Prior background knowledge is essential for human reading and understanding. In this work, we investigate how to leverage external knowledge to improve question answering. We primarily focus on multiple-choice question answering tasks that require external knowledge to answer questions. We investigate the effects of utilizing external in-domain multiple-choice question answering datasets and enriching the reference corpus by external out-domain corpora (i.e., Wikipedia articles). Experimental results demonstrate the effectiveness of external knowledge on two challenging multiple-choice question answering tasks: ARC and OpenBookQA.
http://arxiv.org/abs/1902.00993
In this paper we present a novel lemmatization method based on a sequence-to-sequence neural network architecture and morphosyntactic context representation. In the proposed method, our context-sensitive lemmatizer generates the lemma one character at a time based on the surface form characters and its morphosyntactic features obtained from a morphological tagger. We argue that a sliding window context representation suffers from sparseness, while in majority of cases the morphosyntactic features of a word bring enough information to resolve lemma ambiguities while keeping the context representation dense and more practical for machine learning systems. Additionally, we study two different data augmentation methods utilizing autoencoder training and morphological transducers especially beneficial for low resource languages. We evaluate our lemmatizer on 52 different languages and 76 different treebanks, showing that our system outperforms all latest baseline systems. Compared to the best overall baseline, UDPipe Future, our system outperforms it on 60 out of 76 treebanks reducing errors on average by 18% relative. The lemmatizer together with all trained models is made available as a part of the Turku-neural-parsing-pipeline under the Apache 2.0 license.
http://arxiv.org/abs/1902.00972
There has recently been a surge of work in explanatory artificial intelligence (XAI). This research area tackles the important problem that complex machines and algorithms often cannot provide insights into their behavior and thought processes. XAI allows users and parts of the internal system to be more transparent, providing explanations of their decisions in some level of detail. These explanations are important to ensure algorithmic fairness, identify potential bias/problems in the training data, and to ensure that the algorithms perform as expected. However, explanations produced by these systems is neither standardized nor systematically assessed. In an effort to create best practices and identify open challenges, we provide our definition of explainability and show how it can be used to classify existing literature. We discuss why current approaches to explanatory methods especially for deep neural networks are insufficient. Finally, based on our survey, we conclude with suggested future research directions for explanatory artificial intelligence.
http://arxiv.org/abs/1806.00069
The most straightforward method to accelerate Stochastic Gradient Descent (SGD) is to distribute the randomly selected batch of inputs over multiple processors. To keep the distributed processors fully utilized requires commensurately growing the batch size; however, large batch training usually leads to poor generalization. Existing solutions for large batch training either significantly degrade accuracy or require massive hyper-parameter tuning. To address this issue, we propose a novel large batch training method which combines recent results in adversarial training and second order information. We extensively evaluate our method on Cifar-10/100, SVHN, TinyImageNet, and ImageNet datasets, using multiple NNs, including residual networks as well as smaller networks such as SqueezeNext. Our new approach exceeds the performance of the existing solutions in terms of both accuracy and the number of SGD iterations (up to 1\% and $5\times$, respectively). We emphasize that this is achieved without any additional hyper-parameter tuning to tailor our proposed method in any of these experiments. With slight hyper-parameter tuning, our method can reduce the number of SGD iterations of ResNet18 on Cifar-10/ImageNet to $44.8\times$ and $28.8\times$, respectively. We have open sourced the method including tools for computing Hessian spectrum.
http://arxiv.org/abs/1810.01021
We describe a machine-learning approach to pitch correcting a solo singing performance in a karaoke setting, where the solo voice and accompaniment are on separate tracks. The proposed approach addresses the situation where no musical score of the vocals nor the accompaniment exists: It predicts the amount of correction from the relationship between the spectral contents of the vocal and accompaniment tracks. Hence, the pitch shift in cents suggested by the model can be used to make the voice sound in tune with the accompaniment. This approach differs from commercially used automatic pitch correction systems, where notes in the vocal tracks are shifted to be centered around notes in a user-defined score or mapped to the closest pitch among the twelve equal-tempered scale degrees. We train the model using a dataset of 4,702 amateur karaoke performances selected for good intonation. We present a Convolutional Gated Recurrent Unit (CGRU) model to accomplish this task. This method can be extended into unsupervised pitch correction of a vocal performance, popularly referred to as autotuning.
http://arxiv.org/abs/1902.00956
StarCraft II poses a grand challenge for reinforcement learning. The main difficulties of it include huge state and action space and a long-time horizon. In this paper, we investigate a hierarchical reinforcement learning approach for StarCraft II. The hierarchy involves two levels of abstraction. One is the macro-action automatically extracted from expert’s trajectories, which reduces the action space in an order of magnitude yet remains effective. The other is a two-layer hierarchical architecture which is modular and easy to scale, enabling a curriculum transferring from simpler tasks to more complex tasks. The reinforcement training algorithm for this architecture is also investigated. On a 64x64 map and using restrictive units, we achieve a winning rate of more than 99\% against the difficulty level-1 built-in AI. Through the curriculum transfer learning algorithm and a mixture of combat model, we can achieve over 93\% winning rate of Protoss against the most difficult non-cheating built-in AI (level-7) of Terran, training within two days using a single machine with only 48 CPU cores and 8 K40 GPUs. It also shows strong generalization performance, when tested against never seen opponents including cheating levels built-in AI and all levels of Zerg and Protoss built-in AI. We hope this study could shed some light on the future research of large-scale reinforcement learning.
