In this paper, we propose a novel application of Generative Adversarial Networks (GAN) to the synthesis of cells imaged by fluorescence microscopy. Compared to natural images, cells tend to have a simpler and more geometric global structure that facilitates image generation. However, the correlation between the spatial pattern of different fluorescent proteins reflects important biological functions, and synthesized images have to capture these relationships to be relevant for biological applications. We adapt GANs to the task at hand and propose new models with casual dependencies between image channels that can generate multi-channel images, which would be impossible to obtain experimentally. We evaluate our approach using two independent techniques and compare it against sensible baselines. Finally, we demonstrate that by interpolating across the latent space we can mimic the known changes in protein localization that occur through time during the cell cycle, allowing us to predict temporal evolution from static images.
https://arxiv.org/abs/1708.04692
The photovoltaic properties of (0001) n-InGaN/p-GaN single heterojunctions were investigated numerically and compared with those of conventional p-GaN/i-InGaN/n-GaN structures, employing realistic material parameters. This alternative device architecture exploits the large polarization fields, and high efficiency modules are achieved for In-rich, partially relaxed and coherently strained InGaN films. Conversion efficiencies up to 14% under AM1.5G illumination can be reached, revealing the true potential of InGaN single junction solar cells with proper design.
https://arxiv.org/abs/1709.03729
We demonstrate an InAlN/GaN-on-Si HEMT based UV detector with photo to dark current ratio > 107. Ti/Al/Ni/Au metal stack was evaporated and rapid thermal annealed for Ohmic contacts to the 2D electron gas (2DEG) at the InAlN/GaN interface while the channel + barrier was recess etched to a depth of 20 nm to pinch-off the 2DEG between Source-Drain pads. Spectral responsivity (SR) of 34 A/W at 367 nm was measured at 5 V in conjunction with very high photo to dark current ratio of > 10^7. The photo to dark current ratio at a fixed bias was found to be decreasing with increase in recess length of the PD. The fabricated devices were found to exhibit a UV-to-visible rejection ratio of >103 with a low dark current < 32 pA at 5 V. Transient measurements showed rise and fall times in the range of 3-4 ms. The gain mechanism was investigated and carrier lifetimes were estimated which matched well with those reported elsewhere.
https://arxiv.org/abs/1709.03692
This paper focuses on best-arm identification in multi-armed bandits with bounded rewards. We develop an algorithm that is a fusion of lil-UCB and KL-LUCB, offering the best qualities of the two algorithms in one method. This is achieved by proving a novel anytime confidence bound for the mean of bounded distributions, which is the analogue of the LIL-type bounds recently developed for sub-Gaussian distributions. We corroborate our theoretical results with numerical experiments based on the New Yorker Cartoon Caption Contest.
https://arxiv.org/abs/1709.03570
Most extractive summarization methods focus on the main body of the document from which sentences need to be extracted. However, the gist of the document may lie in side information, such as the title and image captions which are often available for newswire articles. We propose to explore side information in the context of single-document extractive summarization. We develop a framework for single-document summarization composed of a hierarchical document encoder and an attention-based extractor with attention over side information. We evaluate our model on a large scale news dataset. We show that extractive summarization with side information consistently outperforms its counterpart that does not use any side information, in terms of both informativeness and fluency.
https://arxiv.org/abs/1704.04530
Grid maps are widely used in robotics to represent obstacles in the environment and differentiating dynamic objects from static infrastructure is essential for many practical applications. In this work, we present a methods that uses a deep convolutional neural network (CNN) to infer whether grid cells are covering a moving object or not. Compared to tracking approaches, that use e.g. a particle filter to estimate grid cell velocities and then make a decision for individual grid cells based on this estimate, our approach uses the entire grid map as input image for a CNN that inspects a larger area around each cell and thus takes the structural appearance in the grid map into account to make a decision. Compared to our reference method, our concept yields a performance increase from 83.9% to 97.2%. A runtime optimized version of our approach yields similar improvements with an execution time of just 10 milliseconds.
https://arxiv.org/abs/1709.03139
This paper introduces a novel approach to texture synthesis based on generative adversarial networks (GAN) (Goodfellow et al., 2014). We extend the structure of the input noise distribution by constructing tensors with different types of dimensions. We call this technique Periodic Spatial GAN (PSGAN). The PSGAN has several novel abilities which surpass the current state of the art in texture synthesis. First, we can learn multiple textures from datasets of one or more complex large images. Second, we show that the image generation with PSGANs has properties of a texture manifold: we can smoothly interpolate between samples in the structured noise space and generate novel samples, which lie perceptually between the textures of the original dataset. In addition, we can also accurately learn periodical textures. We make multiple experiments which show that PSGANs can flexibly handle diverse texture and image data sources. Our method is highly scalable and it can generate output images of arbitrary large size.
https://arxiv.org/abs/1705.06566
Forward-looking sonar can capture high resolution images of underwater scenes, but their interpretation is complex. Generic object detection in such images has not been solved, specially in cases of small and unknown objects. In comparison, detection proposal algorithms have produced top performing object detectors in real-world color images. In this work we develop a Convolutional Neural Network that can reliably score objectness of image windows in forward-looking sonar images and by thresholding objectness, we generate detection proposals. In our dataset of marine garbage objects, we obtain 94% recall, generating around 60 proposals per image. The biggest strength of our method is that it can generalize to previously unseen objects. We show this by detecting chain links, walls and a wrench without previous training in such objects. We strongly believe our method can be used for class-independent object detection, with many real-world applications such as chain following and mine detection.
