January 25, 2020

3244 words 16 mins read

Paper Group NAWR 39

Paper Group NAWR 39

The Forward-Backward Embedding of Directed Graphs. Disentangled behavioural representations. Viewport Proposal CNN for 360deg Video Quality Assessment. Novel positional encodings to enable tree-based transformers. Sample Adaptive MCMC. Gradient Matching Generative Networks for Zero-Shot Learning. QTUNA: A Corpus for Understanding How Speakers Use Q …

The Forward-Backward Embedding of Directed Graphs

Title The Forward-Backward Embedding of Directed Graphs
Authors Thomas Bonald, Nathan De Lara
Abstract We introduce a novel embedding of directed graphs derived from the singular value decomposition (SVD) of the normalized adjacency matrix. Specifically, we show that, after proper normalization of the singular vectors, the distances between vectors in the embedding space are proportional to the mean commute times between the corresponding nodes by a forward-backward random walk in the graph, which follows the edges alternately in forward and backward directions. In particular, two nodes having many common successors in the graph tend to be represented by close vectors in the embedding space. More formally, we prove that our representation of the graph is equivalent to the spectral embedding of some co-citation graph, where nodes are linked with respect to their common set of successors in the original graph. The interest of our approach is that it does not require to build this co-citation graph, which is typically much denser than the original graph. Experiments on real datasets show the efficiency of the approach.
Tasks
Published 2019-05-01
URL https://openreview.net/forum?id=S1E64jC5tm
PDF https://openreview.net/pdf?id=S1E64jC5tm
PWC https://paperswithcode.com/paper/the-forward-backward-embedding-of-directed
Repo https://github.com/tbonald/directed
Framework none

Disentangled behavioural representations

Title Disentangled behavioural representations
Authors Amir Dezfouli, Hassan Ashtiani, Omar Ghattas, Richard Nock, Peter Dayan, Cheng Soon Ong
Abstract Individual characteristics in human decision-making are often quantified by fitting a parametric cognitive model to subjects’ behavior and then studying differences between them in the associated parameter space. However, these models often fit behavior more poorly than recurrent neural networks (RNNs), which are more flexible and make fewer assumptions about the underlying decision-making processes. Unfortunately, the parameter and latent activity spaces of RNNs are generally high-dimensional and uninterpretable, making it hard to use them to study individual differences. Here, we show how to benefit from the flexibility of RNNs while representing individual differences in a low-dimensional and interpretable space. To achieve this, we propose a novel end-to-end learning framework in which an encoder is trained to map the behavior of subjects into a low-dimensional latent space. These low-dimensional representations are used to generate the parameters of individual RNNs corresponding to the decision-making process of each subject. We introduce terms into the loss function that ensure that the latent dimensions are informative and disentangled, i.e., encouraged to have distinct effects on behavior. This allows them to align with separate facets of individual differences. We illustrate the performance of our framework on synthetic data as well as a dataset including the behavior of patients with psychiatric disorders.
Tasks Decision Making
Published 2019-12-01
URL http://papers.nips.cc/paper/8497-disentangled-behavioural-representations
PDF http://papers.nips.cc/paper/8497-disentangled-behavioural-representations.pdf
PWC https://paperswithcode.com/paper/disentangled-behavioural-representations
Repo https://github.com/adezfouli/rnn_hypercoder
Framework tf

Viewport Proposal CNN for 360deg Video Quality Assessment

Title Viewport Proposal CNN for 360deg Video Quality Assessment
Authors Chen Li, Mai Xu, Lai Jiang, Shanyi Zhang, Xiaoming Tao
Abstract Recent years have witnessed the growing interest in visual quality assessment (VQA) for 360deg video. Unfortunately, the existing VQA approaches do not consider the facts that: 1) Observers only see viewports of 360deg video, rather than patches or whole 360deg frames. 2) Within the viewport, only salient regions can be perceived by observers with high resolution. Thus, this paper proposes a viewport-based convolutional neural network (V-CNN) approach for VQA on 360deg video, considering both auxiliary tasks of viewport proposal and viewport saliency prediction. Our V-CNN approach is composed of two stages, i.e., viewport proposal and VQA. In the first stage, the viewport proposal network (VP-net) is developed to yield several potential viewports, seen as the first auxiliary task. In the second stage, a viewport quality network (VQ-net) is designed to rate the VQA score for each proposed viewport, in which the saliency map of the viewport is predicted and then utilized in VQA score rating. Consequently, another auxiliary task of viewport saliency prediction can be achieved. More importantly, the main task of VQA on 360deg video can be accomplished via integrating the VQA scores of all viewports. The experiments validate the effectiveness of our V-CNN approach in significantly advancing the state-of-the-art performance of VQA on 360deg video. In addition, our approach achieves comparable performance in two auxiliary tasks. The code of our V-CNN approach is available at https://github.com/Archer-Tatsu/V-CNN.
Tasks Saliency Prediction, Video Quality Assessment, Visual Question Answering
Published 2019-06-01
URL http://openaccess.thecvf.com/content_CVPR_2019/html/Li_Viewport_Proposal_CNN_for_360deg_Video_Quality_Assessment_CVPR_2019_paper.html
PDF http://openaccess.thecvf.com/content_CVPR_2019/papers/Li_Viewport_Proposal_CNN_for_360deg_Video_Quality_Assessment_CVPR_2019_paper.pdf
PWC https://paperswithcode.com/paper/viewport-proposal-cnn-for-360deg-video
Repo https://github.com/Archer-Tatsu/V-CNN
Framework pytorch

