Verma, Brahma, and Rai 2020;Demertzis and Iliadis 2020;Soh, Cho, and Cho 2020) directly integrate meta-learning and ZSL without considering the limitations posed by diverse data distributions in ZSL. Thus, the learned models may be misguided towards extremely different distributions.…”