The existing image captioning approaches typically train a one-stage sentence decoder, which is difficult to generate rich fine-grained descriptions. On the other hand, multi-stage image caption model is hard to train due to the vanishing gradient problem. In this paper, we propose a coarse-to-fine multi-stage prediction framework for image captioning, composed of multiple decoders each of which operates on the output of the previous stage, producing increasingly refined image descriptions. Our proposed learning approach addresses the difficulty of vanishing gradients during training by providing a learning objective function that enforces intermediate supervisions. Particularly, we optimize our model with a reinforcement learning approach which utilizes the output of each intermediate decoder's test-time inference algorithm as well as the output of its preceding decoder to normalize the rewards, which simultaneously solves the well-known exposure bias problem and the loss-evaluation mismatch problem. We extensively evaluate the proposed approach on MSCOCO and show that our approach can achieve the state-of-the-art performance.

Deep neural network (DNN) architecture based models have high expressive power and learning capacity. However, they are essentially a black box method since it is not easy to mathematically formulate the functions that are learned within its many layers of representation. Realizing this, many researchers have started to design methods to exploit the drawbacks of deep learning based algorithms questioning their robustness and exposing their singularities. In this paper, we attempt to unravel three aspects related to the robustness of DNNs for face recognition: (i) assessing the impact of deep architectures for face recognition in terms of vulnerabilities to attacks inspired by commonly observed distortions in the real world that are well handled by shallow learning methods along with learning based adversaries; (ii) detecting the singularities by characterizing abnormal filter response behavior in the hidden layers of deep networks; and (iii) making corrections to the processing pipeline to alleviate the problem. Our experimental evaluation using multiple open-source DNN-based face recognition networks, including OpenFace and VGG-Face, and two publicly available databases (MEDS and PaSC) demonstrates that the performance of deep learning based face recognition algorithms can suffer greatly in the presence of such distortions. The proposed method is also compared with existing detection algorithms and the results show that it is able to detect the attacks with very high accuracy by suitably designing a classifier using the response of the hidden layers in the network. Finally, we present several effective countermeasures to mitigate the impact of adversarial attacks and improve the overall robustness of DNN-based face recognition.

In recent years considerable research efforts have been devoted to compression techniques of convolutional neural networks (CNNs). Many works so far have focused on CNN connection pruning methods which produce sparse parameter tensors in convolutional or fully-connected layers. It has been demonstrated in several studies that even simple methods can effectively eliminate connections of a CNN. However, since these methods make parameter tensors just sparser but no smaller, the compression may not transfer directly to acceleration without support from specially designed hardware. In this paper, we propose an iterative approach named Auto-balanced Filter Pruning, where we pre-train the network in an innovative auto-balanced way to transfer the representational capacity of its convolutional layers to a fraction of the filters, prune the redundant ones, then re-train it to restore the accuracy. In this way, a smaller version of the original network is learned and the floating-point operations (FLOPs) are reduced. By applying this method on several common CNNs, we show that a large portion of the filters can be discarded without obvious accuracy drop, leading to significant reduction of computational burdens. Concretely, we reduce the inference cost of LeNet-5 on MNIST, VGG-16 and ResNet-56 on CIFAR-10 by 95.1%, 79.7% and 60.9%, respectively.

Narrated 360° videos are typically provided in many touring scenarios to mimic real-world experience. However, previous work has shown that smart assistance (i.e., providing visual guidance) can significantly help users to follow the Normal Field of View (NFoV) corresponding to the narrative.In this project, we aim at automatically grounding the NFoVs of a 360° video given subtitles of the narrative (referred to as ''NFoV-grounding"). We propose a novel Visual Grounding Model (VGM) to implicitly and efficiently predict the NFoVs given the video content and subtitles. Specifically, at each frame, we efficiently encode the panorama into feature map of candidate NFoVs using a Convolutional Neural Network (CNN) and the subtitles to the same hidden space using an RNN with Gated Recurrent Units (GRU). Then, we apply soft-attention on candidate NFoVs to trigger sentence decoder aiming to minimize the reconstruct loss between the generated and given sentence. Finally, we obtain the NFoV as the candidate NFoV with the maximum attention without any human supervision.To train VGM more robustly, we also generate a reverse sentence conditioning on one minus the soft-attention such that the attention focuses on candidate NFoVs less relevant to the given sentence. The negative log reconstruction loss of the reverse sentence (referred to as ''irrelevant loss") is jointly minimized to encourage the reverse sentence to be different from the given sentence. To evaluate our method, we collect the first narrated 360° videos dataset and achieve state-of-the-art NFoV-grounding performance.

