Image captioning has received significant attention with remarkable improvements in recent advances. Nevertheless, images in the wild encapsulate rich knowledge and cannot be sufficiently described with models built on image-caption pairs containing only in-domain objects. In this paper, we propose to address the problem by augmenting standard deep captioning architectures with object learners. Specifically, we present Long Short-Term Memory with Pointing (LSTM-P) --- a new architecture that facilitates vocabulary expansion and produces novel objects via pointing mechanism. Technically, object learners are initially pre-trained on available object recognition data. Pointing in LSTM-P then balances the probability between generating a word through LSTM and copying a word from the recognized objects at each time step in decoder stage. Furthermore, our captioning encourages global coverage of objects in the sentence. Extensive experiments are conducted on both held-out COCO image captioning and ImageNet datasets for describing novel objects, and superior results are reported when comparing to state-of-the-art approaches. More remarkably, we obtain an average of 60.9% in F1 score on held-out COCO~dataset.
Machine translation systems are conventionally trained on textual resources that do not model phenomena that occur in spoken language. While the evaluation of neural machine translation systems on textual inputs is actively researched in the literature , little has been discovered about the complexities of translating spoken language data with neural models. We introduce and motivate interesting problems one faces when considering the translation of automatic speech recognition (ASR) outputs on neural machine translation (NMT) systems. We test the robustness of sentence encoding approaches for NMT encoder-decoder modeling, focusing on word-based over byte-pair encoding. We compare the translation of utterances containing ASR errors in state-of-the-art NMT encoder-decoder systems against a strong phrase-based machine translation baseline in order to better understand which phenomena present in ASR outputs are better represented under the NMT framework than approaches that represent translation as a linear model.
We present a transductive deep learning-based formulation for the sparse representation-based classification (SRC) method. The proposed network consists of a convolutional autoencoder along with a fully-connected layer. The role of the autoencoder network is to learn robust deep features for classification. On the other hand, the fully-connected layer, which is placed in between the encoder and the decoder networks, is responsible for finding the sparse representation. The estimated sparse codes are then used for classification. Various experiments on three different datasets show that the proposed network leads to sparse representations that give better classification results than state-of-the-art SRC methods. The source code is available at: github.com/mahdiabavisani/DSRC.
Predicting the future is an important aspect for decision-making in robotics or autonomous driving systems, which heavily rely upon visual scene understanding. While prior work attempts to predict future video pixels, anticipate activities or forecast future scene semantic segments from segmentation of the preceding frames, methods that predict future semantic segmentation solely from the previous frame RGB data in a single end-to-end trainable model do not exist. In this paper, we propose a temporal encoder-decoder network architecture that encodes RGB frames from the past and decodes the future semantic segmentation. The network is coupled with a new knowledge distillation training framework specific for the forecasting task. Our method, only seeing preceding video frames, implicitly models the scene segments while simultaneously accounting for the object dynamics to infer the future scene semantic segments. Our results on Cityscapes and Apolloscape outperform the baseline and current state-of-the-art methods. Code is available at https://github.com/eddyhkchiu/segmenting_the_future/.
Connectionist Temporal Classification (CTC) based end-to-end speech recognition system usually need to incorporate an external language model by using WFST-based decoding in order to achieve promising results. This is more essential to Mandarin speech recognition since it owns a special phenomenon, namely homophone, which causes a lot of substitution errors. The linguistic information introduced by language model will help to distinguish these substitution errors. In this work, we propose a transformer based spelling correction model to automatically correct errors especially the substitution errors made by CTC-based Mandarin speech recognition system. Specifically, we investigate using the recognition results generated by CTC-based systems as input and the ground-truth transcriptions as output to train a transformer with encoder-decoder architecture, which is much similar to machine translation. Results in a 20,000 hours Mandarin speech recognition task show that the proposed spelling correction model can achieve a CER of 3.41%, which results in 22.9% and 53.2% relative improvement compared to the baseline CTC-based systems decoded with and without language model respectively.
