Multiple Embedding Research Articles

Ensemble learning via feature selection and multiple transformed subsets: Application to image classification

In the machine learning field, especially in classification tasks, the model’s design and construction are very important. Constructing the model via a limited set of features may sometimes bound the classification performance and lead to non-optimal performances that some algorithms can provide. To this end, Ensemble learning methods were proposed in the literature. These methods’ main goal is to learn a set of models that provide features or predictions whose joint use could lead to a performance better than that obtained by the single model. In this paper, we propose three variants of a new efficient ensemble learning approach that was able to enhance the classification performance of a linear discriminant embedding method. As a case study we consider the efficient “Inter-class sparsity discriminative least square regression” method. We seek the estimation of an enhanced data representation. Instead of deploying multiple classifiers on top of the transformed features, we target the estimation of multiple extracted feature subsets obtained by multiple learned linear embeddings. These are associated with subsets of ranked original features. Multiple feature subsets were used for estimating the transformations. The derived extracted feature subsets were concatenated to form a single data representation vector that is used in the classification process. Many factors were studied and investigated in this paper including (Parameter combinations, number of models, different training percentages, feature selection methods combinations, etc.). Our proposed approach has been benchmarked on different image datasets of various sizes and types (faces, objects and scenes). The proposed scheme achieved competitive performance on four face image datasets (Extended Yale B, LFW-a, Gorgia and FEI) as well as on the COIL20 object dataset and the Outdoor Scene dataset. We measured the performance of our proposed schemes in comparison to (the single model ICS_DLSR, RDA_GD, RSLDA, PCE, LDE, LDA, SVM as well as the KNN algorithm) The conducted experiments showed that the proposed approach can enhance the classification performance in an efficient manner compared to the single-model based learning and was able to outperform its competing methods.

Food recommendation with graph convolutional network

Food recommendation has attracted increasing attentions to various food-related applications and services. The food recommender models aim to match users’ preferences with recipes, where the key lies in the representation learning of users and recipes. However, ranging from early content-based filtering and collaborative filtering methods to recent hybrid methods, the existing work overlooks the various food-related relations, especially the ingredient-ingredient relations, leading to incomprehensive representations. To bridge this gap, we propose a novel model Food recommendation with Graph Convolutional Network (FGCN), which exploits ingredient-ingredient, ingredient-recipe, and recipe-user relations deeply. FGCN employs the information propagation mechanism and adopts multiple embedding propagation layers to model high-order connectivity across different food-related relations and enhance the representations. Specifically, we develop three types of information propagation: (1) ingredient-ingredient information propagation, (2) ingredient-recipe information propagation, and (3) recipe-user information propagation. To validate the effectiveness and rationality of FGCN, we conduct extensive experiments on a real-world dataset. The results show that the proposed FGCN outperforms the state-of-the-art baselines. Further in-depth analyses reveal that FGCN could alleviate the sparsity issue in food recommendation.

Exploring graph capsual network for graph classification

Graph Neural Network (GNN) has received tremendous attention due to their power in learning graph representations by modeling the topological structure and aggregating feature information. However, the scalar node representations learned from GNN may not be sufficient to effectively preserve the attributes of the node/graph features, resulting in sub-optimal graph representation. Repeated averaging gathers too much noise, which makes the features of nodes in different classes over-mixed and leads to the problem of over smoothing. Inspired by the concept of capsule network proposed by Hinton, we propose a new framework for graph classification, named CapsualGNN, which takes full advantage of Graph Neural Network and Capsule Network. Specifically, we firstly represent nodes as groups of capsules, in which each capsule extracts distinctive features of its corresponding node. Then, we exploit routing mechanism to capture important information and properties at the graph level by the generated multiple embeddings for each graph, and utilize attention mechanism to focus on important features. Finally, to solve the problem of over smoothing, we introduce class residual connection for GCN. In addition, we also introduce parameter for distinguishing self-connected nodes and other nodes. We evaluate the framework by using six graph datasets on biological information and social networks, and demonstrate that CapsualGNN outperforms other SOTA techniques on the task of graph classification.

