Representation Learning Framework Research Articles

Discriminative Local Sparse Representation by Robust Adaptive Dictionary Pair Learning.

In this article, we propose a structured robust adaptive dictionary pair learning (RA-DPL) framework for the discriminative sparse representation (SR) learning. To achieve powerful representation ability of the available samples, the setting of RA-DPL seamlessly integrates the robust projective DPL, locality-adaptive SRs, and discriminative coding coefficients learning into a unified learning framework. Specifically, RA-DPL improves existing projective DPL in four perspectives. First, it applies a sparse l2,1 -norm-based metric to encode the reconstruction error to deliver the robust projective dictionary pairs, and the l2,1 -norm has the potential to minimize the error. Second, it imposes the robust l2,1 -norm clearly on the analysis dictionary to ensure the sparse property of the coding coefficients rather than using the costly l0/l1 -norm. As such, the robustness of the data representation and the efficiency of the learning process are jointly considered to guarantee the efficacy of our RA-DPL. Third, RA-DPL conceives a structured reconstruction weight learning paradigm to preserve the local structures of the coding coefficients within each class clearly in an adaptive manner, which encourages to produce the locality preserving representations. Fourth, it also considers improving the discriminating ability of coding coefficients and dictionary by incorporating a discriminating function, which can ensure high intraclass compactness and interclass separation in the code space. Extensive experiments show that our RA-DPL can obtain superior performance over other state of the arts.

Learning Graph Representation With Generative Adversarial Nets

Graph representation learning aims to embed each vertex in a graph into a low-dimensional vector space. Existing graph representation learning methods can be classified into two categories: generative models that learn the underlying connectivity distribution in a graph, and discriminative models that predict the probability of edge between a pair of vertices. In this paper, we propose GraphGAN, an innovative graph representation learning framework unifying the above two classes of methods, in which the generative and the discriminative model play a game-theoretical minimax game. Specifically, for a given vertex, the generative model tries to fit its underlying true connectivity distribution over all other vertices and produces “fake” samples to fool the discriminative model, while the discriminative model tries to detect whether the sampled vertex is from ground truth or generated by the generative model. With the competition between these two models, both of them can alternately and iteratively boost their performance. Moreover, we propose a novel graph softmax as the implementation of the generative model to overcome the limitations of traditional softmax function, which can be proven satisfying desirable properties of normalization, graph structure awareness, and computational efficiency. Through extensive experiments on real-world datasets, we demonstrate that GraphGAN achieves substantial gains in a variety of applications, including graph reconstruction, link prediction, node classification, recommendation, and visualization, over state-of-the-art baselines.

DLTSR: A Deep Learning Framework for Recommendations of Long-Tail Web Services

With the growing popularity of web services, more and more developers are composing multiple services into mashups. Developers show an increasing interest in non-popular services (i.e., long-tail ones), however, there are very scarce studies trying to address the long-tail web service recommendation problem. The major challenges for recommending long-tail services accurately include severe sparsity of historical usage data and unsatisfactory quality of description content. In this paper, we propose to build a deep learning framework to address these challenges and perform accurate long-tail recommendations. To tackle the problem of unsatisfactory quality of description content, we use stacked denoising autoencoders (SDAE) to perform feature extraction. Additionally, we impose the usage records in hot services as a regularization of the encoding output of SDAE, to provide feedback to content extraction. To address the sparsity of historical usage data, we learn the patterns of developers’ preference instead of modeling individual services. Our experimental results on a real-world dataset demonstrate that, with such joint autoencoder based feature representation and content-usage learning framework, the proposed algorithm outperforms the state-of-the-art baselines significantly.