Neural network-based Open-ended conversational agents automatically generate responses based on predictive models learned from a large number of pairs of utterances. The generated responses are typically acceptable as a sentence but are often dull, generic, and certainly devoid of any emotion. In this paper, we present neural models that learn to express a given emotion in the generated response. We propose four models and evaluate them against 3 baselines. An encoder-decoder framework-based model with multiple attention layers provides the best overall performance in terms of expressing the required emotion. While it does not outperform other models on all emotions, it presents promising results in most cases.
Recommender systems play an essential role in music streaming services, prominently in the form of personalized playlists. Exploring the user interactions within these listening sessions can be beneficial to understanding the user preferences in the context of a single session. In the 'Spotify Sequential Skip Prediction Challenge', WSDM, and Spotify are challenging people to understand the way users sequentially interact with music. We describe our solution approach in this paper and also state proposals for further improvements to the model. The proposed model initially generates a fixed vector representation of the session, and this additional information is incorporated into an Encoder-Decoder style architecture. This method achieved the seventh position in the competition, with a mean average accuracy of 0.604 on the test set. The solution code is available at https://github.com/sainathadapa/spotify-sequential-skip-prediction.
Capsule Network is a promising concept in deep learning, yet its true potential is not fully realized thus far, providing sub-par performance on several key benchmark datasets with complex data. Drawing intuition from the success achieved by Convolutional Neural Networks (CNNs) by going deeper, we introduce DeepCaps1, a deep capsule network architecture which uses a novel 3D convolution based dynamic routing algorithm. With DeepCaps, we surpass the state-of-the-art results in the capsule network domain on CIFAR10, SVHN and Fashion MNIST, while achieving a 68% reduction in the number of parameters. Further, we propose a class-independent decoder network, which strengthens the use of reconstruction loss as a regularization term. This leads to an interesting property of the decoder, which allows us to identify and control the physical attributes of the images represented by the instantiation parameters.
We present the workflow of a DSM refinement methodology using a Hybrid-cGAN where the generative part consists of two encoders and a common decoder which blends the spectral and height information within one network. The inputs to the Hybrid-cGAN are single-channel photogrammetric DSMs with continuous values and single-channel pan-chromatic (PAN) half-meter resolution satellite images. Experimental results demonstrate that the earlier information fusion from data with different physical meanings helps to propagate fine details and complete an inaccurate or missing 3D information about building forms. Moreover, it improves the building boundaries making them more rectilinear.
Arbitrary attribute editing generally can be tackled by incorporating encoder-decoder and generative adversarial networks. However, the bottleneck layer in encoder-decoder usually gives rise to blurry and low quality editing result. And adding skip connections improves image quality at the cost of weakened attribute manipulation ability. Moreover, existing methods exploit target attribute vector to guide the flexible translation to desired target domain. In this work, we suggest to address these issues from selective transfer perspective. Considering that specific editing task is certainly only related to the changed attributes instead of all target attributes, our model selectively takes the difference between target and source attribute vectors as input. Furthermore, selective transfer units are incorporated with encoder-decoder to adaptively select and modify encoder feature for enhanced attribute editing. Experiments show that our method (i.e., STGAN) simultaneously improves attribute manipulation accuracy as well as perception quality, and performs favorably against state-of-the-arts in arbitrary facial attribute editing and season translation.
We present a fully convolutional wav-to-wav network for converting between speakers' voices, without relying on text. Our network is based on an encoder-decoder architecture, where the encoder is pre-trained for the task of Automatic Speech Recognition, and a multi-speaker waveform decoder is trained to reconstruct the original signal in an autoregressive manner. We train the network on narrated audiobooks, and demonstrate multi-voice TTS in those voices, by converting the voice of a TTS robot.