https://arxiv.org/abs/1709.02600
Deep neural networks (DNNs) require very large amounts of computation both for training and for inference when deployed in the field. A common approach to implementing DNNs is to recast the most computationally expensive operations as general matrix multiplication (GEMM). However, as we demonstrate in this paper, there are a great many different ways to express DNN convolution operations using GEMM. Although different approaches all perform the same number of operations, the size of temporary data structures differs significantly. Convolution of an input matrix with dimensions $C \times H \times W$, requires $O(K^2CHW)$ additional space using the classical im2col approach. More recently memory-efficient approaches requiring just $O(KCHW)$ auxiliary space have been proposed. We present two novel GEMM-based algorithms that require just $O(MHW)$ and $O(KW)$ additional space respectively, where $M$ is the number of channels in the result of the convolution. These algorithms dramatically reduce the space overhead of DNN convolution, making it much more suitable for memory-limited embedded systems. Experimental evaluation shows that our low-memory algorithms are just as fast as the best patch-building approaches despite requiring just a fraction of the amount of additional memory. Our low-memory algorithms have excellent data locality which gives them a further edge over patch-building algorithms when multiple cores are used. As a result, our low memory algorithms often outperform the best patch-building algorithms using multiple threads.
https://arxiv.org/abs/1709.03395
Detection of objects in cluttered indoor environments is one of the key enabling functionalities for service robots. The best performing object detection approaches in computer vision exploit deep Convolutional Neural Networks (CNN) to simultaneously detect and categorize the objects of interest in cluttered scenes. Training of such models typically requires large amounts of annotated training data which is time consuming and costly to obtain. In this work we explore the ability of using synthetically generated composite images for training state-of-the-art object detectors, especially for object instance detection. We superimpose 2D images of textured object models into images of real environments at variety of locations and scales. Our experiments evaluate different superimposition strategies ranging from purely image-based blending all the way to depth and semantics informed positioning of the object models into real scenes. We demonstrate the effectiveness of these object detector training strategies on two publicly available datasets, the GMU-Kitchens and the Washington RGB-D Scenes v2. As one observation, augmenting some hand-labeled training data with synthetic examples carefully composed onto scenes yields object detectors with comparable performance to using much more hand-labeled data. Broadly, this work charts new opportunities for training detectors for new objects by exploiting existing object model repositories in either a purely automatic fashion or with only a very small number of human-annotated examples.
https://arxiv.org/abs/1702.07836
Recurrent neural networks (RNNs) represent the state of the art in translation, image captioning, and speech recognition. They are also capable of learning algorithmic tasks such as long addition, copying, and sorting from a set of training examples. We demonstrate that RNNs can learn decryption algorithms – the mappings from plaintext to ciphertext – for three polyalphabetic ciphers (Vigenère, Autokey, and Enigma). Most notably, we demonstrate that an RNN with a 3000-unit Long Short-Term Memory (LSTM) cell can learn the decryption function of the Enigma machine. We argue that our model learns efficient internal representations of these ciphers 1) by exploring activations of individual memory neurons and 2) by comparing memory usage across the three ciphers. To be clear, our work is not aimed at ‘cracking’ the Enigma cipher. However, we do show that our model can perform elementary cryptanalysis by running known-plaintext attacks on the Vigenère and Autokey ciphers. Our results indicate that RNNs can learn algorithmic representations of black box polyalphabetic ciphers and that these representations are useful for cryptanalysis.
https://arxiv.org/abs/1708.07576
We propose an automatic unsupervised cell event detection and classification method, which expands convolutional Long Short-Term Memory (LSTM) neural networks, for cellular events in cell video sequences. Cells in images that are captured from various biomedical applications usually have different shapes and motility, which pose difficulties for the automated event detection in cell videos. Current methods to detect cellular events are based on supervised machine learning and rely on tedious manual annotation from investigators with specific expertise. So that our LSTM network could be trained in an unsupervised manner, we designed it with a branched structure where one branch learns the frequent, regular appearance and movements of objects and the second learns the stochastic events, which occur rarely and without warning in a cell video sequence. We tested our network on a publicly available dataset of densely packed stem cell phase-contrast microscopy images undergoing cell division. This dataset is considered to be more challenging that a dataset with sparse cells. We compared our method to several published supervised methods evaluated on the same dataset and to a supervised LSTM method with a similar design and configuration to our unsupervised method. We used an F1-score, which is a balanced measure for both precision and recall. Our results show that our unsupervised method has a higher or similar F1-score when compared to two fully supervised methods that are based on Hidden Conditional Random Fields (HCRF), and has comparable accuracy with the current best supervised HCRF-based method. Our method was generalizable as after being trained on one video it could be applied to videos where the cells were in different conditions. The accuracy of our unsupervised method approached that of its supervised counterpart.
https://arxiv.org/abs/1709.02081
With the proliferation of e-commerce websites and the ubiquitousness of smart phones, cross-domain image retrieval using images taken by smart phones as queries to search products on e-commerce websites is emerging as a popular application. One challenge of this task is to locate the attention of both the query and database images. In particular, database images, e.g. of fashion products, on e-commerce websites are typically displayed with other accessories, and the images taken by users contain noisy background and large variations in orientation and lighting. Consequently, their attention is difficult to locate. In this paper, we exploit the rich tag information available on the e-commerce websites to locate the attention of database images. For query images, we use each candidate image in the database as the context to locate the query attention. Novel deep convolutional neural network architectures, namely TagYNet and CtxYNet, are proposed to learn the attention weights and then extract effective representations of the images. Experimental results on public datasets confirm that our approaches have significant improvement over the existing methods in terms of the retrieval accuracy and efficiency.
