:::
I n recent years, waveform-mapping-based speech enhancement (SE) methods have garnered significant attention. These methods generally use a deep learning model to directly process and reconstruct speech waveforms. Because both the input and output are in waveform format, the waveform-mapping-based SE methods can overcome the distortion caused by imperfect phase estimation, which may be encountered in spectral-mapping-based SE systems. So far, most waveform-mapping-based SE methods have focused on single-channel tasks. In this paper, we propose a novel fully convolutional network (FCN) with Sinc and dilated convolutional layers (termed SDFCN) for multichannel SE that operates in the time domain. We also propose an extended version of SDFCN, called the residual SDFCN (termed rSDFCN). The proposed methods are evaluated on three multichannel SE tasks, namely the dual-channel inner-ear microphones SE task, the distributed microphones SE task, and the CHiME-3 dataset. The experimental results confirm the outstanding denoising capability of the proposed SE systems on both tasks and the benefits of using the residual architecture on the overall SE performance.
Unconstrained video-based face recognition is a challenging problem due to significant within-video variations caused by pose, occlusion and blur. To tackle this problem, an effective idea is to propagate the identity from high-quality faces to low-quality ones through contextual connections, which are constructed based on context such as body appearance. However, previous methods have often propagated erroneous information due to lack of uncertainty modeling of the noisy contextual connections. In this paper, we propose the Uncertainty-Gated Graph (UGG), which conducts graph-based identity propagation between tracklets, which are represented by nodes in a graph. UGG explicitly models the uncertainty of the contextual connections by adaptively updating the weights of the edge gates according to the identity distributions of the nodes during inference. UGG is a generic graphical model that can be applied at only inference time or with end-to-end training. We demonstrate the effectiveness of UGG with state-of-the-art results in the recently released challenging Cast Search in Movies and IARPA Janus Surveillance Video Benchmark dataset.
In recent years, attention models have been extensively used for person and vehicle re-identification. Most re-identification methods are designed to focus attention on key-point locations. However, depending on the orientation, the contribution of each key-point varies. In this paper, we present a novel dual-path adaptive attention model for vehicle re-identification (AAVER). The global appearance path captures macroscopic vehicle features while the orientation conditioned part appearance path learns to capture localized discriminative features by focusing attention on the most informative key-points. Through extensive experimentation, we show that the proposed AAVER method is able to accurately re-identify vehicles in unconstrained scenarios, yielding state of the art results on the challenging dataset VeRi-776. As a byproduct, the proposed system is also able to accurately predict vehicle key-points and shows an improvement of more than 7% over state of the art.
Image deblurring aims to restore the latent sharp images from the corresponding blurred ones. In this paper, we present an unsupervised method for domain-specific single-image deblurring based on disentangled representations. The disentanglement is achieved by splitting the content and blur features in a blurred image using content encoders and blur encoders. We enforce a KL divergence loss to regularize the distribution range of extracted blur attributes such that little content information is contained. Meanwhile, to handle the unpaired training data, a blurring branch and the cycle-consistency loss are added to guarantee that the content structures of the deblurred results match the original images. We also add an adversarial loss on deblurred results to generate visually realistic images and a perceptual loss to further mitigate the artifacts. We perform extensive experiments on the tasks of face and text deblurring using both synthetic datasets and real images, and achieve improved results compared to recent state-of-the-art deblurring methods.
Covariates are factors that have a debilitating influence on face verification performance. In this paper, we comprehensively study two covariate related problems for unconstrained face verification: first, how covariates affect the performance of deep neural networks on the large-scale unconstrained face verification problem; second, how to utilize covariates to improve verification performance. To study the first problem, we implement five state-of-the-art deep convolutional networks and evaluate them on three challenging covariates datasets. In total, seven covariates are considered: pose (yaw and roll), age, facial hair, gender, indoor/outdoor, occlusion (nose and mouth visibility, and forehead visibility), and skin tone. We first report the performance of each individual network on the overall protocol and use the score-level fusion method to analyze each covariate. Some of the results confirm and extend the findings of previous studies, and others are new findings that were rarely mentioned previously or did not show consistent trends. For the second problem, we demonstrate that with the assistance of gender information, the quality of a precurated noisy large-scale face dataset for face recognition can be further improved. After retraining the face recognition model using the curated data, performance improvement is observed at low false acceptance rates.