Haze degrades content and obscures information of images, which can negatively impact vision-based decision-making in real-time systems. In this paper, we propose an efficient fully convolutional neural network (CNN) image dehazing method designed to run on edge graphical processing units (GPUs). We utilize three variants of our architecture to explore the dependency of dehazed image quality on parameter count and model design. The first two variants presented, a small and big version, make use of a single efficient encoder-decoder convolutional feature extractor. The final variant utilizes a pair of encoder-decoders for atmospheric light and transmission map estimation. Each variant ends with an image refinement pyramid pooling network to form the final dehazed image. For the big variant of the single-encoder network, we demonstrate state-of-the-art performance on the NYU Depth dataset. For the small variant, we maintain competitive performance on the super-resolution O/I-HAZE datasets without the need for image cropping. Finally, we examine some challenges presented by the Dense-Haze dataset when leveraging CNN architectures for dehazing of dense haze imagery and examine the impact of loss function selection on image quality. Benchmarks are included to show the feasibility of introducing this approach into real-time systems.
Existing joint optic disc and cup segmentation approaches are developed either in Cartesian or polar coordinate system. However, due to the subtle optic cup, the contextual information exploited from the single domain even by the prevailing CNNs is still insufficient. In this paper, we propose a novel segmentation approach, named Cartesian-polar dual-domain network (DDNet), which for the first time considers the complementary of the Cartesian domain and the polar domain. We propose a two-branch of domain feature encoder and learn translation equivariant representations on rectilinear grid from Cartesian domain and rotation equivariant representations on polar grid from polar domain parallelly. To fuse the features on two different grids, we propose a dual-domain fusion module. This module builds the correspondence between two grids by the differentiable polar transform layer and learns the feature importance across two domains in element-wise to enhance the expressive capability. Finally, the decoder aggregates the fused features from low-level to high-level and makes dense predictions. We validate the state-of-the-art segmentation performances of our DDNet on the public dataset ORIGA. According to the segmentation masks, we estimate the commonly used clinical measure for glaucoma, i.e., the vertical cup-to-disc ratio. The low cup-to-disc ratio estimation error demonstrates the potential application in glaucoma screening.
Existing state-of-the-art salient object detection networks rely on aggregating multi-level features of pre-trained convolutional neural networks (CNNs). Compared to high-level features, low-level features contribute less to performance but cost more computations because of their larger spatial resolutions. In this paper, we propose a novel Cascaded Partial Decoder (CPD) framework for fast and accurate salient object detection. On the one hand, the framework constructs partial decoder which discards larger resolution features of shallower layers for acceleration. On the other hand, we observe that integrating features of deeper layers obtain relatively precise saliency map. Therefore we directly utilize generated saliency map to refine the features of backbone network. This strategy efficiently suppresses distractors in the features and significantly improves their representation ability. Experiments conducted on five benchmark datasets exhibit that the proposed model not only achieves state-of-the-art performance but also runs much faster than existing models. Besides, the proposed framework is further applied to improve existing multi-level feature aggregation models and significantly improve their efficiency and accuracy.
We do not speak word by word from scratch; our brain quickly structures a pattern like \textsc{sth do sth at someplace} and then fill in the detailed descriptions. To render existing encoder-decoder image captioners such human-like reasoning, we propose a novel framework: learning to Collocate Neural Modules (CNM), to generate the `inner pattern' connecting visual encoder and language decoder. Unlike the widely-used neural module networks in visual Q\&A, where the language (ie, question) is fully observable, CNM for captioning is more challenging as the language is being generated and thus is partially observable. To this end, we make the following technical contributions for CNM training: 1) compact module design --- one for function words and three for visual content words (eg, noun, adjective, and verb), 2) soft module fusion and multi-step module execution, robustifying the visual reasoning in partial observation, 3) a linguistic loss for module controller being faithful to part-of-speech collocations (eg, adjective is before noun). Extensive experiments on the challenging MS-COCO image captioning benchmark validate the effectiveness of our CNM image captioner. In particular, CNM achieves a new state-of-the-art 127.9 CIDEr-D on Karpathy split and a single-model 126.0 c40 on the official server. CNM is also robust to few training samples, eg, by training only one sentence per image, CNM can halve the performance loss compared to a strong baseline.