Cultural Embedding: the Logic and Strategy of the Construction of Rural Civilization in the New Era

Culture is the core element of local civilization, and its “embedded” characteristics and value are an important resources for the construction of local civilization. In the new era, the social environment in rural areas has undergone structural changes, and it is difficult to solve the public, organization and participation of the construction of rural customs and civilization to resolve themselves, and the root cause lies in the lack or absence of culture. Under the background of rural revitalization, through the integration of multiple cultural embedding into the rural social structure system, it plays an irreplaceable role in standardizing the rural order and building a rural community, and can inject new elements and new momentum into the construction of rural civilization. Based on this, taking “embeddings” as the theoretical framework and analysis tool, analyzed the internal logic of cultural embedding, and combined with the practice of rural civilization construction, discussed the feasibility of cultural embedding as rural civilization construction in Zhejiang Huzhou.

A heterogeneous stacking ensemble based sentiment analysis framework using multiple word embeddings

AbstractWord embedding techniques have been proposed in the literature to analyze and determine the sentiments expressed in various textual documents such as social media posts, online product reviews, and so forth. However, it is difficult to capture the entire gamut of intricate inter‐dependencies among words in the textual documents using a specific word embedding technique. In this article, we aim to address this issue by proposing a computation‐efficient stacking ensemble based sentiment analysis framework using multiple word embeddings. The proposed framework uses a combination of three distinct word embeddings generated by three different state of the art word embedding techniques, namely, Word2Vec, GloVe, and BERT for performing the sentiment analysis task. It uses an explicitly trained Word2Vec model to generate the first set of 200‐dimensional word embedding. Similarly, pre‐trained GloVe and BERT models are used to generate the other two sets of 200‐dimensional and a 768‐dimensional word embeddings, respectively. These three distinct word embedding sets are then used to train a heterogeneous stacking ensemble based classifier model comprising LSTM, GRU, and Bi‐GRU based base‐level classifiers, and a LSTM based meta‐level classifier. Experimental results on four different datasets, namely, Sentiment140, IMDB Review, Twitter conversation thread, and Twitter Emotion show that the proposed framework achieves high performance with low false positive rate. The proposed framework is also shown to outperform other sentiment analysis frameworks proposed in the literature.

The captivating effect of electric organ discharges: species, sex and orientation are embedded in every single received image.

Some fish communicate using pulsatile, stereotyped electric organ discharges (EODs) that exhibit species- and sex-specific time courses. To ensure reproductive success, they must be able to discriminate conspecifics from sympatric species in the muddy waters they inhabit. We have previously shown that fish in both Gymnotus and Brachyhypopomus genera use the electric field lines as a tracking guide to approach conspecifics (electrotaxis). Here, we show that the social species Brachyhypopomus gauderio uses electrotaxis to arrive abreast a conspecific, coming from behind. Stimulus image analysis shows that, even in a uniform field, every single EOD causes an image in which the gradient and the local field time courses contain enough information to allow the fish to evaluate the conspecific sex, and to find the path to reach it. Using a forced-choice test, we show that sexually mature individuals orient themselves along a uniform field in the direction encoded by the time course characteristic of the opposite sex. This indicates that these fish use the stimulus image profile as a spatial guidance clue to find a mate. Embedding species, sex and orientation cues is a particular example of how species can encode multiple messages in the same self-generated communication signal carrier, allowing for other signal parameters (e.g. EOD timing) to carry additional, often circumstantial, messages. This 'multiple messages' EOD embedding approach expressed in this species is likely to be a common and successful strategy that is widespread across evolutionary lineages and among varied signaling modalities.

Consensus Graph Learning for Multi-View Clustering

Multi-view clustering, which exploits the multi-view information to partition data into their clusters, has attracted intense attention. However, most existing methods directly learn a similarity graph from original multi-view features, which inevitably contain noises and redundancy information. The learned similarity graph is inaccurate and is insufficient to depict the underlying cluster structure of multi-view data. To address this issue, we propose a novel multi-view clustering method that is able to construct an essential similarity graph in a spectral embedding space instead of the original feature space. Concretely, we first obtain multiple spectral embedding matrices from the view-specific similarity graphs, and reorganize the gram matrices constructed by the inner product of the normalized spectral embedding matrices into a tensor. Then, we impose a weighted tensor nuclear norm constraint on the tensor to capture high-order consistent information among multiple views. Furthermore, we unify the spectral embedding and low rank tensor learning into a unified optimization framework to determine the spectral embedding matrices and tensor representation jointly. Finally, we obtain the consensus similarity graph from the gram matrices via an adaptive neighbor manner. An efficient optimization algorithm is designed to solve the resultant optimization problem. Extensive experiments on six benchmark datasets are conducted to verify the efficacy of the proposed method. The code is implemented by using MATLAB R2018a and MindSpore library <xref ref-type="bibr" rid="ref1">[1]</xref>: <uri>https://github.com/guanyuezhen/CGL</uri>.