Interpretable Representation Learning for Speech and Audio Signals Based on Relevance Weighting

The learning of interpretable representations from raw data presents significant challenges for time series data like speech. In this work, we propose a relevance weighting scheme that allows the interpretation of the speech representations during the forward propagation of the model itself. The relevance weighting is achieved using a sub-network approach that performs the task of feature selection. A relevance sub-network, applied on the output of first layer of a convolutional neural network model operating on raw speech signals, acts as an acoustic filterbank (FB) layer with relevance weighting. A similar relevance sub-network applied on the second convolutional layer performs modulation filterbank learning with relevance weighting. The full acoustic model consisting of relevance sub-networks, convolutional layers and feed-forward layers is trained for a speech recognition task on noisy and reverberant speech in the Aurora-4, CHiME-3 and VOiCES datasets. The proposed representation learning framework is also applied for the task of sound classification in the UrbanSound8K dataset. A detailed analysis of the relevance weights learned by the model reveals that the relevance weights capture information regarding the underlying speech/audio content. In addition, speech recognition and sound classification experiments reveal that the incorporation of relevance weighting in the neural network architecture improves the performance significantly.

MSHINE: A Multiple-meta-paths Simultaneous Learning Framework for Heterogeneous Information Network Embedding

Heterogeneous information networks(HINs) become popular in recent years for its strong capability of modelling objects with abundant information using explicit network structure. Network embedding has been proved as an effective method to convert information networks into lower-dimensional space, whereas the core information can be well preserved. However, traditional network embedding algorithms are sub-optimal in capturing rich while potentially incompatible semantics provided by HINs. To address this issue, a novel meta-path-based HIN representation learning framework named mSHINE is designed to simultaneously learn multiple node representations for different meta-paths. More specifically, one representation learning module inspired by the RNN structure is developed and multiple node representations can be learned simultaneously, where each representation is associated with one respective meta-path. By measuring the relevance between nodes with the designed objective function, the learned module can be applied in downstream link prediction tasks. A set of criteria for selecting initial meta-paths is proposed as the other module in mSHINE which is important to reduce the optimal meta-path selection cost when no prior knowledge of suitable meta-paths is available. To corroborate the effectiveness of mSHINE, extensive experimental studies including node classification and link prediction are conducted on five real-world datasets. The results demonstrate that mSHINE outperforms other state-of-the-art HIN embedding methods.

Contrastive Representation Learning: A Framework and Review

Contrastive Learning has recently received interest due to its success in self-supervised representation learning in the computer vision domain. However, the origins of Contrastive Learning date as far back as the 1990s and its development has spanned across many fields and domains including Metric Learning and natural language processing. In this paper we provide a comprehensive literature review and we propose a general Contrastive Representation Learning framework that simplifies and unifies many different contrastive learning methods. We also provide a taxonomy for each of the components of contrastive learning in order to summarise it and distinguish it from other forms of machine learning. We then discuss the inductive biases which are present in any contrastive learning system and we analyse our framework under different views from various sub-fields of Machine Learning. Examples of how contrastive learning has been applied in computer vision, natural language processing, audio processing, and others, as well as in Reinforcement Learning are also presented. Finally, we discuss the challenges and some of the most promising future research directions ahead.

Knowledge Distillation-Based Representation Learning for Short-Utterance Spoken Language Identification

With successful applications of deep feature learning algorithms, spoken language identification (LID) on long utterances obtains satisfactory performance. However, the performance on short utterances is drastically degraded even when the LID system is trained using short utterances. The main reason is due to the large variation of the representation on short utterances which results in high model confusion. To narrow the performance gap between long, and short utterances, we proposed a teacher-student representation learning framework based on a knowledge distillation method to improve LID performance on short utterances. In the proposed framework, in addition to training the student model on short utterances with their true labels, the internal representation from the output of a hidden layer of the student model is supervised with the representation corresponding to their longer utterances. By reducing the distance of internal representations between short, and long utterances, the student model can explore robust discriminative representations for short utterances, which is expected to reduce model confusion. We conducted experiments on our in-house LID dataset, and NIST LRE07 dataset, and showed the effectiveness of the proposed methods for short utterance LID tasks.