Accelerating deep neural networks (DNNs) has been attracting increasing attention as it can benefit a wide range of applications, e.g., enabling mobile systems with limited computing resources to own powerful visual recognition ability. A practical strategy to this goal usually relies on a two-stage process: operating on the trained DNNs (e.g., approximating the convolutional filters with tensor decomposition) and fine-tuning the amended network, leading to difficulty in balancing the trade-off between acceleration and maintaining recognition performance. In this work, aiming at a general and comprehensive way for neural network acceleration, we develop a Wavelet-like Auto-Encoder (WAE) that decomposes the original input image into two low-resolution channels (sub-images) and incorporate the WAE into the classification neural networks for joint training. The two decomposed channels, in particular, are encoded to carry the low-frequency information (e.g., image profiles) and high-frequency (e.g., image details or noises), respectively, and enable reconstructing the original input image through the decoding process. Then, we feed the low-frequency channel into a standard classification network such as VGG or ResNet and employ a very lightweight network to fuse with the high-frequency channel to obtain the classification result. Compared to existing DNN acceleration solutions, our framework has the following advantages: i) it is tolerant to any existing convolutional neural networks for classification without amending their structures; ii) the WAE provides an interpretable way to preserve the main components of the input image for classification.

While a number of research works (Zhang and Zhou 2006; Nam et al. 2014; Gong et al. 2013; Wei et al. 2014; We propose a recurrent neural network (RNN) based model Wang et al. 2016) start to advance the CNN architecture for image multi-label classification. Our model uniquely integrates for multi-label classification, CNN-RNN (Wang et al. and learning of visual attention and Long Short 2016) embeds image and semantic structures by projecting Term Memory (LSTM) layers, which jointly learns the labels both features into a joint embedding space. By further of interest and their co-occurrences, while the associated utilizing the component of Long Short Term Memory image regions are visually attended. Different from existing (LSTM) (Hochreiter and Schmidhuber 1997), a recurrent approaches utilize either model in their network architectures, neural network (RNN) structure is introduced to memorize training of our model does not require predefined long-term label dependency. As a result, CNN-RNN exhibits label orders. Moreover, a robust inference process is introduced promising multi-label classification performance with crosslabel so that prediction errors would not propagate and thus correlation implicitly preserved.

Lateral inhibition in top-down feedback is widely existing in visual neurobiology, but such an important mechanism has not be well explored yet in computer vision. In our recent research, we find that modeling lateral inhibition in convolutional neural network (LICNN) is very useful for visual attention and saliency detection. In this paper, we propose to formulate lateral inhibition inspired by the related studies from neurobiology, and embed it into the top-down gradient computation of a general CNN for classification, i.e. only category-level information is used. After this operation (only conducted once), the network has the ability to generate accurate category-specific attention maps. Further, we apply LICNN for weakly-supervised salient object detection.Extensive experimental studies on a set of databases, e.g., ECSSD, HKU-IS, PASCAL-S and DUT-OMRON, demonstrate the great advantage of LICNN which achieves the state-of-the-art performance. It is especially impressive that LICNN with only category-level supervised information even outperforms some recent methods with segmentation-level supervised learning.

The existing methods for image captioning usually train the language model under the cross entropy loss, which results in the exposure bias and inconsistency of evaluation metric. Recent research has shown these two issues can be well addressed by policy gradient method in reinforcement learning domain attributable to its unique capability of directly optimizing the discrete and non-differentiable evaluation metric. In this paper, we utilize reinforcement learning method to train the image captioning model. Specifically, we train our image captioning model to maximize the overall reward of the sentences by adopting the temporal-difference (TD) learning method, which takes the correlation between temporally successive actions into account. In this way, we assign different values to different words in one sampled sentence by a discounted coefficient when back-propagating the gradient with the REINFORCE algorithm, enabling the correlation between actions to be learned. Besides, instead of estimating a "baseline" to normalize the rewards with another network, we utilize the reward of another Monte-Carlo sample as the "baseline" to avoid high variance. We show that our proposed method can improve the quality of generated captions and outperforms the state-of-the-art methods on the benchmark dataset MS COCO in terms of seven evaluation metrics.

Heterogeneous face recognition (HFR) refers to matching a probe face image taken from one modality to face images acquired from another modality. It plays an important role in security scenarios. However, HFR is still a challenging problem due to great discrepancies between cross-modality images. This paper proposes an asymmetric joint learning (AJL) approach to handle this issue. The proposed method transforms the cross-modality differences mutually by incorporating the synthesized images into the learning process which provides more discriminative information. Although the aggregated data would augment the scale of intra-classes, it also reduces the diversity (i.e. discriminative information) for inter-classes. Then, we develop the AJL model to balance this dilemma. Finally, we could obtain the similarity score between two heterogeneous face images through the log-likelihood ratio. Extensive experiments on viewed sketch database, forensic sketch database and near infrared image database illustrate that the proposed AJL-HFR method achieve superior performance in comparison to state-of-the-art methods.

Propositional satisfiability (SAT) is at the nucleus of state-of-the-art approaches to a variety of computationally hard problems, one of which is cryptanalysis. Moreover, a number of practical applications of SAT can only be tackled efficiently by identifying and exploiting a subset of formula's variables called backdoor set (or simply backdoors). This paper proposes a new class of backdoor sets for SAT used in the context of cryptographic attacks, namely guess-and-determine attacks. The idea is to identify the best set of backdoor variables subject to a statistically estimated hardness of the guess-and-determine attack using a SAT solver. Experimental results on weakened variants of the renowned encryption algorithms exhibit advantage of the proposed approach compared to the state of the art in terms of the estimated hardness of the resulting guess-and-determine attacks.