https://arxiv.org/abs/1709.01784
Neural machine translation (NMT), a new approach to machine translation, has achieved promising results comparable to those of traditional approaches such as statistical machine translation (SMT). Despite its recent success, NMT cannot handle a larger vocabulary because the training complexity and decoding complexity proportionally increase with the number of target words. This problem becomes even more serious when translating patent documents, which contain many technical terms that are observed infrequently. In this paper, we propose to select phrases that contain out-of-vocabulary words using the statistical approach of branching entropy. This allows the proposed NMT system to be applied to a translation task of any language pair without any language-specific knowledge about technical term identification. The selected phrases are then replaced with tokens during training and post-translated by the phrase translation table of SMT. Evaluation on Japanese-to-Chinese, Chinese-to-Japanese, Japanese-to-English and English-to-Japanese patent sentence translation proved the effectiveness of phrases selected with branching entropy, where the proposed NMT model achieves a substantial improvement over a baseline NMT model without our proposed technique. Moreover, the number of translation errors of under-translation by the baseline NMT model without our proposed technique reduces to around half by the proposed NMT model.
https://arxiv.org/abs/1704.04520
NMT systems have problems with large vocabulary sizes. Byte-pair encoding (BPE) is a popular approach to solving this problem, but while BPE allows the system to generate any target-side word, it does not enable effective generalization over the rich vocabulary in morphologically rich languages with strong inflectional phenomena. We introduce a simple approach to overcome this problem by training a system to produce the lemma of a word and its morphologically rich POS tag, which is then followed by a deterministic generation step. We apply this strategy for English-Czech and English-German translation scenarios, obtaining improvements in both settings. We furthermore show that the improvement is not due to only adding explicit morphological information.
https://arxiv.org/abs/1707.06012
The ability to semantically interpret hand-drawn line sketches, although very challenging, can pave way for novel applications in multimedia. We propose SketchParse, the first deep-network architecture for fully automatic parsing of freehand object sketches. SketchParse is configured as a two-level fully convolutional network. The first level contains shared layers common to all object categories. The second level contains a number of expert sub-networks. Each expert specializes in parsing sketches from object categories which contain structurally similar parts. Effectively, the two-level configuration enables our architecture to scale up efficiently as additional categories are added. We introduce a router layer which (i) relays sketch features from shared layers to the correct expert (ii) eliminates the need to manually specify object category during inference. To bypass laborious part-level annotation, we sketchify photos from semantic object-part image datasets and use them for training. Our architecture also incorporates object pose prediction as a novel auxiliary task which boosts overall performance while providing supplementary information regarding the sketch. We demonstrate SketchParse’s abilities (i) on two challenging large-scale sketch datasets (ii) in parsing unseen, semantically related object categories (iii) in improving fine-grained sketch-based image retrieval. As a novel application, we also outline how SketchParse’s output can be used to generate caption-style descriptions for hand-drawn sketches.
https://arxiv.org/abs/1709.01295
We introduce a deep residual recurrent neural network (DR-RNN) as an efficient model reduction technique for nonlinear dynamical systems. The developed DR-RNN is inspired by the iterative steps of line search methods in finding the residual minimiser of numerically discretized differential equations. We formulate this iterative scheme as stacked recurrent neural network (RNN) embedded with the dynamical structure of the emulated differential equations. Numerical examples demonstrate that DR-RNN can effectively emulate the full order models of nonlinear physical systems with a significantly lower number of parameters in comparison to standard RNN architectures. Further, we combined DR-RNN with Proper Orthogonal Decomposition (POD) for model reduction of time dependent partial differential equations. The presented numerical results show the stability of proposed DR-RNN as an explicit reduced order technique. We also show significant gains in accuracy by increasing the depth of proposed DR-RNN similar to other applications of deep learning.
https://arxiv.org/abs/1709.00939
Generating video descriptions in natural language (a.k.a. video captioning) is a more challenging task than image captioning as the videos are intrinsically more complicated than images in two aspects. First, videos cover a broader range of topics, such as news, music, sports and so on. Second, multiple topics could coexist in the same video. In this paper, we propose a novel caption model, topic-guided model (TGM), to generate topic-oriented descriptions for videos in the wild via exploiting topic information. In addition to predefined topics, i.e., category tags crawled from the web, we also mine topics in a data-driven way based on training captions by an unsupervised topic mining model. We show that data-driven topics reflect a better topic schema than the predefined topics. As for testing video topic prediction, we treat the topic mining model as teacher to train the student, the topic prediction model, by utilizing the full multi-modalities in the video especially the speech modality. We propose a series of caption models to exploit topic guidance, including implicitly using the topics as input features to generate words related to the topic and explicitly modifying the weights in the decoder with topics to function as an ensemble of topic-aware language decoders. Our comprehensive experimental results on the current largest video caption dataset MSR-VTT prove the effectiveness of our topic-guided model, which significantly surpasses the winning performance in the 2016 MSR video to language challenge.
https://arxiv.org/abs/1708.09666
Due to its efficiency and stability, Robust Principal Component Analysis (RPCA) has been emerging as a promising tool for moving object detection. Unfortunately, existing RPCA based methods assume static or quasi-static background, and thereby they may have trouble in coping with the background scenes that exhibit a persistent dynamic behavior. In this work, we shall introduce two techniques to fill in the gap. First, instead of using the raw pixel-value as features that are brittle in the presence of dynamic background, we devise a so-called Gaussian max-pooling operator to estimate a “stable-value” for each pixel. Those stable-values are robust to various background changes and can therefore distinguish effectively the foreground objects from the background. Then, to obtain more accurate results, we further propose a Segmentation Constrained RPCA (SC-RPCA) model, which incorporates the temporal and spatial continuity in images into RPCA. The inference process of SC-RPCA is a group sparsity constrained nuclear norm minimization problem, which is convex and easy to solve. Experimental results on seven videos from the CDCNET 2014 database show the superior performance of the proposed method.