Stratified Rule-Aware Network for Abstract Visual Reasoning

Abstract reasoning refers to the ability to analyze information, discover rules at an intangible level, and solve problems in innovative ways. Raven's Progressive Matrices (RPM) test is typically used to examine the capability of abstract reasoning. The subject is asked to identify the correct choice from the answer set to fill the missing panel at the bottom right of RPM (e.g., a 3×3 matrix), following the underlying rules inside the matrix. Recent studies, taking advantage of Convolutional Neural Networks (CNNs), have achieved encouraging progress to accomplish the RPM test. However, they partly ignore necessary inductive biases of RPM solver, such as order sensitivity within each row/column and incremental rule induction. To address this problem, in this paper we propose a Stratified Rule-Aware Network (SRAN) to generate the rule embeddings for two input sequences. Our SRAN learns multiple granularity rule embeddings at different levels, and incrementally integrates the stratified embedding flows through a gated fusion module. With the help of embeddings, a rule similarity metric is applied to guarantee that SRAN can not only be trained using a tuplet loss but also infer the best answer efficiently. We further point out the severe defects existing in the popular RAVEN dataset for RPM test, which prevent from the fair evaluation of the abstract reasoning ability. To fix the defects, we propose an answer set generation algorithm called Attribute Bisection Tree (ABT), forming an improved dataset named Impartial-RAVEN (I-RAVEN for short). Extensive experiments are conducted on both PGM and I-RAVEN datasets, showing that our SRAN outperforms the state-of-the-art models by a considerable margin.

Convolutional Adaptive Network for Link Prediction in Knowledge Bases

Knowledge bases (KBs) have become an integral element in digitalization strategies for intelligent engineering and manufacturing. Existing KBs consist of entities and relations and deal with issues of newly added knowledge and completeness. To predict missing information, we introduce an expressive multi-layer network link prediction framework—namely, the convolutional adaptive network (CANet)—which facilitates adaptive feature recalibration by networks to improve the method’s representational power. In CANet, each entity and relation is encoded into a low-dimensional continuous embedding space, and an interaction operation is adopted to generate multiple specific embeddings. These embeddings are concatenated into input matrices, and an attention mechanism is integrated into the convolutional operation. Finally, we use a score function to measure the likelihood of candidate information and a cross-entropy loss function to speed up computation by reducing the convolution operations. Using five real-world KBs, the experimental results indicate that the proposed method achieves state-of-the-art performance.

Improving Risk Assessment of Miscarriage During Pregnancy with Knowledge Graph Embeddings.

Miscarriages are the most common type of pregnancy loss, mostly occurring in the first 12 weeks of pregnancy. Pregnancy risk assessment aims to quantify evidence to reduce such maternal morbidities, and personalized decision support systems are the cornerstone of high-quality, patient-centered care to improve diagnosis, treatment selection, and risk assessment. However, data sparsity and the increasing number of patient-level observations require more effective forms of representing clinical knowledge to encode known information that enables performing inference and reasoning. Whereas knowledge embedding representation has been widely explored in the open domain data, there are few efforts for its application in the clinical domain. In this study, we contrast differences among multiple embedding strategies, and we demonstrate how these methods can assist in performing risk assessment of miscarriage before and during pregnancy. Our experiments show that simple knowledge embedding approaches that utilize domain-specific metadata perform better than complex embedding strategies, although both can improve results comparatively to a population probabilistic baseline in both AUPRC, F1-score, and a proposed normalized version of these evaluation metrics that better reflects accuracy for unbalanced datasets. Finally, embedding approaches provide evidence about each individual, supporting explainability for its model predictions in such a way that humans understand.

Integrating Extracted Information from Bert and Multiple Embedding Methods with the Deep Neural Network for Humour Detection

Humour detection from sentences has been an interesting and challenging task in the last few years. In attempts to highlight humour detection, most research was conducted using traditional approaches of embedding, e.g., Word2Vec or Glove. Recently BERT sentence embedding has also been used for this task. In this paper, we propose a framework for humour detection in short texts taken from news headlines. Our proposed framework (IBEN) attempts to extract information from written text via the use of different layers of BERT. After several trials, weights were assigned to different layers of the BERT model. The extracted information was then sent to a Bi-GRU neural network as an embedding matrix. We utilized the properties of some external embedding models. A multi-kernel convolution in our neural network was also employed to extract higher-level sentence representations. This framework performed very well on the task of humour detection.