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.
A catastrophe equity put (CatEPut) is constructed to recapitalize an insurance company that suffers huge compensation payouts due to catastrophic events (CEs). The company can exercise its CatEPut to sell its stock to the counterparty at a predetermined price when its accumulated loss due to CEs exceeds a predetermined threshold and its own stock price falls below the strike price. Much literature considers the evaluations of a CatEPut that can only be exercised at maturity; however, most CatEPuts can be exercised early so the company can receive timely funding. This paper adopts lattice approaches to evaluate CatEPuts with early exercise features. To solve the combinatorial exposition problem due to the trigger of CatEPuts’ accumulated loss, our method reduces the possible number of accumulated losses by taking advantage of the closeness of integral additions. We also identify and alleviate a new type of nonlinearity error that yields unstable numerical pricing results by adjusting the lattice structure. We provide a rigorous mathematical proof to show how the proposed lattice can be constructed under a mild condition. Comprehensive numerical experiments are also given to demonstrate the robustness and efficiency of our lattice.
Recommender systems are vital ingredients for many e-commerce services. In the literature, two of the most popular approaches are based on factorization and graph-based models; the former approach captures user preferences by factorizing the observed direct interactions between users and items, and the latter extracts indirect preferences from the graphs constructed by user-item interactions. In this paper we present HOP-Rec , a uni ed and e cient method that incorporates the two approaches. The proposed method in- volves random sur ng on a graph to harvest high-order information among neighborhood items for each user. Instead of factorizing a transition matrix, our method introduces a con dence weighting parameter to simulate all high-order information simultaneously, for which we maintain a sparse user-item interaction matrix and enrich the matrix for each user using random walks. Experimental results show that our approach signi cantly outperforms the state of the art on a range of large-scale real-world datasets.
In this paper, we propose a smartphone-based hearing assistive system (termed SmartHear) to facilitate speech recognition for various target users who could benefit from enhanced listening clarity in the classroom. The SmartHear system consists of transmitter and receiver devices (e.g., smartphone and Bluetooth headset) for voice transmission, and an Android mobile application that controls and connects the different devices via Bluetooth or WiFi technology. The wireless transmission of voice signals between devices overcomes the reverberation and ambient noise effects in the classroom. The main functionalities of SmartHear include: 1) configurable transmitter/receiver assignment, to allow flexible designation of transmitter/receiver roles; 2) advanced noise-reduction techniques; 3) audio recording; and 4) voice-to-text conversion, to give students visual text aid. All the functions are implemented as a mobile application with an easy-to-navigate user interface. Experiments show the effectiveness of the noise-reduction schemes at low signal-to-noise ratios (SNR) in terms of standard speech perception and quality indices, and show the effectiveness of SmartHear in maintaining voice-to-text conversion accuracy regardless of the distance between the speaker and listener. Future applications of SmartHear are also discussed.
This paper proposes an item concept embedding (ICE) framework to model item concepts via textual information. Specifically, in the proposed framework there are two stages: graph construction and embedding learning. In the first stage, we propose a generalized network construction method to build a network involving heterogeneous nodes and a mixture of both homogeneous and heterogeneous relations. The second stage leverages the concept of neighborhood proximity to learn the embeddings of both items and words. With the proposed carefully designed ICE networks, the resulting embedding facilitates both homogeneous and heterogeneous retrieval, including item-to-item and word-to-item retrieval. Moreover, as a distributed embedding approach, the proposed ICE approach not only generates related retrieval results but also delivers more diverse results than traditional keyword-matching-based approaches. As our experiments on two real-world datasets show, ICE encodes useful textual information and thus outperforms traditional methods in various item classification and retrieval tasks.