https://arxiv.org/abs/1709.00657
The topic diversity of open-domain videos leads to various vocabularies and linguistic expressions in describing video contents, and therefore, makes the video captioning task even more challenging. In this paper, we propose an unified caption framework, M&M TGM, which mines multimodal topics in unsupervised fashion from data and guides the caption decoder with these topics. Compared to pre-defined topics, the mined multimodal topics are more semantically and visually coherent and can reflect the topic distribution of videos better. We formulate the topic-aware caption generation as a multi-task learning problem, in which we add a parallel task, topic prediction, in addition to the caption task. For the topic prediction task, we use the mined topics as the teacher to train a student topic prediction model, which learns to predict the latent topics from multimodal contents of videos. The topic prediction provides intermediate supervision to the learning process. As for the caption task, we propose a novel topic-aware decoder to generate more accurate and detailed video descriptions with the guidance from latent topics. The entire learning procedure is end-to-end and it optimizes both tasks simultaneously. The results from extensive experiments conducted on the MSR-VTT and Youtube2Text datasets demonstrate the effectiveness of our proposed model. M&M TGM not only outperforms prior state-of-the-art methods on multiple evaluation metrics and on both benchmark datasets, but also achieves better generalization ability.
https://arxiv.org/abs/1708.09667
Commit messages are a valuable resource in comprehension of software evolution, since they provide a record of changes such as feature additions and bug repairs. Unfortunately, programmers often neglect to write good commit messages. Different techniques have been proposed to help programmers by automatically writing these messages. These techniques are effective at describing what changed, but are often verbose and lack context for understanding the rationale behind a change. In contrast, humans write messages that are short and summarize the high level rationale. In this paper, we adapt Neural Machine Translation (NMT) to automatically “translate” diffs into commit messages. We trained an NMT algorithm using a corpus of diffs and human-written commit messages from the top 1k Github projects. We designed a filter to help ensure that we only trained the algorithm on higher-quality commit messages. Our evaluation uncovered a pattern in which the messages we generate tend to be either very high or very low quality. Therefore, we created a quality-assurance filter to detect cases in which we are unable to produce good messages, and return a warning instead.
https://arxiv.org/abs/1708.09492
The attention model has become a standard component in neural machine translation (NMT) and it guides translation process by selectively focusing on parts of the source sentence when predicting each target word. However, we find that the generation of a target word does not only depend on the source sentence, but also rely heavily on the previous generated target words, especially the distant words which are difficult to model by using recurrent neural networks. To solve this problem, we propose in this paper a novel look-ahead attention mechanism for generation in NMT, which aims at directly capturing the dependency relationship between target words. We further design three patterns to integrate our look-ahead attention into the conventional attention model. Experiments on NIST Chinese-to-English and WMT English-to-German translation tasks show that our proposed look-ahead attention mechanism achieves substantial improvements over state-of-the-art baselines.
https://arxiv.org/abs/1708.09217
In this work, we present a detailed photophysical analysis of recently-discovered optically stable, single photon emitters (SPEs) in Gallium Nitride (GaN). Temperature-resolved photoluminescence measurements reveal that the emission lines at 4 K are three orders of magnitude broader than the transform-limited widths expected from excited state lifetime measurements. The broadening is ascribed to ultra-fast spectral diffusion. Continuing the photophysics study on several emitters at room temperature (RT), a maximum average brightness of ~427 kCounts/s is measured. Furthermore, by determining the decay rates of emitters undergoing three-level optical transitions, radiative and non-radiative lifetimes are calculated at RT. Finally, polarization measurements from 14 emitters are used to determine visibility as well as dipole orientation of defect systems within the GaN crystal. Our results underpin some of the fundamental properties of SPE in GaN both at cryogenic and RT, and define the benchmark for future work in GaN-based single-photon technologies.
https://arxiv.org/abs/1708.09161
Taking an image and question as the input of our method, it can output the text-based answer of the query question about the given image, so called Visual Question Answering (VQA). There are two main modules in our algorithm. Given a natural language question about an image, the first module takes the question as input and then outputs the basic questions of the main given question. The second module takes the main question, image and these basic questions as input and then outputs the text-based answer of the main question. We formulate the basic questions generation problem as a LASSO optimization problem, and also propose a criterion about how to exploit these basic questions to help answer main question. Our method is evaluated on the challenging VQA dataset and yields state-of-the-art accuracy, 60.34% in open-ended task.
https://arxiv.org/abs/1703.06492
In neural image captioning systems, a recurrent neural network (RNN) is typically viewed as the primary generation' component. This view suggests that the image features should be injected’ into the RNN. This is in fact the dominant view in the literature. Alternatively, the RNN can instead be viewed as only encoding the previously generated words. This view suggests that the RNN should only be used to encode linguistic features and that only the final representation should be `merged’ with the image features at a later stage. This paper compares these two architectures. We find that, in general, late merging outperforms injection, suggesting that RNNs are better viewed as encoders, rather than generators.
https://arxiv.org/abs/1708.02043
We propose “Areas of Attention”, a novel attention-based model for automatic image captioning. Our approach models the dependencies between image regions, caption words, and the state of an RNN language model, using three pairwise interactions. In contrast to previous attention-based approaches that associate image regions only to the RNN state, our method allows a direct association between caption words and image regions. During training these associations are inferred from image-level captions, akin to weakly-supervised object detector training. These associations help to improve captioning by localizing the corresponding regions during testing. We also propose and compare different ways of generating attention areas: CNN activation grids, object proposals, and spatial transformers nets applied in a convolutional fashion. Spatial transformers give the best results. They allow for image specific attention areas, and can be trained jointly with the rest of the network. Our attention mechanism and spatial transformer attention areas together yield state-of-the-art results on the MSCOCO dataset.o meaningful latent semantic structure in the generated captions.
