Zero-shot object recognition or zero-shot learning aims to transfer the object recognition ability among the semantically related categories, such as fine-grained animal or bird species. However, the ...images of different fine-grained objects tend to merely exhibit subtle differences in appearance, which will severely deteriorate zero-shot object recognition. To reduce the superfluous information in the fine-grained objects, in this paper, we propose to learn the redundancy-free features for generalized zero-shot learning. We achieve our motivation by projecting the original visual features into a new (redundancy-free) feature space and then restricting the statistical dependence between these two feature spaces. Furthermore, we require the projected features to keep and even strengthen the category relationship in the redundancy-free feature space. In this way, we can remove the redundant information from the visual features without losing the discriminative information. We extensively evaluate the performance on four benchmark datasets. The results show that our redundancy-free feature based generalized zero-shot learning (RFF-GZSL) approach can outperform the state-of-the-arts often by a large margin.
In order to construct a unified permeability prediction model for multi-stage tight gas sandstones with permeability across 6 orders of magnitude and changeable porosity-permeability relationship, ...Bayesian regularization neural network is properly configured with core porosity, conventional logs and a few derivates of them as input items. With high accuracy and excellent generalization, it is promising to be stably and reliably popularized in the study area. The way of model construction, optimization and evaluation may provide underlying insights needed for permeability prediction of similar reservoirs and application of machine learning in reservoir evaluation.
•No consistent porosity-permeability relation exists in channel sandstones.•Universal permeability model is built with Bayesian regularization neural network.•Sufficient petrophysical-geological information is contained in logs.•Prior knowledges enhance interpretability, controllability and applicability.
Permeability can effectively represent the ability of fluids to flow in porous media, and is a crucial physical parameter for the exploration of unconventional oil reservoirs. In low-permeable porous ...media, such as tight rocks, not all pores function as fluid flow channels; meanwhile, the pore surface tends to form unevenly adsorbed water lay. To date, the two phenomena above have not been considered in previous studies on permeability estimation models. In this study, a new fractal permeability model with variable pore sizes was proposed resulting in two advancements: (1) unevenly adsorbed water lay attached to pore surfaces was considered; and (2) instead of the total pore space, the movable fluid space obtained from nuclear magnetic resonance (NMR) experiments was used as the fluid flow channels in the porous media. Helium permeability data of 26 mixed rock samples from tight oil reservoirs were used to establish the accuracy of the proposed permeability model. The results of a sensitivity analysis show that high porosity and large pore diameters play a positive role in generating high permeability in porous media. The permeability decreased with an increase in the fractal dimension of the total pore space and the tortuosity fractal dimension of the total pore space and movable fluid space, whereas it first decreased and then increased as the fractal dimension of the movable fluid space increased. Furthermore, the proposed model indicates that the movable fluid space and unevenly adsorbed water lay greatly influence the fluid flow, specifically at lower permeabilities (K<0.1 mD), while the influence of the unevenness of the adsorbed water lay on fluid flow is negligible at high permeabilities. Overall, the accuracy of the permeability estimation of the proposed model is reliable both in low-permeable and conventional porous media, and can be effectively applied to unconventional oil reservoirs.
•Movable fluid space and unevenly adsorbed water lay affect rock permeability.•High porosity and large pores play a positive role in generating high permeability.•Influence of hydrophilic minerals on the permeability cannot be ignored.•Proposed model helps to estimate the permeability of unconventional oil reservoirs.
The lithofacies characteristics of the Qingshankou Formation (K2qn) shale in the Gulong Depression are crucial for oil exploration and development. This study investigates the K2qn shale lithofacies ...characteristics and their impact on reservoir physical properties using scanning electron microscopy (SEM), high-pressure mercury injection (HPMI), and logging quantification. The results indicate that the main minerals in K2qn shale are quartz, plagioclase, and clay. The sedimentary structures are classified into three types: laminated, layered, and massive. The K2qn shale lithofacies can be categorized into 12 types based on a combination of lithology and sedimentary structure. The main types are laminated clayey shale, layered clayey shale, and layered felsic shale. The larger the average pore size of the K2qn lithofacies, the stronger the heterogeneity of pore size distribution in space and the better the pore-to-throat connectivity. The impact of K2qn shale lithofacies on reservoir physical properties is mainly due to differences in lithology, complemented by variations in the sedimentary structural model. Under certain diagenetic or tectonic conditions, a layered sedimentary structural model of lithofacies may not increase reservoir permeability. Generally, felsic and carbonate rocks in tidal flat environments promote the development of shale with high permeability and porosity, while lithofacies deposited in static water environments below the wave base in lake basins typically exhibit low permeability and porosity. The physical properties of a reservoir are primarily influenced by the differences in pore throat characteristics resulting from variations in lithology.
Generalized zero-shot learning (GZSL) aims to recognize objects from both seen and unseen classes, when only the labeled examples from seen classes are provided. Recent feature generation methods ...learn a generative model that can synthesize the missing visual features of unseen classes to mitigate the data-imbalance problem in GZSL. However, the original visual feature space is suboptimal for GZSL classification since it lacks discriminative information. To tackle this issue, we propose to integrate the generation model with the embedding model, yielding a hybrid GZSL framework. The hybrid GZSL approach maps both the real and the synthetic samples produced by the generation model into an embedding space, where we perform the final GZSL classification. Specifically, we propose a contrastive embedding (CE) for our hybrid GZSL framework. The proposed contrastive embedding can leverage not only the class-wise supervision but also the instance-wise supervision, where the latter is usually neglected by existing GZSL researches. We evaluate our proposed hybrid GZSL framework with contrastive embedding, named CE-GZSL, on five benchmark datasets. The results show that our CEGZSL method can outperform the state-of-the-arts by a significant margin on three datasets. Our codes are available on https://github.com/Hanzy1996/CE-GZSL.
