February 1, 2020

2894 words 14 mins read

Paper Group AWR 364

Investigating Sports Commentator Bias within a Large Corpus of American Football Broadcasts. Emulating Human Developmental Stages with Bayesian Neural Networks. 3D U-Net Based Brain Tumor Segmentation and Survival Days Prediction. Gromov-Wasserstein Averaging in a Riemannian Framework. Sparse Feature Selection in Kernel Discriminant Analysis via Op …

Investigating Sports Commentator Bias within a Large Corpus of American Football Broadcasts


Title	Investigating Sports Commentator Bias within a Large Corpus of American Football Broadcasts
Authors	Jack Merullo, Luke Yeh, Abram Handler, Alvin Grissom II, Brendan O’Connor, Mohit Iyyer
Abstract	Sports broadcasters inject drama into play-by-play commentary by building team and player narratives through subjective analyses and anecdotes. Prior studies based on small datasets and manual coding show that such theatrics evince commentator bias in sports broadcasts. To examine this phenomenon, we assemble FOOTBALL, which contains 1,455 broadcast transcripts from American football games across six decades that are automatically annotated with 250K player mentions and linked with racial metadata. We identify major confounding factors for researchers examining racial bias in FOOTBALL, and perform a computational analysis that supports conclusions from prior social science studies.
Tasks
Published	2019-09-07
URL	https://arxiv.org/abs/1909.03343v4
PDF	https://arxiv.org/pdf/1909.03343v4.pdf
PWC	https://paperswithcode.com/paper/investigating-sports-commentator-bias-within
Repo	https://github.com/jmerullo/football
Framework	none

Emulating Human Developmental Stages with Bayesian Neural Networks


Title	Emulating Human Developmental Stages with Bayesian Neural Networks
Authors	Marcel Binz, Dominik Endres
Abstract	We compare the acquisition of knowledge in humans and machines. Research from the field of developmental psychology indicates, that human-employed hypothesis are initially guided by simple rules, before evolving into more complex theories. This observation is shared across many tasks and domains. We investigate whether stages of development in artificial learning systems are based on the same characteristics. We operationalize developmental stages as the size of the data-set, on which the artificial system is trained. For our analysis we look at the developmental progress of Bayesian Neural Networks on three different data-sets, including occlusion, support and quantity comparison tasks. We compare the results with prior research from developmental psychology and find agreement between the family of optimized models and pattern of development observed in infants and children on all three tasks, indicating common principles for the acquisition of knowledge.
Tasks
Published	2019-02-20
URL	http://arxiv.org/abs/1902.07579v1
PDF	http://arxiv.org/pdf/1902.07579v1.pdf
PWC	https://paperswithcode.com/paper/emulating-human-developmental-stages-with
Repo	https://github.com/marcelbinz/Developmental-Stages-of-BNNs
Framework	pytorch

3D U-Net Based Brain Tumor Segmentation and Survival Days Prediction


Title	3D U-Net Based Brain Tumor Segmentation and Survival Days Prediction
Authors	Feifan Wang, Runzhou Jiang, Liqin Zheng, Chun Meng, Bharat Biswal
Abstract	Past few years have witnessed the prevalence of deep learning in many application scenarios, among which is medical image processing. Diagnosis and treatment of brain tumors requires an accurate and reliable segmentation of brain tumors as a prerequisite. However, such work conventionally requires brain surgeons significant amount of time. Computer vision techniques could provide surgeons a relief from the tedious marking procedure. In this paper, a 3D U-net based deep learning model has been trained with the help of brain-wise normalization and patching strategies for the brain tumor segmentation task in the BraTS 2019 competition. Dice coefficients for enhancing tumor, tumor core, and the whole tumor are 0.737, 0.807 and 0.894 respectively on the validation dataset. These three values on the test dataset are 0.778, 0.798 and 0.852. Furthermore, numerical features including ratio of tumor size to brain size and the area of tumor surface as well as age of subjects are extracted from predicted tumor labels and have been used for the overall survival days prediction task. The accuracy could be 0.448 on the validation dataset, and 0.551 on the final test dataset.
Tasks	Brain Tumor Segmentation
Published	2019-09-15
URL	https://arxiv.org/abs/1909.12901v2
PDF	https://arxiv.org/pdf/1909.12901v2.pdf
PWC	https://paperswithcode.com/paper/brain-wise-tumor-segmentation-and-patient
Repo	https://github.com/woodywff/brats_2019
Framework	tf