https://arxiv.org/abs/1612.01033
The original ImageNet dataset is a popular large-scale benchmark for training Deep Neural Networks. Since the cost of performing experiments (e.g, algorithm design, architecture search, and hyperparameter tuning) on the original dataset might be prohibitive, we propose to consider a downsampled version of ImageNet. In contrast to the CIFAR datasets and earlier downsampled versions of ImageNet, our proposed ImageNet32$\times$32 (and its variants ImageNet64$\times$64 and ImageNet16$\times$16) contains exactly the same number of classes and images as ImageNet, with the only difference that the images are downsampled to 32$\times$32 pixels per image (64$\times$64 and 16$\times$16 pixels for the variants, respectively). Experiments on these downsampled variants are dramatically faster than on the original ImageNet and the characteristics of the downsampled datasets with respect to optimal hyperparameters appear to remain similar. The proposed datasets and scripts to reproduce our results are available at this http URL and this https URL
https://arxiv.org/abs/1707.08819
The main contribution of this paper is a simple semi-supervised pipeline that only uses the original training set without collecting extra data. It is challenging in 1) how to obtain more training data only from the training set and 2) how to use the newly generated data. In this work, the generative adversarial network (GAN) is used to generate unlabeled samples. We propose the label smoothing regularization for outliers (LSRO). This method assigns a uniform label distribution to the unlabeled images, which regularizes the supervised model and improves the baseline. We verify the proposed method on a practical problem: person re-identification (re-ID). This task aims to retrieve a query person from other cameras. We adopt the deep convolutional generative adversarial network (DCGAN) for sample generation, and a baseline convolutional neural network (CNN) for representation learning. Experiments show that adding the GAN-generated data effectively improves the discriminative ability of learned CNN embeddings. On three large-scale datasets, Market-1501, CUHK03 and DukeMTMC-reID, we obtain +4.37%, +1.6% and +2.46% improvement in rank-1 precision over the baseline CNN, respectively. We additionally apply the proposed method to fine-grained bird recognition and achieve a +0.6% improvement over a strong baseline. The code is available at this https URL.
https://arxiv.org/abs/1701.07717
Sequence-to-sequence (Seq2Seq) models with attention have excelled at tasks which involve generating natural language sentences such as machine translation, image captioning and speech recognition. Performance has further been improved by leveraging unlabeled data, often in the form of a language model. In this work, we present the Cold Fusion method, which leverages a pre-trained language model during training, and show its effectiveness on the speech recognition task. We show that Seq2Seq models with Cold Fusion are able to better utilize language information enjoying i) faster convergence and better generalization, and ii) almost complete transfer to a new domain while using less than 10% of the labeled training data.
https://arxiv.org/abs/1708.06426
We propose a simple solution to use a single Neural Machine Translation (NMT) model to translate between multiple languages. Our solution requires no change in the model architecture from our base system but instead introduces an artificial token at the beginning of the input sentence to specify the required target language. The rest of the model, which includes encoder, decoder and attention, remains unchanged and is shared across all languages. Using a shared wordpiece vocabulary, our approach enables Multilingual NMT using a single model without any increase in parameters, which is significantly simpler than previous proposals for Multilingual NMT. Our method often improves the translation quality of all involved language pairs, even while keeping the total number of model parameters constant. On the WMT’14 benchmarks, a single multilingual model achieves comparable performance for English$\rightarrow$French and surpasses state-of-the-art results for English$\rightarrow$German. Similarly, a single multilingual model surpasses state-of-the-art results for French$\rightarrow$English and German$\rightarrow$English on WMT’14 and WMT’15 benchmarks respectively. On production corpora, multilingual models of up to twelve language pairs allow for better translation of many individual pairs. In addition to improving the translation quality of language pairs that the model was trained with, our models can also learn to perform implicit bridging between language pairs never seen explicitly during training, showing that transfer learning and zero-shot translation is possible for neural translation. Finally, we show analyses that hints at a universal interlingua representation in our models and show some interesting examples when mixing languages.
https://arxiv.org/abs/1611.04558
We introduce the Helsinki Neural Machine Translation system (HNMT) and how it is applied in the news translation task at WMT 2017, where it ranked first in both the human and automatic evaluations for English–Finnish. We discuss the success of English–Finnish translations and the overall advantage of NMT over a strong SMT baseline. We also discuss our submissions for English–Latvian, English–Chinese and Chinese–English.
https://arxiv.org/abs/1708.05942
We introduce a new model for building conditional generative models in a semi-supervised setting to conditionally generate data given attributes by adapting the GAN framework. The proposed semi-supervised GAN (SS-GAN) model uses a pair of stacked discriminators to learn the marginal distribution of the data, and the conditional distribution of the attributes given the data respectively. In the semi-supervised setting, the marginal distribution (which is often harder to learn) is learned from the labeled + unlabeled data, and the conditional distribution is learned purely from the labeled data. Our experimental results demonstrate that this model performs significantly better compared to existing semi-supervised conditional GAN models.
https://arxiv.org/abs/1708.05789
Neural machine translation (NMT) approaches have improved the state of the art in many machine translation settings over the last couple of years, but they require large amounts of training data to produce sensible output. We demonstrate that NMT can be used for low-resource languages as well, by introducing more local dependencies and using word alignments to learn sentence reordering during translation. In addition to our novel model, we also present an empirical evaluation of low-resource phrase-based statistical machine translation (SMT) and NMT to investigate the lower limits of the respective technologies. We find that while SMT remains the best option for low-resource settings, our method can produce acceptable translations with only 70000 tokens of training data, a level where the baseline NMT system fails completely.