Multimodal deep representation learning for protein interaction identification and protein family classification

BackgroundProtein-protein interactions(PPIs) engage in dynamic pathological and biological procedures constantly in our life. Thus, it is crucial to comprehend the PPIs thoroughly such that we are able to illuminate the disease occurrence, achieve the optimal drug-target therapeutic effect and describe the protein complex structures. However, compared to the protein sequences obtainable from various species and organisms, the number of revealed protein-protein interactions is relatively limited. To address this dilemma, lots of research endeavor have investigated in it to facilitate the discovery of novel PPIs. Among these methods, PPI prediction techniques that merely rely on protein sequence data are more widespread than other methods which require extensive biological domain knowledge.ResultsIn this paper, we propose a multi-modal deep representation learning structure by incorporating protein physicochemical features with the graph topological features from the PPI networks. Specifically, our method not only bears in mind the protein sequence information but also discerns the topological representations for each protein node in the PPI networks. In our paper, we construct a stacked auto-encoder architecture together with a continuous bag-of-words (CBOW) model based on generated metapaths to study the PPI predictions. Following by that, we utilize the supervised deep neural networks to identify the PPIs and classify the protein families. The PPI prediction accuracy for eight species ranged from 96.76% to 99.77%, which signifies that our multi-modal deep representation learning framework achieves superior performance compared to other computational methods.ConclusionTo the best of our knowledge, this is the first multi-modal deep representation learning framework for examining the PPI networks.

Heterogeneous graph-based joint representation learning for users and POIs in location-based social network

Learning latent representations for users and points of interests (POIs) is an important task in location-based social networks (LBSN), which could largely benefit multiple location-based services, such as POI recommendation and social link prediction. Many contextual factors, like geographical influence, user social relationship and temporal information, are available in LBSN and would be useful for this task. However, incorporating all these contextual factors for user and POI representation learning in LBSN remains challenging, due to their heterogeneous nature. Although the encouraging performance of POI recommendation and social link prediction are delivered, most of the existing representation learning methods for LBSN incorporate only one or two of these contextual factors. In this paper, we propose a novel joint representation learning framework for users and POIs in LBSN, named UP2VEC. In UP2VEC, we present a heterogeneous LBSN graph to incorporate all these aforementioned factors. Specifically, the transition probabilities between nodes inside the heterogeneous graph are derived by jointly considering these contextual factors. The latent representations of users and POIs are then learnt by matching the topological structure of the heterogeneous graph. For evaluating the effectiveness of UP2VEC, a series of experiments are conducted with two real-world datasets (Foursquare and Gowalla) in terms of POI recommendation and social link prediction. Experimental results demonstrate that the proposed UP2VEC significantly outperforms the existing state-of-the-art alternatives. Further experiment shows the superiority of UP2VEC in handling cold-start problem for POI recommendation.

Simultaneous multiple features tracking of beats: A representation learning approach to reduce false alarm rates in ICUs.

The high rate of false alarms is a key challenge related to patient care in intensive care units (ICUs) that can result in delayed responses of the medical staff. Several rule-based and machine learning-based techniques have been developed to address this problem. However, the majority of these methods rely on the availability of different physiological signals such as different electrocardiogram (ECG) leads, arterial blood pressure (ABP), and photoplethysmogram (PPG), where each signal is analyzed by an independent processing unit and the results are fed to an algorithm to determine an alarm. That calls for novel methods that can accurately detect the cardiac events by only accessing one signal (e.g., ECG) with a low level of computation and sensors requirement. We propose a novel and robust representation learning framework for ECG analysis that only rely on a single lead ECG signal and yet achieves considerably better performance compared to the state-of-the-art works in this domain, without relying on an expert knowledge. We evaluate the performance of this method using the "2015 Physionet computing in cardiology challenge" dataset. To the best of our knowledge, the best previously reported performance is based on both expert knowledge and machine learning where all available signals of ECG, ABP and PPG are utilized. Our proposed method reaches the performance of 97.3%, 95.5 %, and 90.8 % in terms of sensitivity, specificity, and the challenge's score, respectively for the detection of five arrhythmias when only one single ECG lead signals is used without any expert knowledge.