Gromov-Wasserstein Averaging in a Riemannian Framework


Title	Gromov-Wasserstein Averaging in a Riemannian Framework
Authors	Samir Chowdhury, Tom Needham
Abstract	We introduce a theoretical framework for performing statistical tasks—including, but not limited to, averaging and principal component analysis—on the space of (possibly asymmetric) matrices with arbitrary entries and sizes. This is carried out under the lens of the Gromov-Wasserstein (GW) distance, and our methods translate the Riemannian framework of GW distances developed by Sturm into practical, implementable tools for network data analysis. Our methods are illustrated on datasets of asymmetric stochastic blockmodel networks and planar shapes viewed as metric spaces. On the theoretical front, we supplement the work of Sturm by producing additional results on the tangent structure of this “space of spaces”, as well as on the gradient flow of the Fr'{e}chet functional on this space.
Tasks
Published	2019-10-10
URL	https://arxiv.org/abs/1910.04308v1
PDF	https://arxiv.org/pdf/1910.04308v1.pdf
PWC	https://paperswithcode.com/paper/gromov-wasserstein-averaging-in-a-riemannian
Repo	https://github.com/trneedham/gromov-wasserstein-statistics
Framework	none

Sparse Feature Selection in Kernel Discriminant Analysis via Optimal Scoring


Title	Sparse Feature Selection in Kernel Discriminant Analysis via Optimal Scoring
Authors	Alexander F. Lapanowski, Irina Gaynanova
Abstract	We consider the two-group classification problem and propose a kernel classifier based on the optimal scoring framework. Unlike previous approaches, we provide theoretical guarantees on the expected risk consistency of the method. We also allow for feature selection by imposing structured sparsity using weighted kernels. We propose fully-automated methods for selection of all tuning parameters, and in particular adapt kernel shrinkage ideas for ridge parameter selection. Numerical studies demonstrate the superior classification performance of the proposed approach compared to existing nonparametric classifiers.
Tasks	Feature Selection
Published	2019-02-12
URL	http://arxiv.org/abs/1902.04248v1
PDF	http://arxiv.org/pdf/1902.04248v1.pdf
PWC	https://paperswithcode.com/paper/sparse-feature-selection-in-kernel
Repo	https://github.com/aflapan/sparseKOS
Framework	none

HINT: Hierarchical Invertible Neural Transport for Density Estimation and Bayesian Inference


Title	HINT: Hierarchical Invertible Neural Transport for Density Estimation and Bayesian Inference
Authors	Jakob Kruse, Gianluca Detommaso, Robert Scheichl, Ullrich Köthe
Abstract	A large proportion of recent invertible neural architectures is based on a coupling block design. It operates by dividing incoming variables into two sub-spaces, one of which parameterizes an easily invertible (usually affine) transformation that is applied to the other. While the Jacobian of such a transformation is triangular, it is very sparse and thus may lack expressiveness. This work presents a simple remedy by noting that (affine) coupling can be repeated recursively within the resulting sub-spaces, leading to an efficiently invertible block with dense triangular Jacobian. By formulating our recursive coupling scheme via a hierarchical architecture, HINT allows sampling from a joint distribution p(y,x) and the corresponding posterior p(xy) using a single invertible network. We demonstrate the power of our method for density estimation and Bayesian inference on a novel data set of 2D shapes in Fourier parameterization, which enables consistent visualization of samples for different dimensionalities.
Tasks	Bayesian Inference, Density Estimation
Published	2019-05-25
URL	https://arxiv.org/abs/1905.10687v3
PDF	https://arxiv.org/pdf/1905.10687v3.pdf
PWC	https://paperswithcode.com/paper/hint-hierarchical-invertible-neural-transport
Repo	https://github.com/VLL-HD/HINT
Framework	pytorch