Novel positional encodings to enable tree-based transformers

Title Novel positional encodings to enable tree-based transformers
Authors Vighnesh Shiv, Chris Quirk
Abstract Neural models optimized for tree-based problems are of great value in tasks like SQL query extraction and program synthesis. On sequence-structured data, transformers have been shown to learn relationships across arbitrary pairs of positions more reliably than recurrent models. Motivated by this property, we propose a method to extend transformers to tree-structured data, enabling sequence-to-tree, tree-to-sequence, and tree-to-tree mappings. Our approach abstracts the transformer’s sinusoidal positional encodings, allowing us to instead use a novel positional encoding scheme to represent node positions within trees. We evaluated our model in tree-to-tree program translation and sequence-to-tree semantic parsing settings, achieving superior performance over both sequence-to-sequence transformers and state-of-the-art tree-based LSTMs on several datasets. In particular, our results include a 22% absolute increase in accuracy on a JavaScript to CoffeeScript translation dataset.
Tasks Program Synthesis, Semantic Parsing
Published 2019-12-01
URL http://papers.nips.cc/paper/9376-novel-positional-encodings-to-enable-tree-based-transformers
PDF http://papers.nips.cc/paper/9376-novel-positional-encodings-to-enable-tree-based-transformers.pdf
PWC https://paperswithcode.com/paper/novel-positional-encodings-to-enable-tree
Repo https://github.com/microsoft/icecaps
Framework tf

Sample Adaptive MCMC

Title Sample Adaptive MCMC
Authors Michael Zhu
Abstract For MCMC methods like Metropolis-Hastings, tuning the proposal distribution is important in practice for effective sampling from the target distribution \pi. In this paper, we present Sample Adaptive MCMC (SA-MCMC), a MCMC method based on a reversible Markov chain for \pi^{\otimes N} that uses an adaptive proposal distribution based on the current state of N points and a sequential substitution procedure with one new likelihood evaluation per iteration and at most one updated point each iteration. The SA-MCMC proposal distribution automatically adapts within its parametric family to best approximate the target distribution, so in contrast to many existing MCMC methods, SA-MCMC does not require any tuning of the proposal distribution. Instead, SA-MCMC only requires specifying the initial state of N points, which can often be chosen a priori, thereby automating the entire sampling procedure with no tuning required. Experimental results demonstrate the fast adaptation and effective sampling of SA-MCMC.
Tasks
Published 2019-12-01
URL http://papers.nips.cc/paper/9107-sample-adaptive-mcmc
PDF http://papers.nips.cc/paper/9107-sample-adaptive-mcmc.pdf
PWC https://paperswithcode.com/paper/sample-adaptive-mcmc
Repo https://github.com/michaelhzhu/SampleAdaptiveMCMC
Framework none