http://arxiv.org/abs/1809.09095
There is a growing interest in designing models that can deal with images from different visual domains. If there exists a universal structure in different visual domains that can be captured via a common parameterization, then we can use a single model for all domains rather than one model per domain. A model aware of the relationships between different domains can also be trained to work on new domains with less resources. However, to identify the reusable structure in a model is not easy. In this paper, we propose a multi-domain learning architecture based on depthwise separable convolution. The proposed approach is based on the assumption that images from different domains share cross-channel correlations but have domain-specific spatial correlations. The proposed model is compact and has minimal overhead when being applied to new domains. Additionally, we introduce a gating mechanism to promote soft sharing between different domains. We evaluate our approach on Visual Decathlon Challenge, a benchmark for testing the ability of multi-domain models. The experiments show that our approach can achieve the highest score while only requiring 50% of the parameters compared with the state-of-the-art approaches.
http://arxiv.org/abs/1902.00927
Convolutional networks reach top quality in pixel-level video object segmentation but require a large amount of training data (1k~100k) to deliver such results. We propose a new training strategy which achieves state-of-the-art results across three evaluation datasets while using 20x~1000x less annotated data than competing methods. Our approach is suitable for both single and multiple object segmentation. Instead of using large training sets hoping to generalize across domains, we generate in-domain training data using the provided annotation on the first frame of each video to synthesize (“lucid dream”) plausible future video frames. In-domain per-video training data allows us to train high quality appearance- and motion-based models, as well as tune the post-processing stage. This approach allows to reach competitive results even when training from only a single annotated frame, without ImageNet pre-training. Our results indicate that using a larger training set is not automatically better, and that for the video object segmentation task a smaller training set that is closer to the target domain is more effective. This changes the mindset regarding how many training samples and general “objectness” knowledge are required for the video object segmentation task.
http://arxiv.org/abs/1703.09554
State-of-the-art deep model compression methods exploit the low-rank approximation and sparsity pruning to remove redundant parameters from a learned hidden layer. However, they process each hidden layer individually while neglecting the common components across layers, and thus are not able to fully exploit the potential redundancy space for compression. To solve the above problem and enable further compression of a model, removing the cross-layer redundancy and mining the layer-wise inheritance knowledge is necessary. In this paper, we introduce a holistic model compression framework, namely MIning Cross-layer Inherent similarity Knowledge (MICIK), to fully excavate the potential redundancy space. The proposed MICIK framework simultaneously, (1) learns the common and unique weight components across deep neural network layers to increase compression rate; (2) preserves the inherent similarity knowledge of nearby layers and distant layers to minimize the accuracy loss and (3) can be complementary to other existing compression techniques such as knowledge distillation. Extensive experiments on large-scale convolutional neural networks demonstrate that MICIK is superior over state-of-the-art model compression approaches with 16X parameter reduction on VGG-16 and 6X on GoogLeNet, all without accuracy loss.
http://arxiv.org/abs/1902.00918
Knowledge graphs have recently become the state-of-the-art tool for representing the diverse and complex knowledge of the world. Examples include the proprietary knowledge graphs of companies such as Google, Facebook, IBM, or Microsoft, but also freely available ones such as YAGO, DBpedia, and Wikidata. A distinguishing feature of Wikidata is that the knowledge is collaboratively edited and curated. While this greatly enhances the scope of Wikidata, it also makes it impossible for a single individual to grasp complex connections between properties or understand the global impact of edits in the graph. We apply Formal Concept Analysis to efficiently identify comprehensible implications that are implicitly present in the data. Although the complex structure of data modelling in Wikidata is not amenable to a direct approach, we overcome this limitation by extracting contextual representations of parts of Wikidata in a systematic fashion. We demonstrate the practical feasibility of our approach through several experiments and show that the results may lead to the discovery of interesting implicational knowledge. Besides providing a method for obtaining large real-world data sets for FCA, we sketch potential applications in offering semantic assistance for editing and curating Wikidata.
http://arxiv.org/abs/1902.00916
We consider the task of inferring is-a relationships from large text corpora. For this purpose, we propose a new method combining hyperbolic embeddings and Hearst patterns. This approach allows us to set appropriate constraints for inferring concept hierarchies from distributional contexts while also being able to predict missing is-a relationships and to correct wrong extractions. Moreover – and in contrast with other methods – the hierarchical nature of hyperbolic space allows us to learn highly efficient representations and to improve the taxonomic consistency of the inferred hierarchies. Experimentally, we show that our approach achieves state-of-the-art performance on several commonly-used benchmarks.
http://arxiv.org/abs/1902.00913
In this paper, we will study the simplest kind of beauty which can be found in simple visual patterns. The proposed approach shows that aesthetically appealing patterns deliver higher amount of information over multiple levels in comparison with less aesthetically appealing patterns when the same amount of energy is used. The proposed approach is used to classify aesthetically appealing patterns.
http://arxiv.org/abs/1705.08244
Stochastic gradient descent (SGD) is a popular and efficient method with wide applications in training deep neural nets and other nonconvex models. While the behavior of SGD is well understood in the convex learning setting, the existing theoretical results for SGD applied to nonconvex objective functions are far from mature. For example, existing results require to impose a nontrivial assumption on the uniform boundedness of gradients for all iterates encountered in the learning process, which is hard to verify in practical implementations. In this paper, we establish a rigorous theoretical foundation for SGD in nonconvex learning by showing that this boundedness assumption can be removed without affecting convergence rates. In particular, we establish sufficient conditions for almost sure convergence as well as optimal convergence rates for SGD applied to both general nonconvex objective functions and gradient-dominated objective functions. A linear convergence is further derived in the case with zero variances.