Semi-supervised representation learning via dual autoencoders for domain adaptation

Domain adaptation aims to exploit the knowledge in source domain to promote the learning tasks in target domain, which plays a critical role in real-world applications. Recently, lots of deep learning approaches based on autoencoders have achieved a significance performance in domain adaptation. However, most existing methods focus on minimizing the distribution divergence by putting the source and target data together to learn global feature representations, while they do not consider the local relationship between instances in the same category from different domains. To address this problem, we propose a novel Semi-Supervised Representation Learning framework via Dual Autoencoders for domain adaptation, named SSRLDA. More specifically, we extract richer feature representations by learning the global and local feature representations simultaneously using two novel autoencoders, which are referred to as marginalized denoising autoencoder with adaptation distribution (MDAad) and multi-class marginalized denoising autoencoder (MMDA) respectively. Meanwhile, we make full use of label information to optimize feature representations. Experimental results show that our proposed approach outperforms several state-of-the-art baseline methods.

Driver Drowsiness Detection Using Condition-Adaptive Representation Learning Framework

We propose a condition-adaptive representation learning framework for the driver drowsiness detection based on 3D-deep convolutional neural network. The proposed framework consists of four models: spatio-temporal representation learning, scene condition understanding, feature fusion, and drowsiness detection. The spatio-temporal representation learning extracts features that can describe motions and appearances in video simultaneously. The scene condition understanding classifies the scene conditions related to various conditions about the drivers and driving situations such as statuses of wearing glasses, illumination condition of driving, and motion of facial elements such as head, eye, and mouth. The feature fusion generates a condition-adaptive representation using two features extracted from above models. The detection model recognizes drivers drowsiness status using the condition-adaptive representation. The condition-adaptive representation learning framework can extract more discriminative features focusing on each scene condition than the general representation so that the drowsiness detection method can provide more accurate results for the various driving situations. The proposed framework is evaluated with the NTHU Drowsy Driver Detection video dataset. The experimental results show that our framework outperforms the existing drowsiness detection methods based on visual analysis.

Hypergraph Neural Networks

In this paper, we present a hypergraph neural networks (HGNN) framework for data representation learning, which can encode high-order data correlation in a hypergraph structure. Confronting the challenges of learning representation for complex data in real practice, we propose to incorporate such data structure in a hypergraph, which is more flexible on data modeling, especially when dealing with complex data. In this method, a hyperedge convolution operation is designed to handle the data correlation during representation learning. In this way, traditional hypergraph learning procedure can be conducted using hyperedge convolution operations efficiently. HGNN is able to learn the hidden layer representation considering the high-order data structure, which is a general framework considering the complex data correlations. We have conducted experiments on citation network classification and visual object recognition tasks and compared HGNN with graph convolutional networks and other traditional methods. Experimental results demonstrate that the proposed HGNN method outperforms recent state-of-theart methods. We can also reveal from the results that the proposed HGNN is superior when dealing with multi-modal data compared with existing methods.

Semantic Relationships Guided Representation Learning for Facial Action Unit Recognition

Facial action unit (AU) recognition is a crucial task for facial expressions analysis and has attracted extensive attention in the field of artificial intelligence and computer vision. Existing works have either focused on designing or learning complex regional feature representations, or delved into various types of AU relationship modeling. Albeit with varying degrees of progress, it is still arduous for existing methods to handle complex situations. In this paper, we investigate how to integrate the semantic relationship propagation between AUs in a deep neural network framework to enhance the feature representation of facial regions, and propose an AU semantic relationship embedded representation learning (SRERL) framework. Specifically, by analyzing the symbiosis and mutual exclusion of AUs in various facial expressions, we organize the facial AUs in the form of structured knowledge-graph and integrate a Gated Graph Neural Network (GGNN) in a multi-scale CNN framework to propagate node information through the graph for generating enhanced AU representation. As the learned feature involves both the appearance characteristics and the AU relationship reasoning, the proposed model is more robust and can cope with more challenging cases, e.g., illumination change and partial occlusion. Extensive experiments on the two public benchmarks demonstrate that our method outperforms the previous work and achieves state of the art performance.