Gradient Matching Generative Networks for Zero-Shot Learning

Title Gradient Matching Generative Networks for Zero-Shot Learning
Authors Mert Bulent Sariyildiz, Ramazan Gokberk Cinbis
Abstract Zero-shot learning (ZSL) is one of the most promising problems where substantial progress can potentially be achieved through unsupervised learning, due to distributional differences between supervised and zero-shot classes. For this reason, several works investigate the incorporation of discriminative domain adaptation techniques into ZSL, which, however, lead to modest improvements in ZSL accuracy. In contrast, we propose a generative model that can naturally learn from unsupervised examples, and synthesize training examples for unseen classes purely based on their class embeddings, and therefore, reduce the zero-shot learning problem into a supervised classification task. The proposed approach consists of two important components: (i) a conditional Generative Adversarial Network that learns to produce samples that mimic the characteristics of unsupervised data examples, and (ii) the Gradient Matching (GM) loss that measures the quality of the gradient signal obtained from the synthesized examples. Using our GM loss formulation, we enforce the generator to produce examples from which accurate classifiers can be trained. Experimental results on several ZSL benchmark datasets show that our approach leads to significant improvements over the state of the art in generalized zero-shot classification.
Tasks Domain Adaptation, Generalized Zero-Shot Learning - Unseen, Zero-Shot Learning
Published 2019-06-01
URL http://openaccess.thecvf.com/content_CVPR_2019/html/Sariyildiz_Gradient_Matching_Generative_Networks_for_Zero-Shot_Learning_CVPR_2019_paper.html
PDF http://openaccess.thecvf.com/content_CVPR_2019/papers/Sariyildiz_Gradient_Matching_Generative_Networks_for_Zero-Shot_Learning_CVPR_2019_paper.pdf
PWC https://paperswithcode.com/paper/gradient-matching-generative-networks-for
Repo https://github.com/mbsariyildiz/gmn-zsl
Framework pytorch

QTUNA: A Corpus for Understanding How Speakers Use Quantification

Title QTUNA: A Corpus for Understanding How Speakers Use Quantification
Authors Guanyi Chen, Kees van Deemter, Silvia Pagliaro, Louk Smalbil, Chenghua Lin
Abstract A prominent strand of work in formal semantics investigates the ways in which human languages quantify over the elements of a set, as when we say {}\textit{All $A$ are $B$ }{''}, {}\textit{All except two $A$ are $B$ }{''}, {``}\textit{Only a few of the $A$ are $B$ }{''} and so on. Our aim is to build Natural Language Generation algorithms that mimic humans{'} use of quantified expressions. To inform these algorithms, we conducted on a series of elicitation experiments in which human speakers were asked to perform a linguistic task that invites the use of quantified expressions. We discuss how these experiments were conducted and what corpora they gave rise to. We conduct an informal analysis of the corpora, and offer an initial assessment of the challenges that these corpora pose for Natural Language Generation. The dataset is available at:{\textasciitilde}\url{https://github.com/a-quei/qtuna}. |
Tasks Text Generation
Published 2019-10-01
URL https://www.aclweb.org/anthology/W19-8616/
PDF https://www.aclweb.org/anthology/W19-8616
PWC https://paperswithcode.com/paper/qtuna-a-corpus-for-understanding-how-speakers
Repo https://github.com/a-quei/qtuna
Framework none

Fully Dynamic Consistent Facility Location

Title Fully Dynamic Consistent Facility Location
Authors Vincent Cohen-Addad, Niklas Oskar D. Hjuler, Nikos Parotsidis, David Saulpic, Chris Schwiegelshohn
Abstract We consider classic clustering problems in fully dynamic data streams, where data elements can be both inserted and deleted. In this context, several parameters are of importance: (1) the quality of the solution after each insertion or deletion, (2) the time it takes to update the solution, and (3) how different consecutive solutions are. The question of obtaining efficient algorithms in this context for facility location, $k$-median and $k$-means has been raised in a recent paper by Hubert-Chan et al. [WWW’18] and also appears as a natural follow-up on the online model with recourse studied by Lattanzi and Vassilvitskii [ICML’17] (i.e.: in insertion-only streams). In this paper, we focus on general metric spaces and mainly on the facility location problem. We give an arguably simple algorithm that maintains a constant factor approximation, with $O(n\log n)$ update time, and total recourse $O(n)$. This improves over the naive algorithm which consists in recomputing a solution at each time step and that can take up to $O(n^2)$ update time, and $O(n^2)$ total recourse. These bounds are nearly optimal: in general metric space, inserting a point take $O(n)$ times to describe the distances to other points, and we give a simple lower bound of $O(n)$ for the recourse. Moreover, we generalize this result for the $k$-medians and $k$-means problems: our algorithm maintains a constant factor approximation in time $\widetilde{O}(n+k^2)$. We complement our analysis with experiments showing that the cost of the solution maintained by our algorithm at any time $t$ is very close to the cost of a solution obtained by quickly recomputing a solution from scratch at time $t$ while having a much better running time.
Tasks
Published 2019-12-01
URL http://papers.nips.cc/paper/8588-fully-dynamic-consistent-facility-location
PDF http://papers.nips.cc/paper/8588-fully-dynamic-consistent-facility-location.pdf
PWC https://paperswithcode.com/paper/fully-dynamic-consistent-facility-location
Repo https://github.com/NikosParotsidis/Fully-dynamic_facility_location-NeurIPS2019
Framework none

