Zero-shot learning aims to classify unseen image categories by learning a visual-semantic embedding space. In most cases, the traditional methods adopt a separated two-step pipeline that extracts image features are utilized to learn the embedding space. It leads to the lack of specific structural semantic information of image features for zero-shot learning task. In this paper, we propose an end-to-end trainable Deep Semantic Structural Constraints model to address this issue. The proposed model contains the Image Feature Structure constraint and the Semantic Embedding Structure constraint, which aim to learn structure-preserving image features and endue the learned embedding space with stronger generalization ability respectively. With the assistance of semantic structural information, the model gains more auxiliary clues for zero-shot learning. The state-of-the-art performance certifies the effectiveness of our proposed method.

Multimodal models have been proven to outperform text-based models on learning semantic word representations. Almost all previous multimodal models typically treat the representations from different modalities equally. However, it is obvious that information from different modalities contributes differently to the meaning of words. This motivates us to build a multimodal model that can dynamically fuse the semantic representations from different modalities according to different types of words. To that end, we propose three novel dynamic fusion methods to assign importance weights to each modality, in which weights are learned under the weak supervision of word association pairs. The extensive experiments have demonstrated that the proposed methods outperform strong unimodal baselines and state-of-the-art multimodal models.

A novel deep neural network training paradigm that exploits the conjoint information in multiple heterogeneous sources is proposed. Specifically, in a RGB-D based action recognition task, it cooperatively trains a single convolutional neural network (named c-ConvNet) on both RGB visual features and depth features, and deeply aggregates the two kinds of features for action recognition. Differently from the conventional ConvNet that learns the deep separable features for homogeneous modality-based classification with only one softmax loss function, the c-ConvNet enhances the discriminative power of the deeply learned features and weakens the undesired modality discrepancy by jointly optimizing a ranking loss and a softmax loss for both homogeneous and heterogeneous modalities. The ranking loss consists of intra-modality and cross-modality triplet losses, and it reduces both the intra-modality and cross-modality feature variations. Furthermore, the correlations between RGB and depth data are embedded in the c-ConvNet, and can be retrieved by either of the modalities and contribute to the recognition in the case even only one of the modalities is available. The proposed method was extensively evaluated on two large RGB-D action recognition datasets, ChaLearn LAP IsoGD and NTU RGB+D datasets, and one small dataset, SYSU 3D HOI, and achieved state-of-the-art results.

Makeup is widely used to improve facial attractiveness and is well accepted by the public. However, different makeup styles will result in significant facial appearance changes. It remains a challenging problem to match makeup and non-makeup face images. This paper proposes a learning from generation approach for makeup-invariant face verification by introducing a bi-level adversarial network (BLAN). To alleviate the negative effects from makeup, we first generate non-makeup images from makeup ones, and then use the synthesized non-makeup images for further verification. Two adversarial networks in BLAN are integrated in an end-to-end deep network, with the one on pixel level for reconstructing appealing facial images and the other on feature level for preserving identity information. These two networks jointly reduce the sensing gap between makeup and non-makeup images. Moreover, we make the generator well constrained by incorporating multiple perceptual losses. Experimental results on three benchmark makeup face datasets demonstrate that our method achieves state-of-the-art verification accuracy across makeup status and can produce photo-realistic non-makeup face images.

Heterogeneous face matching is a challenge issue in face recognition due to large domain difference as well as insufficient pairwise images in different modalities during training. This paper proposes a coupled deep learning (CDL) approach for the heterogeneous face matching. CDL seeks a shared feature space in which the heterogeneous face matching problem can be approximately treated as a homogeneous face matching problem. The objective function of CDL mainly includes two parts. The first part contains a trace norm and a block-diagonal prior as relevance constraints, which not only make unpaired images from multiple modalities be clustered and correlated, but also regularize the parameters to alleviate overfitting. An approximate variational formulation is introduced to deal with the difficulties of optimizing low-rank constraint directly. The second part contains a cross modal ranking among triplet domain specific images to maximize the margin for different identities and increase data for a small amount of training samples. Besides, an alternating minimization method is employed to iteratively update the parameters of CDL. Experimental results show that CDL achieves better performance on the challenging CASIA NIR-VIS 2.0 face recognition database, the IIIT-D Sketch database, the CUHK Face Sketch (CUFS), and the CUHK Face Sketch FERET (CUFSF), which significantly outperforms state-of-the-art heterogeneous face recognition methods.