http://arxiv.org/abs/1902.00908
We present DeepISP, a full end-to-end deep neural model of the camera image signal processing (ISP) pipeline. Our model learns a mapping from the raw low-light mosaiced image to the final visually compelling image and encompasses low-level tasks such as demosaicing and denoising as well as higher-level tasks such as color correction and image adjustment. The training and evaluation of the pipeline were performed on a dedicated dataset containing pairs of low-light and well-lit images captured by a Samsung S7 smartphone camera in both raw and processed JPEG formats. The proposed solution achieves state-of-the-art performance in objective evaluation of PSNR on the subtask of joint denoising and demosaicing. For the full end-to-end pipeline, it achieves better visual quality compared to the manufacturer ISP, in both a subjective human assessment and when rated by a deep model trained for assessing image quality.
http://arxiv.org/abs/1801.06724
We present a robust real-time LiDAR 3D object detector that leverages heteroscedastic aleatoric uncertainties to significantly improve its detection performance. A multi-loss function is designed to incorporate uncertainty estimations predicted by auxiliary output layers. Using our proposed method, the network ignores to train from noisy samples, and focuses more on informative ones. We validate our method on the KITTI object detection benchmark. Our method surpasses the baseline method which does not explicitly estimate uncertainties by up to nearly 9% in terms of Average Precision (AP). It also produces state-of-the-art results compared to other methods while running with an inference time of only 72 ms. In addition, we conduct extensive experiments to understand how aleatoric uncertainties behave. Extracting aleatoric uncertainties brings almost no additional computation cost during the deployment, making our method highly desirable for autonomous driving applications.
https://arxiv.org/abs/1809.05590
Recent neural network approaches to summarization are largely either sentence-extractive, choosing a set of sentences as the summary, or abstractive, generating the summary from a seq2seq model. In this work, we present a neural model for single-document summarization based on joint extraction and compression. Following recent successful extractive models, we frame the summarization problem as a series of local decisions. Our model chooses sentences from the document and then decides which of a set of compression options to apply to each selected sentence. We compute this set of options using discrete compression rules based on syntactic constituency parses; however, our approach is modular and can flexibly use any available source of compressions. For learning, we construct oracle extractive-compressive summaries that reflect uncertainty over our model’s decision sequence, then learn both of our components jointly with this supervision. Experimental results on the CNN/Daily Mail and New York Times datasets show that our model achieves the state-of-the-art performance on content selection evaluated by ROUGE. Moreover, human and manual evaluation show that our model’s output generally remains grammatical.
http://arxiv.org/abs/1902.00863
In this paper, we address the single image haze removal problem in a nighttime scene. The night haze removal is a severely ill-posed problem especially due to the presence of various visible light sources with varying colors and non-uniform illumination. These light sources are of different shapes and introduce noticeable glow in night scenes. To address these effects we introduce a deep learning based DeGlow-DeHaze iterative architecture which accounts for varying color illumination and glows. First, our convolution neural network (CNN) based DeGlow model is able to remove the glow effect significantly and on top of it a separate DeHaze network is included to remove the haze effect. For our recurrent network training, the hazy images and the corresponding transmission maps are synthesized from the NYU depth datasets and consequently restored a high-quality haze-free image. The experimental results demonstrate that our hybrid CNN model outperforms other state-of-the-art methods in terms of computation speed and image quality. We also show the effectiveness of our model on a number of real images and compare our results with the existing night haze heuristic models.
http://arxiv.org/abs/1902.00855
In this paper we consider the problem of how a reinforcement learning agent tasked with solving a set of related Markov decision processes can use knowledge acquired early in its lifetime to improve its ability to more rapidly solve novel, but related, tasks. One way of exploiting this experience is by identifying recurrent patterns in trajectories obtained from well-performing policies. We propose a three-step framework in which an agent 1) generates a set of candidate open-loop macros by compressing trajectories drawn from near-optimal policies; 2) evaluates the value of each macro; and 3) selects a maximally diverse subset of macros that spans the space of policies typically required for solving the set of related tasks. Our experiments show that extending the original primitive action-set of the agent with the identified macros allows it to more rapidly learn an optimal policy in unseen, but similar MDPs.
http://arxiv.org/abs/1711.09048
Mobile robots navigating in indoor and outdoor environments must be able to identify and avoid unsafe terrain. Although a significant amount of work has been done on the detection of standing obstacles (solid obstructions), not much work has been done on the detection of negative obstacles (e.g. dropoffs, ledges, downward stairs). We propose a method of terrain safety segmentation using deep convolutional networks. Our custom semantic segmentation architecture uses a single camera as input and creates a freespace map distinguishing safe terrain and obstacles. We then show how this freespace map can be used for real-time navigation on an indoor robot. The results show that our system generalizes well, is suitable for real-time operation, and runs at around 55 fps on a small indoor robot powered by a low-power embedded GPU.
http://arxiv.org/abs/1902.00842
Complex motion patterns of natural systems, such as fish schools, bird flocks, and cell groups, have attracted great attention from scientists for years. Trajectory measurement of individuals is vital for quantitative and high-throughput study of their collective behaviors. However, such data are rare mainly due to the challenges of detection and tracking of large numbers of objects with similar visual features and frequent occlusions. We present an automatic and effective framework to measure trajectories of large numbers of crowded oval-shaped objects, such as fish and cells. We first use a novel dual ellipse locator to detect the coarse position of each individual and then propose a variance minimization active contour method to obtain the optimal segmentation results. For tracking, cost matrix of assignment between consecutive frames is trainable via a random forest classifier with many spatial, texture, and shape features. The optimal trajectories are found for the whole image sequence by solving two linear assignment problems. We evaluate the proposed method on many challenging data sets.