Joint Representation Learning for Multi-Modal Transportation Recommendation

Multi-modal transportation recommendation has a goal of recommending a travel plan which considers various transportation modes, such as walking, cycling, automobile, and public transit, and how to connect among these modes. The successful development of multi-modal transportation recommendation systems can help to satisfy the diversified needs of travelers and improve the efficiency of transport networks. However, existing transport recommender systems mainly focus on unimodal transport planning. To this end, in this paper, we propose a joint representation learning framework for multi-modal transportation recommendation based on a carefully-constructed multi-modal transportation graph. Specifically, we first extract a multi-modal transportation graph from large-scale map query data to describe the concurrency of users, Origin-Destination (OD) pairs, and transport modes. Then, we provide effective solutions for the optimization problem and develop an anchor embedding for transport modes to initialize the embeddings of transport modes. Moreover, we infer user relevance and OD pair relevance, and incorporate them to regularize the representation learning. Finally, we exploit the learned representations for online multimodal transportation recommendations. Indeed, our method has been deployed into one of the largest navigation Apps to serve hundreds of millions of users, and extensive experimental results with real-world map query data demonstrate the enhanced performance of the proposed method for multimodal transportation recommendations.

UGSD: User Generated Sentiment Dictionaries from Online Customer Reviews

Customer reviews on platforms such as TripAdvisor and Amazon provide rich information about the ways that people convey sentiment on certain domains. Given these kinds of user reviews, this paper proposes UGSD, a representation learning framework for constructing domain-specific sentiment dictionaries from online customer reviews, in which we leverage the relationship between user-generated reviews and the ratings of the reviews to associate the reviewer sentiment with certain entities. The proposed framework has the following three main advantages. First, no additional annotations of words or external dictionaries are needed for the proposed framework; the only resources needed are the review texts and entity ratings. Second, the framework is applicable across a variety of user-generated content from different domains to construct domain-specific sentiment dictionaries. Finally, each word in the constructed dictionary is associated with a low-dimensional dense representation and a degree of relatedness to a certain rating, which enable us to obtain more fine-grained dictionaries and enhance the application scalability of the constructed dictionaries as the word representations can be adopted for various tasks or applications, such as entity ranking and dictionary expansion. The experimental results on three real-world datasets show that the framework is effective in constructing high-quality domain-specific sentiment dictionaries from customer reviews.

Letting Logos Speak: Leveraging Multiview Representation Learning for Data-Driven Logo Design

Logos serve a fundamental role as the visual figureheads of brands. Yet, due to the difficulty of using unstructured image data, prior research on logo design has largely been limited to non-quantitative studies. In this work, we explore the interplay between logo design and brand identity creation from a data-driven perspective. We develop both a novel logo feature extraction algorithm that uses modern image processing tools to decompose pixel-level image data into meaningful features, and a multiview representation learning framework that links these visual features to textual descriptions, consumer ratings of brand personality, and other high-level tags describing firms. We apply this framework to a unique dataset of brands, to understand which brands use which logo features, and how consumers evaluate these brands’ personalities. Moreover, we show that manipulating the model’s learned representations through what we term “brand arithmetic” yields new brand identities, and can help with ideation. Finally, through an application to fast food branding, we show how our model can be used as a decision support tool for suggesting typical logo features for a brand, and for predicting consumers’ reactions to new brands or rebranding efforts.

Learning Domain-Independent Deep Representations by Mutual Information Minimization.