Identifying event instance in text plays a critical role in building NLP applications such as Information Extraction (IE) system. However, most existing methods for this task focus only on monolingual clues of a specific language and ignore the massive information provided by other languages. Data scarcity and monolingual ambiguity hinder the performance of these monolingual approaches. In this paper, we propose a novel multilingual approach---dubbed as Gated Multilingual Attention (GMLATT) framework---to address the two issues simultaneously. In specific, to alleviate data scarcity problem, we exploit the consistent information in multilingual data via context attention mechanism. Which takes advantage of the consistent evidence in multilingual data other than learning only from monolingual data. To deal with monolingual ambiguity problem, we propose gated cross-lingual attention to exploit the complement information conveyed by multilingual data, which is helpful for the disambiguation. The cross-lingual attention gate serves as a sentinel modelling the confidence of the clues provided by other languages and controls the information integration of various languages. We have conducted extensive experiments on the ACE 2005 benchmark. Experimental results show that our approach significantly outperforms state-of-the-art methods.

Cross-domain data reconstruction methods derive a shared transformation across source and target domains. These methods usually make a specific assumption on noise, which exhibits limited ability when the target data are contaminated by different kinds of complex noise in practice. To enhance the robustness of domain adaptation under severe noise conditions, this paper proposes a novel reconstruction based algorithm in an information-theoretic setting. Specifically, benefiting from the theoretical property of correntropy, the proposed algorithm is distinguished with: detecting the contaminated target samples without making any specific assumption on noise; greatly suppressing the negative influence of noise on cross-domain transformation. Moreover, a relative entropy based regularization of the transformation is incorporated to avoid trivial solutions with the reaped theoretic advantages, i.e., non-negativity and scale-invariance. For optimization, a half-quadratic technique is developed to minimize the non-convex information-theoretic objectives with explicitly guaranteed convergence. Experiments on two real-world domain adaptation tasks demonstrate the superiority of our method.

Multimodal models have been proven to outperform text-based approaches on learning semantic representations. However, it still remains unclear what properties are encoded in multimodal representations, in what aspects do they outperform the single-modality representations, and what happened in the process of semantic compositionality in different input modalities. Considering that multimodal models are originally motivated by human concept representations, we assume that correlating multimodal representations with brain-based semantics would interpret their inner properties to answer the above questions. To that end, we propose simple interpretation methods based on brain-based componential semantics. First we investigate the inner properties of multimodal representations by correlating them with corresponding brain-based property vectors. Then we map the distributed vector space to the interpretable brain-based componential space to explore the inner properties of semantic compositionality. Ultimately, the present paper sheds light on the fundamental questions of natural language understanding, such as how to represent the meaning of words and how to combine word meanings into larger units.

Many efforts have been taken to train spiking neural networks (SNNs), but most of them still need improvements due to the discontinuous and non-differential characteristics of SNNs. While the mammalian brains solve these kinds of problems by integrating a series of biological plasticity learning rules. In this paper, we will focus on two biological plausible methodologies and try to solve these catastrophic training problems in SNNs. Firstly, the biological neural network will try to keep a balance between inputs and outputs on both the neuron and the network levels. Secondly, the biological synaptic weights will be passively updated by the changes of the membrane potentials of the neighbour-hood neurons, and the plasticity of synapses will not propagate back to other previous layers. With these biological inspirations, we propose Voltage-driven Plasticity-centric SNN (VPSNN), which includes four steps, namely: feed forward inference, unsupervised equilibrium state learning, supervised last layer learning and passively updating synaptic weights based on spike-timing dependent plasticity (STDP). Finally we get the accuracy of 98.52% on the hand-written digits classification task on MNIST. In addition, with the help of a visualization tool, we try to analyze the black box of SNN and get better understanding of what benefits have been acquired by the proposed method.

Deep convolutional neural networks (CNNs) have shown appealing performance on various computer vision tasks in recent years. This motivates people to deploy CNNs to real-world applications. However, most of state-of-art CNNs require large memory and computational resources, which hinders the deployment on mobile devices. Recent studies show that low-bit weight representation can reduce much storage and memory demand, and also can achieve efficient network inference. To achieve this goal, we propose a novel approach named BWNH to train Binary Weight Networks via Hashing. In this paper, we first reveal the strong connection between inner-product preserving hashing and binary weight networks, and show that training binary weight networks can be intrinsically regarded as a hashing problem. Based on this perspective, we propose an alternating optimization method to learn the hash codes instead of directly learning binary weights. Extensive experiments on CIFAR10, CIFAR100 and ImageNet demonstrate that our proposed BWNH outperforms current state-of-art by a large margin.