A word embedding-based approach to cross-lingual topic modeling

The cross-lingual topic analysis aims at extracting latent topics from corpora of different languages. Early approaches rely on high-cost multilingual resources (e.g., a parallel corpus), which is hard to come by in many real cases. Some works only require a translation dictionary as a linkage between languages; however, when given an inappropriate dictionary (e.g., small coverage of dictionary), the cross-lingual topic model would shrink to a monolingual topic model and generate less diversified topics. Therefore, it is imperative to investigate a cross-lingual topic model requiring fewer bilingual resources. Recently, some space-mapping techniques have been proposed to help align multiple word embedding of different languages into a quality cross-lingual word embedding by referring to a small number of translation pairs. This work proposes a cross-lingual topic model, called Cb-CLTM, which incorporates with cross-lingual word embedding. To leverage the power of word semantics and the linkage between languages from the cross-lingual word embedding, the Cb-CLTM considers each word as a continuous embedding vector rather than a discrete word type. The experiments demonstrate that, when cross-lingual word space exhibits strong isomorphism, Cb-CLTM can generate more coherent topics with higher diversity and induce better representations of documents across languages for further tasks such as cross-lingual document clustering and classification. When the cross-lingual word space is less isomorphic, Cb-CLTM generates less coherent topics yet still prevails in topic diversity and document classification.

Multiple Visual-Semantic Embedding for Video Retrieval from Query Sentence

Visual-semantic embedding aims to learn a joint embedding space where related video and sentence instances are located close to each other. Most existing methods put instances in a single embedding space. However, they struggle to embed instances due to the difficulty of matching visual dynamics in videos to textual features in sentences. A single space is not enough to accommodate various videos and sentences. In this paper, we propose a novel framework that maps instances into multiple individual embedding spaces so that we can capture multiple relationships between instances, leading to compelling video retrieval. We propose to produce a final similarity between instances by fusing similarities measured in each embedding space using a weighted sum strategy. We determine the weights according to a sentence. Therefore, we can flexibly emphasize an embedding space. We conducted sentence-to-video retrieval experiments on a benchmark dataset. The proposed method achieved superior performance, and the results are competitive to state-of-the-art methods. These experimental results demonstrated the effectiveness of the proposed multiple embedding approach compared to existing methods.

Modeling uncertainty to improve personalized recommendations via Bayesian deep learning

Modeling uncertainty has been a major challenge in developing Machine Learning solutions to solve real world problems in various domains. In Recommender Systems, a typical usage of uncertainty is to balance exploration and exploitation, where the uncertainty helps to guide the selection of new options in exploration. Recent advances in combining Bayesian methods with deep learning enable us to express uncertain status in deep learning models. In this paper, we investigate an approach based on Bayesian deep learning to improve personalized recommendations. We first build deep learning architectures to learn useful representation of user and item inputs for predicting their interactions. We then explore multiple embedding components to accommodate different types of user and item inputs. Based on Bayesian deep learning techniques, a key novelty of our approach is to capture the uncertainty associated with the model output and further utilize it to boost exploration in the context of Recommender Systems. We test the proposed approach in both a Collaborative Filtering and a simulated online recommendation setting. Experimental results on publicly available benchmarks demonstrate the benefits of our approach in improving the recommendation performance.

Repeatable Data Hiding: Towards the Reusability of Digital Images

This article proposes a repeatable data hiding framework for digital images, in which the distortion caused by data hiding is invariable no matter how many times the embedding operation is executed. As a result, the usability of images can be always guaranteed. To achieve repeatable data hiding, we deduce the theoretical modification probabilities of the elements in cover image (an image used for data hiding). Then we design an embedding framework to make the practical modification probabilities of cover elements equal to the deduced ones. When additional data is embedded into cover image with the deduced probabilities, the modification-trace of current embedding is replaced by the latter embedding. That means multiple embeddings do not produce new modification-trace, so that the distortion caused by data hiding is invariable. Therefore, the repeatability of data hiding can be guaranteed. Finally, we describe some applications of repeatable data hiding, e.g., logistics management, authentication database management, and steganography, to show the practicability of our framework.