Domain transfer learning aims to learn common data representations from a source domain and a target domain so that the source domain data can help the classification of the target domain. Conventional transfer representation learning imposes the distributions of source and target domain representations to be similar, which heavily relies on the characterization of the distributions of domains and the distribution matching criteria. In this paper, we proposed a novel framework for domain transfer representation learning. Our motive is to make the learned representations of data points independent from the domains which they belong to. In other words, from an optimal cross-domain representation of a data point, it is difficult to tell which domain it is from. In this way, the learned representations can be generalized to different domains. To measure the dependency between the representations and the corresponding domain which the data points belong to, we propose to use the mutual information between the representations and the domain-belonging indicators. By minimizing such mutual information, we learn the representations which are independent from domains. We build a classwise deep convolutional network model as a representation model and maximize the margin of each data point of the corresponding class, which is defined over the intraclass and interclass neighborhood. To learn the parameters of the model, we construct a unified minimization problem where the margins are maximized while the representation-domain mutual information is minimized. In this way, we learn representations which are not only discriminate but also independent from domains. An iterative algorithm based on the Adam optimization method is proposed to solve the minimization to learn the classwise deep model parameters and the cross-domain representations simultaneously. Extensive experiments over benchmark datasets show its effectiveness and advantage over existing domain transfer learning methods.

A Unified Framework for Community Detection and Network Representation Learning

Network representation learning (NRL) aims to learn low-dimensional vectors for vertices in a network. Most existing NRL methods focus on learning representations from local context of vertices (such as their neighbors). Nevertheless, vertices in many complex networks also exhibit significant global patterns widely known as communities. It's intuitive that vertices in the same community tend to connect densely and share common attributes. These patterns are expected to improve NRL and benefit relevant evaluation tasks, such as link prediction and vertex classification. Inspired by the analogy between network representation learning and text modeling, we propose a unified NRL framework by introducing community information of vertices, named as Community-enhanced Network Representation Learning (CNRL). CNRL simultaneously detects community distribution of each vertex and learns embeddings of both vertices and communities. Moreover, the proposed community enhancement mechanism can be applied to various existing NRL models. In experiments, we evaluate our model on vertex classification, link prediction, and community detection using several real-world datasets. The results demonstrate that CNRL significantly and consistently outperforms other state-of-the-art methods while verifying our assumptions on the correlations between vertices and communities.

Representation learning for automated spectroscopic redshift estimation

Context. Determining the radial positions of galaxies up to a high accuracy depends on the correct identification of salient features in their spectra. Classical techniques for spectroscopic redshift estimation make use of template matching with cross-correlation. These templates are usually constructed from empirical spectra or simulations based on the modeling of local galaxies. Aims. We propose two new spectroscopic redshift estimation schemes based on new learning techniques for galaxy spectra representation, using either a dictionary learning technique for sparse representation or denoising autoencoders. We investigate how these representations impact redshift estimation. Methods. We first explored dictionary learning to obtain a sparse representation of the rest-frame galaxy spectra modeling both the continuum and line emissions. As an alternative, denoising autoencoders were considered to learn non-linear representations from rest-frame emission lines extracted from the data. In both cases, the redshift was then determined by redshifting the learnt representation and selecting the redshift that gave the lowest approximation error among the tested values. Results. These methods have been tested on realistic simulated galaxy spectra, with photometry modeled after the Large Synoptic Survey Telescope (LSST) and spectroscopy reproducing properties of the Sloan Digital Sky Survey (SDSS). They were compared to Darth Fader, a robust technique extracting line features and estimating redshift through eigentemplates cross-correlations. We show that both dictionary learning and denoising autoencoders provide improved accuracy and reliability across all signal-to-noise (S/N) regimes and galaxy types. Furthermore, the former is more robust at high noise levels; the latter is more accurate on high S/N regimes. Combining both estimators improves results at low S/N. Conclusions. The representation learning framework for spectroscopic redshift analysis introduced in this work offers high performance in feature extraction and redshift estimation, improving on a classical eigentemplates approach. This is a necessity for next-generation galaxy surveys, and we demonstrate a successful application in realistic simulated survey data.