http://arxiv.org/abs/1902.00835
Seeking information about products and services is an important activity of online consumers before making a purchase decision. Inspired by recent research on conversational reading comprehension (CRC) on formal documents, this paper studies the task of leveraging knowledge from a huge amount of reviews to answer multi-turn questions from consumers or users. Questions spanning multiple turns in a dialogue enables users to ask more specific questions that are hard to ask within a single question as in traditional machine reading comprehension (MRC). In this paper, we first build a dataset and then propose a novel task-adaptation approach to encoding the formulation of CRC task into a pre-trained language model. This task-adaptation approach is unsupervised and can greatly enhance the performance of the end CRC task that has only limited supervision. Experimental results show that the proposed approach is highly effective and has competitive performance as supervised approach. We plan to release the datasets and the code in May 2019.
http://arxiv.org/abs/1902.00821
This paper presents a novel unsupervised probabilistic model estimation of visual background in video sequences using a variational autoencoder framework. Due to the redundant nature of the backgrounds in surveillance videos, visual information of the background can be compressed into a low-dimensional subspace in the encoder part of the variational autoencoder, while the highly variant information of its moving foreground gets filtered throughout its encoding-decoding process. Our deep probabilistic background model (DeepPBM) estimation approach is enabled by the power of deep neural networks in learning compressed representations of video frames and reconstructing them back to the original domain. We evaluated the performance of our DeepPBM in background subtraction on 9 surveillance videos from the background model challenge (BMC2012) dataset, and compared that with a standard subspace learning technique, robust principle component analysis (RPCA), which similarly estimates a deterministic low dimensional representation of the background in videos and is widely used for this application. Our method outperforms RPCA on BMC2012 dataset with 23% in average in F-measure score, emphasizing that background subtraction using the trained model can be done in more than 10 times faster.
http://arxiv.org/abs/1902.00820
The types of sound events that occur in a situation are limited, and some sound events are likely to co-occur; for instance, dishes'' andglass jingling.’’ In this paper, we introduce a technique of sound event detection utilizing graph Laplacian regularization taking the sound event co-occurrence into account. To consider the co-occurrence of sound events in a sound event detection system, we first represent sound event occurrences as a graph whose nodes indicate the frequency of event occurrence and whose edges indicate the co-occurrence of sound events. We then utilize this graph structure for sound event modeling, which is optimized under an objective function with a regularization term considering the graph structure. Experimental results obtained using TUT Acoustic Scenes 2016 development and 2017 development datasets indicate that the proposed method improves the detection performance of sound events by 7.9 percentage points compared to that of the conventional CNN-BiGRU-based method in terms of the segment-based F1-score. Moreover, the results show that the proposed method can detect co-occurring sound events more accurately than the conventional method.
http://arxiv.org/abs/1902.00816
Generative adversarial networks (GANs) have shown great promise in generating complex data such as images. A standard practice in GANs is to discard the discriminator after training and use only the generator for sampling. However, this loses valuable information of real data distribution learned by the discriminator. In this work, we propose a collaborative sampling scheme between the generator and discriminator for improved data generation. Guided by the discriminator, our approach refines generated samples through gradient-based optimization, shifting the generator distribution closer to the real data distribution. Additionally, we present a practical discriminator shaping method that can further improve the sample refinement process. Orthogonal to existing GAN variants, our proposed method offers a new degree of freedom in GAN sampling. We demonstrate its efficacy through experiments on synthetic data and image generation tasks.
http://arxiv.org/abs/1902.00813
While non-suicidal self-injury (NSSI) is not a new phenomenon, there is still a limited yet little is still known about understanding of the behavior, the intent behind the behavior and what the individuals themselves say about their behavior. This study collected pro-NSSI public blog posts from Reddit on pro-NSSI and analyzed the content linguistically using LIWC software, in order to examine the use of NSSI specific words, linguistic properties and the psychological linguistic properties. were examined. The results inform current counseling practices by dispelling myths and providing insight into the inner world of people who engage in use NSSII to cope. The most frequently appearing category of For NSSI specific words categories, in the Reddit blogs was the reasons in which one engagesfor engaging in NSSI was the most frequently used in the Reddit blogs. The linguistic properties found in the analysis reflected the predicted results; authors of pro-NSSI posts used demonstrated expected results of first-person singular pronouns extensively, which indicatesing high levels of mental health distress and isolation. The psychological linguistic properties that could be observed of in these public Reddit posts were dominantly in a negative emotional tone which demonstrates youth and impulsivity. The linguistic properties found when these posts were analyzed supports the work of earlier studies that dispelled common myths about NSSI that were circulating in the mental health community. These findings suggest that the language of people who engage in NSSI supports research findings in dispelling common myths about NSSI.