An optimized robust watermarking technique using CKGSA in frequency domain

The existing SVD based watermarking techniques insert watermark into the singular component which causes the false positive problem (FPP). To overcome this problem, a robust watermarking technique using a chaotic kbest gravitational search algorithm (CKGSA) in DCT and SVD domain is proposed. To eliminate the FPP, the watermark is embedded into the principal component (PC) of the cover image rather than the singular value. The objective to use CKGSA is to optimize the value of multiple embedding factors so that robustness and imperceptibility can be balanced. Arnold transform (AT) is applied to the watermark image before embedding, to enhance security and robustness. The performance of the proposed technique has been experimented on a standard dataset and compared with the six recent state-of-the-art techniques including DS-DE, DS-PSO, DCT-SVD, DDS-PSO, DS-GSA, and DS-TLBO in terms of imperceptibility (PSNR, SSIM) and robustness (PSNR, NC). The experimental results demonstrate that the proposed technique outperforms the other existing techniques.

A Graph Gaussian Embedding Method for Predicting Alzheimer's Disease Progression With MEG Brain Networks.

Characterizing the subtle changes of functional brain networks associated with the pathological cascade of Alzheimer's disease (AD) is important for early diagnosis and prediction of disease progression prior to clinical symptoms. We developed a new deep learning method, termed multiple graph Gaussian embedding model (MG2G), which can learn highly informative network features by mapping high-dimensional resting-state brain networks into a low-dimensional latent space. These latent distribution-based embeddings enable a quantitative characterization of subtle and heterogeneous brain connectivity patterns at different regions, and can be used as input to traditional classifiers for various downstream graph analytic tasks, such as AD early stage prediction, and statistical evaluation of between-group significant alterations across brain regions. We used MG2G to detect the intrinsic latent dimensionality of MEG brain networks, predict the progression of patients with mild cognitive impairment (MCI) to AD, and identify brain regions with network alterations related to MCI.

Multiple Embeddings Enhanced Multi-Graph Neural Networks for Chinese Healthcare Named Entity Recognition.

Named Entity Recognition (NER) is a natural language processing task for recognizing named entities in a given sentence. Chinese NER is difficult due to the lack of delimited spaces and conventional features for determining named entity boundaries and categories. This study proposes the ME-MGNN (Multiple Embeddings enhanced Multi-Graph Neural Networks) model for Chinese NER in the healthcare domain. We integrate multiple embeddings at different granularities from the radical, character to word levels for an extended character representation, and this is fed into multiple gated graph sequence neural networks to identify named entities and classify their types. The experimental datasets were collected from health-related news, digital health magazines and medical question/answer forums. Manual annotation was conducted for a total of 68,460 named entities across 10 entity types (body, symptom, instrument, examination, chemical, disease, drug, supplement, treatment and time) in 30,692 sentences. Experimental results indicated our ME-MGNN model achieved an F1-score result of 75.69, outperforming previous methods. In practice, a series of model analysis implied that our method is effective and efficient for Chinese healthcare NER.

Cross-Lingual Neural Network Speech Synthesis Based on Multiple Embeddings

Cross-Lingual Neural Network Speech Synthesis Based on Multiple Embeddings

Fine-Grained Fashion Similarity Prediction by Attribute-Specific Embedding Learning.

This paper strives to predict fine-grained fashion similarity. In this similarity paradigm, one should pay more attention to the similarity in terms of a specific design/attribute between fashion items. For example, whether the collar designs of the two clothes are similar. It has potential value in many fashion related applications, such as fashion copyright protection. To this end, we propose an Attribute-Specific Embedding Network (ASEN) to jointly learn multiple attribute-specific embeddings, thus measure the fine-grained similarity in the corresponding space. The proposed ASEN is comprised of a global branch and a local branch. The global branch takes the whole image as input to extract features from a global perspective, while the local branch takes as input the zoomed-in region-of-interest (RoI) w.r.t. the specified attribute thus able to extract more fine-grained features. As the global branch and the local branch extract the features from different perspectives, they are complementary to each other. Additionally, in each branch, two attention modules, i.e., Attribute-aware Spatial Attention and Attribute-aware Channel Attention, are integrated to make ASEN be able to locate the related regions and capture the essential patterns under the guidance of the specified attribute, thus make the learned attribute-specific embeddings better reflect the fine-grained similarity. Extensive experiments on three fashion-related datasets, i.e., FashionAI, DARN, and DeepFashion, show the effectiveness of ASEN for fine-grained fashion similarity prediction and its potential for fashion reranking. Code and data are available at https://github.com/maryeon/asenpp.