Detecting Bot Behaviour in Social Media using Digital DNA Compression

Title Detecting Bot Behaviour in Social Media using Digital DNA Compression
Authors Nivranshu Pasricha, Conor Hayes
Abstract A major challenge faced by online social networks such as Facebook and Twitter is the remarkable rise of fake and automated bot accounts over the last few years. Some of these accounts have been reported to engage in undesirable activities such as spamming, political campaigning and spreading falsehood on the platform. We present an approach to detect bot-like behaviour among Twitter accounts by analyzing their past tweeting activity. We build upon an existing technique of analysis of Twitter accounts called Digital DNA. Digital DNA models the behaviour of Twitter accounts by encoding the post history of a user account as a sequence of characters analogous to an actual DNA sequence. In our approach, we employ a lossless compression algorithm on these Digital DNA sequences and use the compression statistics as a measure of predictability in the behaviour of a group of Twitter accounts. We leverage the information conveyed by the compression statistics to visually represent the posting behaviour by a simple two dimensional scatter plot and categorize the user accounts as bots and genuine users by using an off-the-shelf implementation of the logistic regression classification algorithm.
Tasks Twitter Bot Detection
Published 2019-12-05
URL https://aran.library.nuigalway.ie/handle/10379/15683
PDF http://aics2019.datascienceinstitute.ie/papers/aics_35.pdf
PWC https://paperswithcode.com/paper/detecting-bot-behaviour-in-social-media-using
Repo https://github.com/pasricha/bot-dna-compression
Framework none

Computational Ad Hominem Detection

Title Computational Ad Hominem Detection
Authors Pieter Delobelle, Murilo Cunha, Eric Massip Cano, Jeroen Peperkamp, Bettina Berendt
Abstract Fallacies like the personal attack{—}also known as the ad hominem attack{—}are introduced in debates as an easy win, even though they provide no rhetorical contribution. Although their importance in argumentation mining is acknowledged, automated mining and analysis is still lacking. We show TF-IDF approaches are insufficient to detect the ad hominem attack. Therefore we present a machine learning approach for information extraction, which has a recall of 80{%} for a social media data source. We also demonstrate our approach with an application that uses online learning.
Tasks
Published 2019-07-01
URL https://www.aclweb.org/anthology/P19-2028/
PDF https://www.aclweb.org/anthology/P19-2028
PWC https://paperswithcode.com/paper/computational-ad-hominem-detection
Repo https://github.com/iPieter/G0B34a_knowledge_and_the_web
Framework none

Face Reconstruction from Voice using Generative Adversarial Networks

Title Face Reconstruction from Voice using Generative Adversarial Networks
Authors Yandong Wen, Bhiksha Raj, Rita Singh
Abstract Voice profiling aims at inferring various human parameters from their speech, e.g. gender, age, etc. In this paper, we address the challenge posed by a subtask of voice profiling - reconstructing someone’s face from their voice. The task is designed to answer the question: given an audio clip spoken by an unseen person, can we picture a face that has as many common elements, or associations as possible with the speaker, in terms of identity? To address this problem, we propose a simple but effective computational framework based on generative adversarial networks (GANs). The network learns to generate faces from voices by matching the identities of generated faces to those of the speakers, on a training set. We evaluate the performance of the network by leveraging a closely related task - cross-modal matching. The results show that our model is able to generate faces that match several biometric characteristics of the speaker, and results in matching accuracies that are much better than chance. The code is publicly available in https://github.com/cmu-mlsp/reconstructing_faces_from_voices
Tasks Face Reconstruction
Published 2019-12-01
URL http://papers.nips.cc/paper/8768-face-reconstruction-from-voice-using-generative-adversarial-networks
PDF http://papers.nips.cc/paper/8768-face-reconstruction-from-voice-using-generative-adversarial-networks.pdf
PWC https://paperswithcode.com/paper/face-reconstruction-from-voice-using
Repo https://github.com/cmu-mlsp/reconstructing_faces_from_voices
Framework pytorch

Cross-lingual CCG Induction

Title Cross-lingual CCG Induction
Authors Kilian Evang
Abstract Combinatory categorial grammars are linguistically motivated and useful for semantic parsing, but costly to acquire in a supervised way and difficult to acquire in an unsupervised way. We propose an alternative making use of cross-lingual learning: an existing source-language parser is used together with a parallel corpus to induce a grammar and parsing model for a target language. On the PASCAL benchmark, cross-lingual CCG induction outperforms CCG induction from gold-standard POS tags on 3 out of 8 languages, and unsupervised CCG induction on 6 out of 8 languages. We also show that cross-lingually induced CCGs reflect syntactic properties of the target languages.
Tasks Semantic Parsing
Published 2019-06-01
URL https://www.aclweb.org/anthology/N19-1160/
PDF https://www.aclweb.org/anthology/N19-1160
PWC https://paperswithcode.com/paper/cross-lingual-ccg-induction
Repo https://github.com/texttheater/xlci
Framework pytorch