http://arxiv.org/abs/1902.06689
Early detection of skin cancer, particularly melanoma, is crucial to enable advanced treatment. Due to the rapid growth of skin cancers, there is a growing need of computerized analysis for skin lesions. These processes including detection, classification, and segmentation. The state-of-the-art public available datasets for skin lesions are often accompanied with a very limited amount of segmentation ground truth labeling as it is laborious and expensive. The lesion boundary segmentation is vital to locate the lesion accurately in dermoscopic images and lesion diagnosis of different skin lesion types. In this work, we propose the use of fully automated deep learning ensemble methods for accurate lesion boundary segmentation in dermoscopic images. We trained the Mask-RCNN and DeeplabV3+ methods on ISIC-2017 segmentation training dataset and evaluate the various ensemble performance of both networks on ISIC-2017 testing set, PH2 dataset. Our results showed that the proposed ensemble method segmented the skin lesions with Jaccard index of 79.58% for the ISBI 2017 test dataset. In comparison to FrCN, FCN, U-Net, and SegNet, the proposed ensemble method outperformed them by 2.48%, 7.42%, 17.95%, and 9.96% for the Jaccard index, respectively. Furthermore, the proposed ensemble method achieved a segmentation accuracy of 95.6% for some representative clinical benign cases, 90.78\% for the melanoma cases, and 91.29% for the seborrheic keratosis cases in the ISBI 2017 test dataset, exhibiting better performance than those of FrCN, FCN, U-Net, and SegNet.
http://arxiv.org/abs/1902.00809
Dialogue systems have many applications such as customer support or question answering. Typically they have been limited to shallow single turn interactions. However more advanced applications such as career coaching or planning a trip require a much more complex multi-turn dialogue. Current limitations of conversational systems have made it difficult to support applications that require personalization, customization and context dependent interactions. We tackle this challenging problem by using domain-independent AI planning to automatically create dialogue plans, customized to guide a dialogue towards achieving a given goal. The input includes a library of atomic dialogue actions, an initial state of the dialogue, and a goal. Dialogue plans are plugged into a dialogue system capable to orchestrate their execution. Use cases demonstrate the viability of the approach. Our work on dialogue planning has been integrated into a product, and it is in the process of being deployed into another.
http://arxiv.org/abs/1902.00771
In this paper we propose a convolutional neural network that is designed to upsample a series of sparse range measurements based on the contextual cues gleaned from a high resolution intensity image. Our approach draws inspiration from related work on super-resolution and in-painting. We propose a novel architecture that seeks to pull contextual cues separately from the intensity image and the depth features and then fuse them later in the network. We argue that this approach effectively exploits the relationship between the two modalities and produces accurate results while respecting salient image structures. We present experimental results to demonstrate that our approach is comparable with state of the art methods and generalizes well across multiple datasets.
http://arxiv.org/abs/1902.00761
Recognizing packaged grocery products based solely on appearance is still an open issue for modern computer vision systems due to peculiar challenges. Firstly, the number of different items to be recognized is huge (i.e., in the order of thousands) and rapidly changing over time. Moreover, there exist a significant domain shift between the images that should be recognized at test time, taken in stores by cheap cameras, and those available for training, usually just one or a few studio-quality images per product. We propose an end-to-end architecture comprising a GAN to address the domain shift at training time and a deep CNN trained on the samples generated by the GAN to learn an embedding of product images that enforces a hierarchy between product categories. At test time, we perform recognition by means of K-NN search against a database consisting of just one reference image per product. Experiments addressing recognition of products present in the training datasets as well as different ones unseen at training time show that our approach compares favourably to state-of-the-art methods on the grocery recognition task and generalize fairly well to similar ones.
http://arxiv.org/abs/1902.00760
Scientific competitions are important in robotics because they foster knowledge exchange and allow teams to test their research in unstandardized scenarios and compare result. In the field of service robotics its role becomes crucial. Competitions like RoboCup@Home bring robots to people, a fundamental step to integrate them into society. In this paper we summarize and discuss the differences between the achievements claimed by teams in their team description papers, and the results observed during the competition^1 from a qualitative perspective. We conclude with a set of important challenges to be conquered first in order to take robots to people’s homes. We believe that competitions are also an excellent opportunity to collect data of direct and unbiased interactions for further research. ^1 The authors belong to several teams who have participated in RoboCup@Home as early as 2007
http://arxiv.org/abs/1902.00758
Recently, progress has been made towards improving relational reasoning in machine learning field. Among existing models, graph neural networks (GNNs) is one of the most effective approaches for multi-hop relational reasoning. In fact, multi-hop relational reasoning is indispensable in many natural language processing tasks such as relation extraction. In this paper, we propose to generate the parameters of graph neural networks (GP-GNNs) according to natural language sentences, which enables GNNs to process relational reasoning on unstructured text inputs. We verify GP-GNNs in relation extraction from text. Experimental results on a human-annotated dataset and two distantly supervised datasets show that our model achieves significant improvements compared to baselines. We also perform a qualitative analysis to demonstrate that our model could discover more accurate relations by multi-hop relational reasoning.
http://arxiv.org/abs/1902.00756
On the one hand, speech is a key aspect to people’s communication. On the other, it is widely acknowledged that language proficiency is related to intelligence. Therefore, intelligent robots should be able to understand, at least, people’s orders within their application domain. These insights are not new in RoboCup@Home, but we lack of a long-term plan to evaluate this approach. In this paper we conduct a brief review of the achievements on automated speech recognition and natural language understanding in RoboCup@Home. Furthermore, we discuss main challenges to tackle in spoken human-robot interaction within the scope of this competition. Finally, we contribute by presenting a pipelined road map to engender research in the area of natural language understanding applied to domestic service robotics.