https://arxiv.org/abs/1708.05729
We uncover the origin of blue luminescence (BL) peak in Mg doped GaN thin film using a combination of experimental X-ray absorption near edge spectroscopy (XANES), first-principles calculations based on density functional theory and full multiple scattering theoretical analysis of various possible defect complexes and their XANES signatures. We demonstrate that a defect complex composed of Mg substituted at Ga site (MgGa) and Mg at interstitial site (Mgi) is primarily responsible for the observed BL by Donor-Acceptor Pair transition (DAP) associated with a deep donor state in the gap. It correlates with a higher (lower) oxidation state of N (Ga) in heavily Mg doped GaN than in its pristine structure, evident in our experiments as well as calculations. Physical and chemical mechanisms identified here point out a route to achieving efficient p-type GaN.
https://arxiv.org/abs/1708.04036
Salient object detection has seen remarkable progress driven by deep learning techniques. However, most of deep learning based salient object detection methods are black-box in nature and lacking in interpretability. This paper proposes the first self-explanatory saliency detection network that explicitly exploits low- and high-level features for salient object detection. We demonstrate that such supportive clues not only significantly enhances performance of salient object detection but also gives better justified detection results. More specifically, we develop a multi-stage saliency encoder to extract multi-scale features which contain both low- and high-level saliency context. Dense short- and long-range connections are introduced to reuse these features iteratively. Benefiting from the direct access to low- and high-level features, the proposed saliency encoder can not only model the object context but also preserve the boundary. Furthermore, a self-explanatory generator is proposed to interpret how the proposed saliency encoder or other deep saliency models making decisions. The generator simulates the absence of interesting features by preventing these features from contributing to the saliency classifier and estimates the corresponding saliency prediction without these features. A comparison function, saliency explanation, is defined to measure the prediction changes between deep saliency models and corresponding generator. Through visualizing the differences, we can interpret the capability of different deep neural networks based saliency detection models and demonstrate that our proposed model indeed uses more reasonable structure for salient object detection. Extensive experiments on five popular benchmark datasets and the visualized saliency explanation demonstrate that the proposed method provides new state-of-the-art.
https://arxiv.org/abs/1708.05595
Extending state-of-the-art object detectors from image to video is challenging. The accuracy of detection suffers from degenerated object appearances in videos, e.g., motion blur, video defocus, rare poses, etc. Existing work attempts to exploit temporal information on box level, but such methods are not trained end-to-end. We present flow-guided feature aggregation, an accurate and end-to-end learning framework for video object detection. It leverages temporal coherence on feature level instead. It improves the per-frame features by aggregation of nearby features along the motion paths, and thus improves the video recognition accuracy. Our method significantly improves upon strong single-frame baselines in ImageNet VID, especially for more challenging fast moving objects. Our framework is principled, and on par with the best engineered systems winning the ImageNet VID challenges 2016, without additional bells-and-whistles. The proposed method, together with Deep Feature Flow, powered the winning entry of ImageNet VID challenges 2017. The code is available at this https URL.
https://arxiv.org/abs/1703.10025
Deep Neural Network (DNN) are currently of great inter- est in research and application. The training of these net- works is a compute intensive and time consuming task. To reduce training times to a bearable amount at reasonable cost we extend the popular Caffe toolbox for DNN with an efficient distributed memory communication pattern. To achieve good scalability we emphasize the overlap of computation and communication and prefer fine granu- lar synchronization patterns over global barriers. To im- plement these communication patterns we rely on the the Global address space Programming Interface version 2 (GPI-2) communication library. This interface provides a light-weight set of asynchronous one-sided communica- tion primitives supplemented by non-blocking fine gran- ular data synchronization mechanisms. Therefore, Caf- feGPI is the name of our parallel version of Caffe. First benchmarks demonstrate better scaling behavior com- pared with other extensions, e.g., the Intel TM Caffe. Even within a single symmetric multiprocessing machine with four graphics processing units, the CaffeGPI scales bet- ter than the standard Caffe toolbox. These first results demonstrate that the use of standard High Performance Computing (HPC) hardware is a valid cost saving ap- proach to train large DDNs. I/O is an other bottleneck to work with DDNs in a standard parallel HPC setting, which we will consider in more detail in a forthcoming paper.
https://arxiv.org/abs/1706.00095
In this paper, we make a simple observation that questions about images often contain premises - objects and relationships implied by the question - and that reasoning about premises can help Visual Question Answering (VQA) models respond more intelligently to irrelevant or previously unseen questions. When presented with a question that is irrelevant to an image, state-of-the-art VQA models will still answer purely based on learned language biases, resulting in non-sensical or even misleading answers. We note that a visual question is irrelevant to an image if at least one of its premises is false (i.e. not depicted in the image). We leverage this observation to construct a dataset for Question Relevance Prediction and Explanation (QRPE) by searching for false premises. We train novel question relevance detection models and show that models that reason about premises consistently outperform models that do not. We also find that forcing standard VQA models to reason about premises during training can lead to improvements on tasks requiring compositional reasoning.