Point-Based Multi-View Stereo Network


Title	Point-Based Multi-View Stereo Network
Authors	Rui Chen, Songfang Han, Jing Xu, Hao Su
Abstract	We introduce Point-MVSNet, a novel point-based deep framework for multi-view stereo (MVS). Distinct from existing cost volume approaches, our method directly processes the target scene as point clouds. More specifically, our method predicts the depth in a coarse-to-fine manner. We first generate a coarse depth map, convert it into a point cloud and refine the point cloud iteratively by estimating the residual between the depth of the current iteration and that of the ground truth. Our network leverages 3D geometry priors and 2D texture information jointly and effectively by fusing them into a feature-augmented point cloud, and processes the point cloud to estimate the 3D flow for each point. This point-based architecture allows higher accuracy, more computational efficiency and more flexibility than cost-volume-based counterparts. Experimental results show that our approach achieves a significant improvement in reconstruction quality compared with state-of-the-art methods on the DTU and the Tanks and Temples dataset. Our source code and trained models are available at https://github.com/callmeray/PointMVSNet .
Tasks
Published	2019-08-12
URL	https://arxiv.org/abs/1908.04422v1
PDF	https://arxiv.org/pdf/1908.04422v1.pdf
PWC	https://paperswithcode.com/paper/point-based-multi-view-stereo-network
Repo	https://github.com/callmeray/PointMVSNet
Framework	pytorch

Discrete Flows: Invertible Generative Models of Discrete Data


Title	Discrete Flows: Invertible Generative Models of Discrete Data
Authors	Dustin Tran, Keyon Vafa, Kumar Krishna Agrawal, Laurent Dinh, Ben Poole
Abstract	While normalizing flows have led to significant advances in modeling high-dimensional continuous distributions, their applicability to discrete distributions remains unknown. In this paper, we show that flows can in fact be extended to discrete events—and under a simple change-of-variables formula not requiring log-determinant-Jacobian computations. Discrete flows have numerous applications. We consider two flow architectures: discrete autoregressive flows that enable bidirectionality, allowing, for example, tokens in text to depend on both left-to-right and right-to-left contexts in an exact language model; and discrete bipartite flows that enable efficient non-autoregressive generation as in RealNVP. Empirically, we find that discrete autoregressive flows outperform autoregressive baselines on synthetic discrete distributions, an addition task, and Potts models; and bipartite flows can obtain competitive performance with autoregressive baselines on character-level language modeling for Penn Tree Bank and text8.
Tasks	Language Modelling
Published	2019-05-24
URL	https://arxiv.org/abs/1905.10347v1
PDF	https://arxiv.org/pdf/1905.10347v1.pdf
PWC	https://paperswithcode.com/paper/discrete-flows-invertible-generative-models
Repo	https://github.com/TrentBrick/PyTorchDiscreteFlows
Framework	pytorch

Combining Sentiment Lexica with a Multi-View Variational Autoencoder


Title	Combining Sentiment Lexica with a Multi-View Variational Autoencoder
Authors	Alexander Hoyle, Lawrence Wolf-Sonkin, Hanna Wallach, Ryan Cotterell, Isabelle Augenstein
Abstract	When assigning quantitative labels to a dataset, different methodologies may rely on different scales. In particular, when assigning polarities to words in a sentiment lexicon, annotators may use binary, categorical, or continuous labels. Naturally, it is of interest to unify these labels from disparate scales to both achieve maximal coverage over words and to create a single, more robust sentiment lexicon while retaining scale coherence. We introduce a generative model of sentiment lexica to combine disparate scales into a common latent representation. We realize this model with a novel multi-view variational autoencoder (VAE), called SentiVAE. We evaluate our approach via a downstream text classification task involving nine English-Language sentiment analysis datasets; our representation outperforms six individual sentiment lexica, as well as a straightforward combination thereof.
Tasks	Sentiment Analysis, Text Classification
Published	2019-04-05
URL	http://arxiv.org/abs/1904.02839v1
PDF	http://arxiv.org/pdf/1904.02839v1.pdf
PWC	https://paperswithcode.com/paper/combining-sentiment-lexica-with-a-multi-view
Repo	https://github.com/ahoho/SentiVAE
Framework	pytorch