http://arxiv.org/abs/1902.00754
This work investigates the role of factors like training method, training corpus size and thematic relevance of texts in the performance of word embedding features on sentiment analysis of tweets, song lyrics, movie reviews and item reviews. We also explore specific training or post-processing methods that can be used to enhance the performance of word embeddings in certain tasks or domains. Our empirical observations indicate that models trained with multithematic texts that are large and rich in vocabulary are the best in answering syntactic and semantic word analogy questions. We further observe that influence of thematic relevance is stronger on movie and phone reviews, but weaker on tweets and lyrics. These two later domains are more sensitive to corpus size and training method, with Glove outperforming Word2vec. “Injecting” extra intelligence from lexicons or generating sentiment specific word embeddings are two prominent alternatives for increasing performance of word embedding features.
http://arxiv.org/abs/1902.00753
Fine-tuning large pre-trained models is an effective transfer mechanism in NLP. However, in the presence of many downstream tasks, fine-tuning is parameter inefficient: an entire new model is required for every task. As an alternative, we propose transfer with adapter modules. Adapter modules yield a compact and extensible model; they add only a few trainable parameters per task, and new tasks can be added without revisiting previous ones. The parameters of the original network remain fixed, yielding a high degree of parameter sharing. To demonstrate adapter’s effectiveness, we transfer the recently proposed BERT Transformer model to 26 diverse text classification tasks, including the GLUE benchmark. Adapters attain near state-of-the-art performance, whilst adding only a few parameters per task. On GLUE, we attain within 0.4% of the performance of full fine-tuning, adding only 3.6% parameters per task. By contrast, fine-tuning trains 100% of the parameters per task.
http://arxiv.org/abs/1902.00751
Rapid development of Internet technologies promotes traditional newspapers to report news on social networks. However, people on social networks may have different needs which naturally arises the question: whether can we analyze the influence of writing style on news quality automatically and assist writers in improving news quality? It’s challenging due to writing style and ‘quality’ are hard to measure. First, we use ‘popularity’ as the measure of ‘quality’. It is natural on social networks but brings new problems: popularity are also influenced by event and publisher. So we design two methods to alleviate their influence. Then, we proposed eight types of linguistic features (53 features in all) according eight writing guidelines and analyze their relationship with news quality. The experimental results show these linguistic features influence greatly on news quality. Based on it, we design a news quality assessment model on social network (SNQAM). SNQAM performs excellently on predicting quality, presenting interpretable quality score and giving accessible suggestions on how to improve it according to writing guidelines we referred to.
http://arxiv.org/abs/1902.00750
In this paper, we propose an online Multi-Object Tracking (MOT) approach which integrates the merits of single object tracking and data association methods in a unified framework to handle noisy detections and frequent interactions between targets. Specifically, for applying single object tracking in MOT, we introduce a cost-sensitive tracking loss based on the state-of-the-art visual tracker, which encourages the model to focus on hard negative distractors during online learning. For data association, we propose Dual Matching Attention Networks (DMAN) with both spatial and temporal attention mechanisms. The spatial attention module generates dual attention maps which enable the network to focus on the matching patterns of the input image pair, while the temporal attention module adaptively allocates different levels of attention to different samples in the tracklet to suppress noisy observations. Experimental results on the MOT benchmark datasets show that the proposed algorithm performs favorably against both online and offline trackers in terms of identity-preserving metrics.
http://arxiv.org/abs/1902.00749
A new conceptual foundation for the notion of “information” is proposed, based on the concept of a “distinction graph”: a graph in which two nodes are connected iff they cannot be distinguished by a particular observer. The “graphtropy” of a distinction graph is defined as the average connection probability of two nodes; in the case where the distinction graph is a composed of disconnected components that are fully connected subgraphs, this is equivalent to Ellerman’s logical entropy, which has straightforward relationships to Shannon entropy. Probabilistic distinction graphs and probabilistic graphtropy are also considered, as well as connections between graphtropy and thermodynamic and quantum entropy. The semantics of the Second Law of Thermodynamics and the Maximum Entropy Production Principle are unfolded in a novel way, via analysis of the cognitive processes underlying the making of distinction graphs This evokes an interpretation in which complex intelligence is seen to correspond to states of consciousness with intermediate graphtropy, which are associated with memory imperfections that violate the assumptions leading to derivation of the Second Law. In the case where nodes of a distinction graph are labeled by computable entities, graphtropy is shown to be monotonically related to the average algorithmic information of the nodes (relative to to the algorithmic information of the observer). A quantum-mechanical version of distinction graphs is considered, in which distinctions can exist in a superposed state; this yields to graphtropy as a measure of the impurity of a mixed state, and to a concept of “quangraphtropy.” Finally, a novel computational model called Dynamic Distinction Graphs (DDGs) is formulated, via enhancing distinction graphs with additional links expressing causal implications, enabling a distinction-based model of “observers.”
http://arxiv.org/abs/1902.00741
We present a method to train self-binarizing neural networks, that is, networks that evolve their weights and activations during training to become binary. To obtain similar binary networks, existing methods rely on the sign activation function. This function, however, has no gradients for non-zero values, which makes standard backpropagation impossible. To circumvent the difficulty of training a network relying on the sign activation function, these methods alternate between floating-point and binary representations of the network during training, which is sub-optimal and inefficient. We approach the binarization task by training on a unique representation involving a smooth activation function, which is iteratively sharpened during training until it becomes a binary representation equivalent to the sign activation function. Additionally, we introduce a new technique to perform binary batch normalization that simplifies the conventional batch normalization by transforming it into a simple comparison operation. This is unlike existing methods, which are forced to the retain the conventional floating-point-based batch normalization. Our binary networks, apart from displaying advantages of lower memory and computation as compared to conventional floating-point and binary networks, also show higher classification accuracy than existing state-of-the-art methods on multiple benchmark datasets.