Generalized zero-shot learning (GZSL) aims to recognize objects from both seen and unseen classes when only the labeled examples from seen classes are provided. Recent feature generation methods ...learn a generative model that can synthesize the missing visual features of unseen classes to mitigate the data-imbalance problem in GZSL. However, the original visual feature space is suboptimal for GZSL recognition since it lacks semantic information, which is vital for recognizing the unseen classes. To tackle this issue, we propose to integrate the feature generation model with an embedding model. Our GZSL framework maps both the real and the synthetic samples produced by the generation model into an embedding space, where we perform the final GZSL classification. Specifically, we propose a semantic contrastive embedding (SCE) for our GZSL framework. Our SCE consists of attribute-level contrastive embedding and class-level contrastive embedding. They aim to obtain the transferable and discriminative information, respectively, in the embedding space. We evaluate our GZSL method with semantic contrastive embedding, named SCE-GZSL, on four benchmark datasets. The results show that our SCE-GZSL method can achieve the state-of-the-art or the second-best on these datasets.
Block-diagonal representation (BDR) is an effective subspace clustering method. The existing BDR methods usually obtain a self-expression coefficient matrix from the original features by a shallow ...linear model. However, the underlying structure of real-world data is often nonlinear, thus those methods cannot faithfully reflect the intrinsic relationship among samples. To address this problem, we propose a novel latent BDR (LBDR) model to perform the subspace clustering on a nonlinear structure, which jointly learns an autoencoder and a BDR matrix. The autoencoder, which consists of a nonlinear encoder and a linear decoder, plays an important role to learn features from the nonlinear samples. Meanwhile, the learned features are used as a new dictionary for a linear model with block-diagonal regularization, which can ensure good performances for spectral clustering. Moreover, we theoretically prove that the learned features are located in the linear space, thus ensuring the effectiveness of the linear model using self-expression. Extensive experiments on various real-world datasets verify the superiority of our LBDR over the state-of-the-art subspace clustering approaches.
In zero-shot learning (ZSL), attribute knowledge plays a vital role in transferring knowledge from seen classes to unseen classes. However, most existing ZSL methods learn biased attribute knowledge, ...which usually results in biased attribute prediction and a decline in zero-shot recognition performance. To solve this problem and learn unbiased attribute knowledge, we propose a visual attribute Transformer for zero-shot recognition (ZS-VAT), which is an effective and interpretable Transformer designed specifically for ZSL. In ZS-VAT, we design an attribute-head self-attention (AHSA) that is capable of learning unbiased attribute knowledge. Specifically, each attribute head in AHSA first transforms the local features into attribute-reinforced features and then accumulates the attribute knowledge from all corresponding reinforced features, reducing the mutual influence between attributes and avoiding information loss. AHSA finally preserves unbiased attribute knowledge through attribute embeddings. We also propose an attribute fusion model (AFM) that learns to recover the correct category knowledge from the attribute knowledge. In particular, AFM takes all features from AHSA as input and generates global embeddings. We carried out experiments to demonstrate that the attribute knowledge from AHSA and the category knowledge from AFM are able to assist each other. During the final semantic prediction, we combine the attribute embedding prediction (AEP) and global embedding prediction (GEP). We evaluated the proposed scheme on three benchmark datasets. ZS-VAT outperformed the state-of-the-art generalized ZSL (GZSL) methods on two datasets and achieved competitive results on the other dataset.
Generalized zero-shot learning suffers from an extreme data imbalance problem, that is, the training data only come from seen classes while no unseen class data are available. Recently, a number of ...feature generation methods based on generative adversarial networks (GAN) have been proposed to address this problem. Existing feature generation methods, however, have never considered the under-constrained problem, and thus could generate an unrestricted visual feature corresponding to no meaningful object class. In this paper, we propose to equip the feature generation framework with a parallel inference network that projects visual feature to the semantic descriptor space, constraining to avoid the generation of unrestricted visual features. The two-parallel-stream framework (1) enables our method, termed inference guided feature generation (Inf-FG), to mitigate the under-constrained problem and (2) makes our Inf-FG applicable to transductive ZSL. Our Inf-FG learns the feature generator and the inference network simultaneously by aligning the joint distribution of visual features and semantic descriptors from the feature generator and the joint distribution from the inference network. We evaluate our approach on four benchmark ZSL datasets, including AWA, CUB, SUN, and FLO, on which our method improves our baselines on generalized zero-shot learning.
Zero-shot object recognition or zero-shot learning aims to transfer the object recognition ability among the semantically related categories, such as fine-grained animal or bird species. However, the ...images of different fine-grained objects tend to merely exhibit subtle differences in appearance, which will severely deteriorate zero-shot object recognition. To reduce the superfluous information in the fine-grained objects, in this paper, we propose to learn the redundancy-free features for generalized zero-shot learning. We achieve our motivation by projecting the original visual features into a new (redundancy-free) feature space and then restricting the statistical dependence between these two feature spaces. Furthermore, we require the projected features to keep and even strengthen the category relationship in the redundancy-free feature space. In this way, we can remove the redundant information from the visual features without losing the discriminative information. We extensively evaluate the performance on four benchmark datasets. The results show that our redundancy-free feature based generalized zero-shot learning (RFF-GZSL) approach can achieve competitive results compared with the state-of-the-arts.