HoMM: Higher-order Moment Matching for Unsupervised Domain Adaptation


Title	HoMM: Higher-order Moment Matching for Unsupervised Domain Adaptation
Authors	Chao Chen, Zhihang Fu, Zhihong Chen, Sheng Jin, Zhaowei Cheng, Xinyu Jin, Xian-Sheng Hua
Abstract	Minimizing the discrepancy of feature distributions between different domains is one of the most promising directions in unsupervised domain adaptation. From the perspective of distribution matching, most existing discrepancy-based methods are designed to match the second-order or lower statistics, which however, have limited expression of statistical characteristic for non-Gaussian distributions. In this work, we explore the benefits of using higher-order statistics (mainly refer to third-order and fourth-order statistics) for domain matching. We propose a Higher-order Moment Matching (HoMM) method, and further extend the HoMM into reproducing kernel Hilbert spaces (RKHS). In particular, our proposed HoMM can perform arbitrary-order moment tensor matching, we show that the first-order HoMM is equivalent to Maximum Mean Discrepancy (MMD) and the second-order HoMM is equivalent to Correlation Alignment (CORAL). Moreover, the third-order and the fourth-order moment tensor matching are expected to perform comprehensive domain alignment as higher-order statistics can approximate more complex, non-Gaussian distributions. Besides, we also exploit the pseudo-labeled target samples to learn discriminative representations in the target domain, which further improves the transfer performance. Extensive experiments are conducted, showing that our proposed HoMM consistently outperforms the existing moment matching methods by a large margin. Codes are available at \url{https://github.com/chenchao666/HoMM-Master}
Tasks	Domain Adaptation, Unsupervised Domain Adaptation
Published	2019-12-27
URL	https://arxiv.org/abs/1912.11976v1
PDF	https://arxiv.org/pdf/1912.11976v1.pdf
PWC	https://paperswithcode.com/paper/homm-higher-order-moment-matching-for
Repo	https://github.com/chenchao666/HoMM-Master
Framework	tf

Coursera Corpus Mining and Multistage Fine-Tuning for Improving Lectures Translation


Title	Coursera Corpus Mining and Multistage Fine-Tuning for Improving Lectures Translation
Authors	Haiyue Song, Raj Dabre, Atsushi Fujita, Sadao Kurohashi
Abstract	Lectures translation is a case of spoken language translation and there is a lack of publicly available parallel corpora for this purpose. To address this, we examine a language independent framework for parallel corpus mining which is a quick and effective way to mine a parallel corpus from publicly available lectures at Coursera. Our approach determines sentence alignments, relying on machine translation and cosine similarity over continuous-space sentence representations. We also show how to use the resulting corpora in a multistage fine-tuning based domain adaptation for high-quality lectures translation. For Japanese–English lectures translation, we extracted parallel data of approximately 40,000 lines and created development and test sets through manual filtering for benchmarking translation performance. We demonstrate that the mined corpus greatly enhances the quality of translation when used in conjunction with out-of-domain parallel corpora via multistage training. This paper also suggests some guidelines to gather and clean corpora, mine parallel sentences, address noise in the mined data, and create high-quality evaluation splits. For the sake of reproducibility, we will release our code for parallel data creation.
Tasks	Domain Adaptation, Machine Translation, Parallel Corpus Mining
Published	2019-12-26
URL	https://arxiv.org/abs/1912.11739v2
PDF	https://arxiv.org/pdf/1912.11739v2.pdf
PWC	https://paperswithcode.com/paper/coursera-corpus-mining-and-multistage-fine
Repo	https://github.com/shyyhs/CourseraParallelCorpusMining
Framework	none

Adversarial Domain Adaptation with Domain Mixup


Title	Adversarial Domain Adaptation with Domain Mixup
Authors	Minghao Xu, Jian Zhang, Bingbing Ni, Teng Li, Chengjie Wang, Qi Tian, Wenjun Zhang
Abstract	Recent works on domain adaptation reveal the effectiveness of adversarial learning on filling the discrepancy between source and target domains. However, two common limitations exist in current adversarial-learning-based methods. First, samples from two domains alone are not sufficient to ensure domain-invariance at most part of latent space. Second, the domain discriminator involved in these methods can only judge real or fake with the guidance of hard label, while it is more reasonable to use soft scores to evaluate the generated images or features, i.e., to fully utilize the inter-domain information. In this paper, we present adversarial domain adaptation with domain mixup (DM-ADA), which guarantees domain-invariance in a more continuous latent space and guides the domain discriminator in judging samples’ difference relative to source and target domains. Domain mixup is jointly conducted on pixel and feature level to improve the robustness of models. Extensive experiments prove that the proposed approach can achieve superior performance on tasks with various degrees of domain shift and data complexity.
Tasks	Domain Adaptation
Published	2019-12-04
URL	https://arxiv.org/abs/1912.01805v1
PDF	https://arxiv.org/pdf/1912.01805v1.pdf
PWC	https://paperswithcode.com/paper/adversarial-domain-adaptation-with-domain
Repo	https://github.com/ChrisAllenMing/Mixup_for_UDA
Framework	pytorch