http://arxiv.org/abs/1902.00730
Interactive Machine Learning is concerned with creating systems that operate in environments alongside humans to achieve a task. A typical use is to extend or amplify the capabilities of a human in cognitive or physical ways, requiring the machine to adapt to the users’ intentions and preferences. Often, this takes the form of a human operator providing some type of feedback to the user, which can be explicit feedback, implicit feedback, or a combination of both. Explicit feedback, such as through a mouse click, carries a high cognitive load. The focus of this study is to extend the current state of the art in interactive machine learning by demonstrating that agents can learn a human user’s behavior and adapt to preferences with a reduced amount of explicit human feedback in a mixed feedback setting. The learning agent perceives a value of its own behavior from hand gestures given via a spatial interface. This feedback mechanism is termed Spatial Interface Valuing. This method is evaluated experimentally in a simulated environment for a grasping task using a robotic arm with variable grip settings. Preliminary results indicate that learning agents using spatial interface valuing can learn a value function mapping spatial gestures to expected future rewards much more quickly as compared to those same agents just receiving explicit feedback, demonstrating that an agent perceiving feedback from a human user via a spatial interface can serve as an effective complement to existing approaches.
http://arxiv.org/abs/1902.00719
Previous research on quantum computing/mechanics and the arts has usually been in simulation. The small amount of work done in hardware or with actual physical systems has not utilized any of the advantages of quantum computation: the main advantage being the potential speed increase of quantum algorithms. This paper introduces a way of utilizing Grover’s algorithm - which has been shown to provide a quadratic speed-up over its classical equivalent - in algorithmic rule-based music composition. The system introduced - qgMuse - is simple but scalable. It lays some groundwork for new ways of addressing a significant problem in computer music research: unstructured random search for desired music features. Example melodies are composed using qgMuse using the ibmqx4 quantum hardware, and the paper concludes with discussion on how such an approach can grow with the improvement of quantum computer hardware and software.
http://arxiv.org/abs/1902.04237
The social media platform provides an opportunity to gain valuable insights into user behaviour. Users mimic their internal feelings and emotions in a disinhibited fashion using natural language. Techniques in Natural Language Processing have helped researchers decipher standard documents and cull together inferences from massive amount of data. A representative corpus is a prerequisite for NLP and one of the challenges we face today is the non-standard and noisy language that exists on the internet. Our work focuses on building a corpus from social media that is focused on detecting mental illness. We use depression as a case study and demonstrate the effectiveness of using such a corpus for helping practitioners detect such cases. Our results show a high correlation between our Social Media Corpus and the standard corpus for depression.
http://arxiv.org/abs/1902.00702
Machine Learning is an important sub-field of the Artificial Intelligence and it has been become a very critical task to train Machine Learning techniques via effective method or techniques. Recently, researchers try to use alternative techniques to improve ability of Machine Learning techniques. Moving from the explanations, objective of this study is to introduce a novel SVM-CoDOA (Cognitive Development Optimization Algorithm trained Support Vector Machines) system for general medical diagnosis. In detail, the system consists of a SVM, which is trained by CoDOA, a newly developed optimization algorithm. As it is known, use of optimization algorithms is an essential task to train and improve Machine Learning techniques. In this sense, the study has provided a medical diagnosis oriented problem scope in order to show effectiveness of the SVM-CoDOA hybrid formation.
http://arxiv.org/abs/1902.00685
Recent advances in Big Data has prompted health care practitioners to utilize the data available on social media to discern sentiment and emotions expression. Health Informatics and Clinical Analytics depend heavily on information gathered from diverse sources. Traditionally, a healthcare practitioner will ask a patient to fill out a questionnaire that will form the basis of diagnosing the medical condition. However, medical practitioners have access to many sources of data including the patients writings on various media. Natural Language Processing (NLP) allows researchers to gather such data and analyze it to glean the underlying meaning of such writings. The field of sentiment analysis (applied to many other domains) depend heavily on techniques utilized by NLP. This work will look into various prevalent theories underlying the NLP field and how they can be leveraged to gather users sentiments on social media. Such sentiments can be culled over a period of time thus minimizing the errors introduced by data input and other stressors. Furthermore, we look at some applications of sentiment analysis and application of NLP to mental health. The reader will also learn about the NLTK toolkit that implements various NLP theories and how they can make the data scavenging process a lot easier.
http://arxiv.org/abs/1902.00679
Animal behavior is not driven simply by its current observations, but is strongly influenced by internal states. Estimating the structure of these internal states is crucial for understanding the neural basis of behavior. In principle, internal states can be estimated by inverting behavior models, as in inverse model-based Reinforcement Learning. However, this requires careful parameterization and risks model-mismatch to the animal. Here we take a data-driven approach to infer latent states directly from observations of behavior, using a partially observable switching semi-Markov process. This process has two elements critical for capturing animal behavior: it captures non-exponential distribution of times between observations, and transitions between latent states depend on the animal’s actions, features that require more complex non-markovian models to represent. To demonstrate the utility of our approach, we apply it to the observations of a simulated optimal agent performing a foraging task, and find that latent dynamics extracted by the model has correspondences with the belief dynamics of the agent. Finally, we apply our model to identify latent states in the behaviors of monkey performing a foraging task, and find clusters of latent states that identify periods of time consistent with expectant waiting. This data-driven behavioral model will be valuable for inferring latent cognitive states, and thereby for measuring neural representations of those states.