https://arxiv.org/abs/1705.00601
Designing architectures for deep neural networks requires expert knowledge and substantial computation time. We propose a technique to accelerate architecture selection by learning an auxiliary HyperNet that generates the weights of a main model conditioned on that model’s architecture. By comparing the relative validation performance of networks with HyperNet-generated weights, we can effectively search over a wide range of architectures at the cost of a single training run. To facilitate this search, we develop a flexible mechanism based on memory read-writes that allows us to define a wide range of network connectivity patterns, with ResNet, DenseNet, and FractalNet blocks as special cases. We validate our method (SMASH) on CIFAR-10 and CIFAR-100, STL-10, ModelNet10, and Imagenet32x32, achieving competitive performance with similarly-sized hand-designed networks. Our code is available at this https URL
https://arxiv.org/abs/1708.05344
Image captioning often requires a large set of training image-sentence pairs. In practice, however, acquiring sufficient training pairs is always expensive, making the recent captioning models limited in their ability to describe objects outside of training corpora (i.e., novel objects). In this paper, we present Long Short-Term Memory with Copying Mechanism (LSTM-C) — a new architecture that incorporates copying into the Convolutional Neural Networks (CNN) plus Recurrent Neural Networks (RNN) image captioning framework, for describing novel objects in captions. Specifically, freely available object recognition datasets are leveraged to develop classifiers for novel objects. Our LSTM-C then nicely integrates the standard word-by-word sentence generation by a decoder RNN with copying mechanism which may instead select words from novel objects at proper places in the output sentence. Extensive experiments are conducted on both MSCOCO image captioning and ImageNet datasets, demonstrating the ability of our proposed LSTM-C architecture to describe novel objects. Furthermore, superior results are reported when compared to state-of-the-art deep models.
https://arxiv.org/abs/1708.05271
We present LBW-Net, an efficient optimization based method for quantization and training of the low bit-width convolutional neural networks (CNNs). Specifically, we quantize the weights to zero or powers of two by minimizing the Euclidean distance between full-precision weights and quantized weights during backpropagation. We characterize the combinatorial nature of the low bit-width quantization problem. For 2-bit (ternary) CNNs, the quantization of $N$ weights can be done by an exact formula in $O(N\log N)$ complexity. When the bit-width is three and above, we further propose a semi-analytical thresholding scheme with a single free parameter for quantization that is computationally inexpensive. The free parameter is further determined by network retraining and object detection tests. LBW-Net has several desirable advantages over full-precision CNNs, including considerable memory savings, energy efficiency, and faster deployment. Our experiments on PASCAL VOC dataset show that compared with its 32-bit floating-point counterpart, the performance of the 6-bit LBW-Net is nearly lossless in the object detection tasks, and can even do better in some real world visual scenes, while empirically enjoying more than 4$\times$ faster deployment.
https://arxiv.org/abs/1612.06052
Learning image representations with ConvNets by pre-training on ImageNet has proven useful across many visual understanding tasks including object detection, semantic segmentation, and image captioning. Although any image representation can be applied to video frames, a dedicated spatiotemporal representation is still vital in order to incorporate motion patterns that cannot be captured by appearance based models alone. This paper presents an empirical ConvNet architecture search for spatiotemporal feature learning, culminating in a deep 3-dimensional (3D) Residual ConvNet. Our proposed architecture outperforms C3D by a good margin on Sports-1M, UCF101, HMDB51, THUMOS14, and ASLAN while being 2 times faster at inference time, 2 times smaller in model size, and having a more compact representation.
https://arxiv.org/abs/1708.05038
In this paper we present an approach to estimate Free Space from a Stereo image pair using stochastic occupancy grids. We do this in the domain of autonomous driving on the famous benchmark dataset KITTI. Later based on the generated occupancy grid we match 2 image sequences to compute the top view representation of the map. We do this to map the environment. We compute a transformation between the occupancy grids of two successive images and use it to compute the top view map. Two issues need to be addressed for mapping are discussed - computing a map and dealing with dynamic objects for computing the map. Dynamic Objects are detected in successive images based on an idea similar to tracking of foreground objects from the background objects based on motion flow. A novel RANSAC based segmentation approach has been proposed here to address this issue.
https://arxiv.org/abs/1708.04989
We describe studies on the nanoscale transport dynamics of carriers in strained AlN/GaN/AlN quantum wells: an electron-hole bilayer charge system with large difference in transport properties between the two charge layers. From electronic band diagram analysis, the presence of spatially separated two-dimensional electron and hole charge layers is predicted at opposite interfaces. Since these charge layers exhibit distinct spectral signatures at terahertz frequencies, a combination of terahertz and far-infrared spectroscopy enables us to extract (a) individual contributions to the total conductivity, as well as (b) effective scattering rates for charge-carriers in each layer. Furthermore, by comparing direct-current and terahertz extracted conductivity levels, we are able to determine the extent to which structural defects affect charge transport. Our results evidence that (i) a non-unity Hall-factor and (ii) the considerable contribution of holes to the overall conductivity, lead to a lower apparent mobility in Hall-effect measurements. Overall, our work demonstrates that terahertz spectroscopy is a suitable technique for the study of bilayer charge systems with large differences in transport properties between layers, such as quantum wells in III-Nitride semiconductors.
https://arxiv.org/abs/1708.04752
Rich and dense human labeled datasets are among the main enabling factors for the recent advance on vision-language understanding. Many seemingly distant annotations (e.g., semantic segmentation and visual question answering (VQA)) are inherently connected in that they reveal different levels and perspectives of human understandings about the same visual scenes — and even the same set of images (e.g., of COCO). The popularity of COCO correlates those annotations and tasks. Explicitly linking them up may significantly benefit both individual tasks and the unified vision and language modeling. We present the preliminary work of linking the instance segmentations provided by COCO to the questions and answers (QAs) in the VQA dataset, and name the collected links visual questions and segmentation answers (VQS). They transfer human supervision between the previously separate tasks, offer more effective leverage to existing problems, and also open the door for new research problems and models. We study two applications of the VQS data in this paper: supervised attention for VQA and a novel question-focused semantic segmentation task. For the former, we obtain state-of-the-art results on the VQA real multiple-choice task by simply augmenting the multilayer perceptrons with some attention features that are learned using the segmentation-QA links as explicit supervision. To put the latter in perspective, we study two plausible methods and compare them to an oracle method assuming that the instance segmentations are given at the test stage.