Cross-Domain Complementary Learning with Synthetic Data for Multi-Person Part Segmentation


Title	Cross-Domain Complementary Learning with Synthetic Data for Multi-Person Part Segmentation
Authors	Kevin Lin, Lijuan Wang, Kun Luo, Yinpeng Chen, Zicheng Liu, Ming-Ting Sun
Abstract	The success of supervised deep learning depends on the training labels. However, data labeling at pixel-level is very expensive, and people have been exploring synthetic data as an alternative. Even though it is easy to generate labels for synthetic data, the quality gap makes it challenging to transfer knowledge from synthetic data to real data. In this paper, we propose a novel technique, called cross-domain complementary learning that takes advantage of the rich variations of real data and the easily obtainable labels of synthetic data to learn multi-person part segmentation on real images without any human-annotated segmentation labels. To make sure the synthetic data and real data are aligned in a common latent space, we use an auxiliary task of human pose estimation to bridge the two domains. Without any real part segmentation training data, our method performs comparably to several supervised state-of-the-art approaches which require real part segmentation training data on Pascal-Person-Parts and COCO-DensePose datasets. We further demonstrate the generalizability of our method on predicting novel keypoints in the wild where no real data labels are available for the novel keypoints.
Tasks	Domain Adaptation, Human Part Segmentation, Multi-Human Parsing, Pose Estimation
Published	2019-07-11
URL	https://arxiv.org/abs/1907.05193v1
PDF	https://arxiv.org/pdf/1907.05193v1.pdf
PWC	https://paperswithcode.com/paper/cross-domain-complementary-learning-with
Repo	https://github.com/kevinlin311tw/CDCL-human-part-segmentation
Framework	tf

Random Projection in Neural Episodic Control


Title	Random Projection in Neural Episodic Control
Authors	Daichi Nishio, Satoshi Yamane
Abstract	End-to-end deep reinforcement learning has enabled agents to learn with little preprocessing by humans. However, it is still difficult to learn stably and efficiently because the learning method usually uses a nonlinear function approximation. Neural Episodic Control (NEC), which has been proposed in order to improve sample efficiency, is able to learn stably by estimating action values using a non-parametric method. In this paper, we propose an architecture that incorporates random projection into NEC to train with more stability. In addition, we verify the effectiveness of our architecture by Atari’s five games. The main idea is to reduce the number of parameters that have to learn by replacing neural networks with random projection in order to reduce dimensions while keeping the learning end-to-end.
Tasks
Published	2019-04-03
URL	http://arxiv.org/abs/1904.01790v2
PDF	http://arxiv.org/pdf/1904.01790v2.pdf
PWC	https://paperswithcode.com/paper/random-projection-in-neural-episodic-control
Repo	https://github.com/dnishio/NEC-RP
Framework	none

Detecting Urban Changes with Recurrent Neural Networks from Multitemporal Sentinel-2 Data


Title	Detecting Urban Changes with Recurrent Neural Networks from Multitemporal Sentinel-2 Data
Authors	Maria Papadomanolaki, Sagar Verma, Maria Vakalopoulou, Siddharth Gupta, Konstantinos Karantzalos
Abstract	\begin{abstract} The advent of multitemporal high resolution data, like the Copernicus Sentinel-2, has enhanced significantly the potential of monitoring the earth’s surface and environmental dynamics. In this paper, we present a novel deep learning framework for urban change detection which combines state-of-the-art fully convolutional networks (similar to U-Net) for feature representation and powerful recurrent networks (such as LSTMs) for temporal modeling. We report our results on the recently publicly available bi-temporal Onera Satellite Change Detection (OSCD) Sentinel-2 dataset, enhancing the temporal information with additional images of the same region on different dates. Moreover, we evaluate the performance of the recurrent networks as well as the use of the additional dates on the unseen test-set using an ensemble cross-validation strategy. All the developed models during the validation phase have scored an overall accuracy of more than 95%, while the use of LSTMs and further temporal information, boost the F1 rate of the change class by an additional 1.5%.
Tasks
Published	2019-10-17
URL	https://arxiv.org/abs/1910.07778v1
PDF	https://arxiv.org/pdf/1910.07778v1.pdf
PWC	https://paperswithcode.com/paper/detecting-urban-changes-with-recurrent-neural
Repo	https://github.com/mpapadomanolaki/UNetLSTM
Framework	pytorch