Recent research in the design of end to end communication system using deep learning has produced models which can outperform traditional communication schemes. Most of these architectures leveraged autoencoders to design the encoder at the transmitter and decoder at the receiver and train them jointly by modeling transmit symbols as latent codes from the encoder. However, in communication systems, the receiver has to work with noise corrupted versions of transmit symbols. Traditional autoencoders are not designed to work with latent codes corrupted with noise. In this work, we provide a framework to design end to end communication systems which accounts for the existence of noise corrupted transmit symbols. The proposed method uses deep neural architecture. An objective function for optimizing these models is derived based on the concepts of variational inference. Further, domain knowledge such as channel type can be systematically integrated into the objective. Through numerical simulation, the proposed method is shown to consistently produce models with better packing density and achieving it faster in multiple popular channel models as compared to the previous works leveraging deep learning models.
This article reports on an investigation of the use of convolutional neural networks to predict the visual attention of chess players. The visual attention model described in this article has been created to generate saliency maps that capture hierarchical and spatial features of chessboard, in order to predict the probability fixation for individual pixels Using a skip-layer architecture of an autoencoder, with a unified decoder, we are able to use multiscale features to predict saliency of part of the board at different scales, showing multiple relations between pieces. We have used scan path and fixation data from players engaged in solving chess problems, to compute 6600 saliency maps associated to the corresponding chess piece configurations. This corpus is completed with synthetically generated data from actual games gathered from an online chess platform. Experiments realized using both scan-paths from chess players and the CAT2000 saliency dataset of natural images, highlights several results. Deep features, pretrained on natural images, were found to be helpful in training visual attention prediction for chess. The proposed neural network architecture is able to generate meaningful saliency maps on unseen chess configurations with good scores on standard metrics. This work provides a baseline for future work on visual attention prediction in similar contexts.
Deep image completion usually fails to harmonically blend the restored image into existing content, especially in the boundary area. This paper handles with this problem from a new perspective of creating a smooth transition and proposes a concise Deep Fusion Network (DFNet). Firstly, a fusion block is introduced to generate a flexible alpha composition map for combining known and unknown regions. The fusion block not only provides a smooth fusion between restored and existing content, but also provides an attention map to make network focus more on the unknown pixels. In this way, it builds a bridge for structural and texture information, so that information can be naturally propagated from known region into completion. Furthermore, fusion blocks are embedded into several decoder layers of the network. Accompanied by the adjustable loss constraints on each layer, more accurate structure information are achieved. We qualitatively and quantitatively compare our method with other state-of-the-art methods on Places2 and CelebA datasets. The results show the superior performance of DFNet, especially in the aspects of harmonious texture transition, texture detail and semantic structural consistency. Our source code will be avaiable at: \url{https://github.com/hughplay/DFNet}
Counting people or objects with significantly varying scales and densities has attracted much interest from the research community and yet it remains an open problem. In this paper, we propose a simple but an efficient and effective network, named DENet, which is composed of two components, i.e., a detection network (DNet) and an encoder-decoder estimation network (ENet). We first run DNet on an input image to detect and count individuals who can be segmented clearly. Then, ENet is utilized to estimate the density maps of the remaining areas, where the numbers of individuals cannot be detected. We propose a modified Xception as an encoder for feature extraction and a combination of dilated convolution and transposed convolution as a decoder. In the ShanghaiTech Part A, UCF and WorldExpo'10 datasets, our DENet achieves lower Mean Absolute Error (MAE) than those of the state-of-the-art methods.
We propose a data-driven 3D shape design method that can learn a generative model from a corpus of existing designs, and use this model to produce a wide range of new designs. The approach learns an encoding of the samples in the training corpus using an unsupervised variational autoencoder-decoder architecture, without the need for an explicit parametric representation of the original designs. To facilitate the generation of smooth final surfaces, we develop a 3D shape representation based on a distance transformation of the original 3D data, rather than using the commonly utilized binary voxel representation. Once established, the generator maps the latent space representations to the high-dimensional distance transformation fields, which are then automatically surfaced to produce 3D representations amenable to physics simulations or other objective function evaluation modules. We demonstrate our approach for the computational design of gliders that are optimized to attain prescribed performance scores. Our results show that when combined with genetic optimization, the proposed approach can generate a rich set of candidate concept designs that achieve prescribed functional goals, even when the original dataset has only a few or no solutions that achieve these goals.