https://arxiv.org/abs/1708.04686
One of the major challenges in object detection is to propose detectors with highly accurate localization of objects. The online sampling of high-loss region proposals (hard examples) uses the multitask loss with equal weight settings across all loss types (e.g, classification and localization, rigid and non-rigid categories) and ignores the influence of different loss distributions throughout the training process, which we find essential to the training efficacy. In this paper, we present the Stratified Online Hard Example Mining (S-OHEM) algorithm for training higher efficiency and accuracy detectors. S-OHEM exploits OHEM with stratified sampling, a widely-adopted sampling technique, to choose the training examples according to this influence during hard example mining, and thus enhance the performance of object detectors. We show through systematic experiments that S-OHEM yields an average precision (AP) improvement of 0.5% on rigid categories of PASCAL VOC 2007 for both the IoU threshold of 0.6 and 0.7. For KITTI 2012, both results of the same metric are 1.6%. Regarding the mean average precision (mAP), a relative increase of 0.3% and 0.5% (1% and 0.5%) is observed for VOC07 (KITTI12) using the same set of IoU threshold. Also, S-OHEM is easy to integrate with existing region-based detectors and is capable of acting with post-recognition level regressors.
https://arxiv.org/abs/1705.02233
Image captioning has so far been explored mostly in English, as most available datasets are in this language. However, the application of image captioning should not be restricted by language. Only few studies have been conducted for image captioning in a cross-lingual setting. Different from these works that manually build a dataset for a target language, we aim to learn a cross-lingual captioning model fully from machine-translated sentences. To conquer the lack of fluency in the translated sentences, we propose in this paper a fluency-guided learning framework. The framework comprises a module to automatically estimate the fluency of the sentences and another module to utilize the estimated fluency scores to effectively train an image captioning model for the target language. As experiments on two bilingual (English-Chinese) datasets show, our approach improves both fluency and relevance of the generated captions in Chinese, but without using any manually written sentences from the target language.
https://arxiv.org/abs/1708.04390
Impressive image captioning results are achieved in domains with plenty of training image and sentence pairs (e.g., MSCOCO). However, transferring to a target domain with significant domain shifts but no paired training data (referred to as cross-domain image captioning) remains largely unexplored. We propose a novel adversarial training procedure to leverage unpaired data in the target domain. Two critic networks are introduced to guide the captioner, namely domain critic and multi-modal critic. The domain critic assesses whether the generated sentences are indistinguishable from sentences in the target domain. The multi-modal critic assesses whether an image and its generated sentence are a valid pair. During training, the critics and captioner act as adversaries – captioner aims to generate indistinguishable sentences, whereas critics aim at distinguishing them. The assessment improves the captioner through policy gradient updates. During inference, we further propose a novel critic-based planning method to select high-quality sentences without additional supervision (e.g., tags). To evaluate, we use MSCOCO as the source domain and four other datasets (CUB-200-2011, Oxford-102, TGIF, and Flickr30k) as the target domains. Our method consistently performs well on all datasets. In particular, on CUB-200-2011, we achieve 21.8% CIDEr-D improvement after adaptation. Utilizing critics during inference further gives another 4.5% boost.
https://arxiv.org/abs/1705.00930
Recurrent networks are trained to memorize their input better, often in the hopes that such training will increase the ability of the network to predict. We show that networks designed to memorize input can be arbitrarily bad at prediction. We also find, for several types of inputs, that one-node networks optimized for prediction are nearly at upper bounds on predictive capacity given by Wiener filters, and are roughly equivalent in performance to randomly generated five-node networks. Our results suggest that maximizing memory capacity leads to very different networks than maximizing predictive capacity, and that optimizing recurrent weights can decrease reservoir size by half an order of magnitude.
https://arxiv.org/abs/1706.09382
Simultaneous administration of multiple drugs can have synergistic or antagonistic effects as one drug can affect activities of other drugs. Synergistic effects lead to improved therapeutic outcomes, whereas, antagonistic effects can be life-threatening, may lead to increased healthcare cost, or may even cause death. Thus identification of unknown drug-drug interaction (DDI) is an important concern for efficient and effective healthcare. Although multiple resources for DDI exist, they are often unable to keep pace with rich amount of information available in fast growing biomedical texts. Most existing methods model DDI extraction from text as a classification problem and mainly rely on handcrafted features. Some of these features further depend on domain specific tools. Recently neural network models using latent features have been shown to give similar or better performance than the other existing models dependent on handcrafted features. In this paper, we present three models namely, {\it B-LSTM}, {\it AB-LSTM} and {\it Joint AB-LSTM} based on long short-term memory (LSTM) network. All three models utilize word and position embedding as latent features and thus do not rely on explicit feature engineering. Further use of bidirectional long short-term memory (Bi-LSTM) networks allow implicit feature extraction from the whole sentence. The two models, {\it AB-LSTM} and {\it Joint AB-LSTM} also use attentive pooling in the output of Bi-LSTM layer to assign weights to features. Our experimental results on the SemEval-2013 DDI extraction dataset show that the {\it Joint AB-LSTM} model outperforms all the existing methods, including those relying on handcrafted features. The other two proposed LSTM models also perform competitively with state-of-the-art methods.
https://arxiv.org/abs/1701.08303