http://arxiv.org/abs/1902.00673
Existing graph- and hypergraph-based algorithms for document summarization represent the sentences of a corpus as the nodes of a graph or a hypergraph in which the edges represent relationships of lexical similarities between sentences. Each sentence of the corpus is then scored individually, using popular node ranking algorithms, and a summary is produced by extracting highly scored sentences. This approach fails to select a subset of jointly relevant sentences and it may produce redundant summaries that are missing important topics of the corpus. To alleviate this issue, a new hypergraph-based summarizer is proposed in this paper, in which each node is a sentence and each hyperedge is a theme, namely a group of sentences sharing a topic. Themes are weighted in terms of their prominence in the corpus and their relevance to a user-defined query. It is further shown that the problem of identifying a subset of sentences covering the relevant themes of the corpus is equivalent to that of finding a hypergraph transversal in our theme-based hypergraph. Two extensions of the notion of hypergraph transversal are proposed for the purpose of summarization, and polynomial time algorithms building on the theory of submodular functions are proposed for solving the associated discrete optimization problems. The worst-case time complexity of the proposed algorithms is squared in the number of terms, which makes it cheaper than the existing hypergraph-based methods. A thorough comparative analysis with related models on DUC benchmark datasets demonstrates the effectiveness of our approach, which outperforms existing graph- or hypergraph-based methods by at least 6% of ROUGE-SU4 score.
http://arxiv.org/abs/1902.00672
The visual world we sense, interpret and interact everyday is a complex composition of interleaved physical entities. Therefore, it is a very challenging task to generate vivid scenes of similar complexity using computers. In this work, we present a scene generation framework based on Generative Adversarial Networks (GANs) to sequentially compose a scene, breaking down the underlying problem into smaller ones. Different than the existing approaches, our framework offers an explicit control over the elements of a scene through separate background and foreground generators. Starting with an initially generated background, foreground objects then populate the scene one-by-one in a sequential manner. Via quantitative and qualitative experiments on a subset of the MS-COCO dataset, we show that our proposed framework produces not only more diverse images but also copes better with affine transformations and occlusion artifacts of foreground objects than its counterparts.
http://arxiv.org/abs/1902.00671
In this paper, we propose a novel model with a hierarchical photo-scene encoder and a reconstructor for the task of album storytelling. The photo-scene encoder contains two sub-encoders, namely the photo and scene encoders, which are stacked together and behave hierarchically to fully exploit the structure information of the photos within an album. Specifically, the photo encoder generates semantic representation for each photo while exploiting temporal relationships among them. The scene encoder, relying on the obtained photo representations, is responsible for detecting the scene changes and generating scene representations. Subsequently, the decoder dynamically and attentively summarizes the encoded photo and scene representations to generate a sequence of album representations, based on which a story consisting of multiple coherent sentences is generated. In order to fully extract the useful semantic information from an album, a reconstructor is employed to reproduce the summarized album representations based on the hidden states of the decoder. The proposed model can be trained in an end-to-end manner, which results in an improved performance over the state-of-the-arts on the public visual storytelling (VIST) dataset. Ablation studies further demonstrate the effectiveness of the proposed hierarchical photo-scene encoder and reconstructor.
http://arxiv.org/abs/1902.00669
Fairness, through its many forms and definitions, has become an important issue facing the machine learning community. In this work, we consider how to incorporate group fairness constraints in kernel regression methods, applicable to Gaussian processes, support vector machines, neural network regression and decision tree regression. Further, we focus on examining the effect of incorporating these constraints in decision tree regression, with direct applications to random forests and boosted trees amongst other widespread popular inference techniques. We show that the order of complexity of memory and computation is preserved for such models and tightly bound the expected perturbations to the model in terms of the number of leaves of the trees. Importantly, the approach works on trained models and hence can be easily applied to models in current use and group labels are only required on training data.
http://arxiv.org/abs/1810.05041
Deep language models learning a hierarchical representation proved to be a powerful tool for natural language processing, text mining and information retrieval. However, representations that perform well for retrieval must capture semantic meaning at different levels of abstraction or context-scopes. In this paper, we propose a new method to generate multi-resolution word embedding representing documents at multiple resolutions in term of context-scopes. In order to investigate its performance, we use the Stanford Question Answering Dataset (SQuAD) and the Question Answering by Search And Reading (QUASAR) in an open-domain question-answering setting, where the first task is to find documents useful for answering a given question. To this end, we first compare the quality of various text-embedding methods for retrieval performance and give an extensive empirical comparison with the performance of various non-augmented base embeddings with and without multi-resolution representation. We argue that multi-resolution word embeddings are consistently superior to the original counterparts and deep residual neural models specifically trained for retrieval purposes can yield further significant gains when they are used for augmenting those embeddings.
http://arxiv.org/abs/1902.00663
Acoustic-to-word (A2W) models that allow direct mapping from acoustic signals to word sequences are an appealing approach to end-to-end automatic speech recognition due to their simplicity. However, prior works have shown that modelling A2W typically encounters issues of data sparsity that prevent training such a model directly. So far, pre-training initialization is the only approach proposed to deal with this issue. In this work, we propose to build a shared neural network and optimize A2W and conventional hybrid models in a multi-task manner. Our results show that training an A2W model is much more stable with our multi-task model without pre-training initialization, and results in a significant improvement compared to a baseline model. Experiments also reveal that the performance of a hybrid acoustic model can be further improved when jointly training with a sequence-level optimization criterion such as acoustic-to-word.
http://arxiv.org/abs/1902.01951