RNA secondary structure prediction using an ensemble of two-dimensional deep neural networks and transfer learning

Title RNA secondary structure prediction using an ensemble of two-dimensional deep neural networks and transfer learning
Authors Jaswinder Singh, Jack Hanson, Kuldip Paliwal, Yaoqi Zhou
Abstract The majority of our human genome transcribes into noncoding RNAs with unknown structures and functions. Obtaining functional clues for noncoding RNAs requires accurate base-pairing or secondary-structure prediction. However, the performance of such predictions by current folding-based algorithms has been stagnated for more than a decade. Here, we propose the use of deep contextual learning for base-pair prediction including those noncanonical and non-nested (pseudoknot) base pairs stabilized by tertiary interactions. Since only <250 nonredundant, high-resolution RNA structures are available for model training, we utilize transfer learning from a model initially trained with a recent high-quality bpRNA dataset of >10,000 nonredundant RNAs made available through comparative analysis. The resulting method achieves large, statistically significant improvement in predicting all base pairs, noncanonical and non-nested base pairs in particular. The proposed method (SPOT-RNA), with a freely available server and standalone software, should be useful for improving RNA structure modeling, sequence alignment, and functional annotations.
Tasks Transfer Learning
Published 2019-11-27
URL https://doi.org/10.1038/s41467-019-13395-9
PDF https://doi.org/10.1038/s41467-019-13395-9
PWC https://paperswithcode.com/paper/rna-secondary-structure-prediction-using-an
Repo https://github.com/jaswindersingh2/SPOT-RNA
Framework tf

Sleep quality prediction in caregivers using physiological signals

Title Sleep quality prediction in caregivers using physiological signals
Authors Reza Sadeghi, Tanvi Banerjee, Jennifer C. Hughes, Larry W. Lawhorne
Abstract Most caregivers of people with dementia (CPWD) experience a high degree of stress due to the demands of providing care, especially when addressing unpredictable behavioral and psychological symptoms of dementia. Such challenging responsibilities make caregivers susceptible to poor sleep quality with detrimental effects on their overall health. Hence, monitoring caregivers’ sleep quality can provide important CPWD stress assessment. Most current sleep studies are based on polysomnography, which is expensive and potentially disrupts the caregiving routine. To address these issues, we propose a clinical decision support system to predict sleep quality based on trends of physiological signals in the deep sleep stage. This system utilizes four raw physiological signals using a wearable device (E4 wristband): heart rate variability, electrodermal activity, body movement, and skin temperature. To evaluate the performance of the proposed method, analyses were conducted on a two-week period of sleep monitored on eight CPWD. The best performance is achieved using the random forest classifier with an accuracy of 75% for sleep quality, and 73% for restfulness, respectively. We found that the most important features to detect these measures are sleep efficiency (ratio of amount of time asleep to the amount of time in bed) and skin temperature. The results from our sleep analysis system demonstrate the capability of using wearable sensors to measure sleep quality and restfulness in CPWD.
Tasks Heart Rate Variability, Sleep Quality Prediction
Published 2019-05-20
URL https://www.sciencedirect.com/science/article/pii/S001048251930160X
PDF https://www.sciencedirect.com/science/article/pii/S001048251930160X
PWC https://paperswithcode.com/paper/sleep-quality-prediction-in-caregivers-using
Repo https://github.com/RezaSadeghiWSU/Sleep-quality-in-caregivers
Framework none

Integral Object Mining via Online Attention Accumulation

Title Integral Object Mining via Online Attention Accumulation
Authors Peng-Tao Jiang, Qibin Hou, Yang Cao, Ming-Ming Cheng, Yunchao Wei, Hong-Kai Xiong
Abstract Object attention maps generated by image classifiers are usually used as priors for weakly-supervised segmentation approaches. However, normal image classifiers produce attention only at the most discriminative object parts, which limits the performance of weakly-supervised segmentation task. Therefore, how to effectively identify entire object regions in a weakly-supervised manner has always been a challenging and meaningful problem. We observe that the attention maps produced by a classification network continuously focus on different object parts during training. In order to accumulate the discovered different object parts, we propose an online attention accumulation (OAA) strategy which maintains a cumulative attention map for each target category in each training image so that the integral object regions can be gradually promoted as the training goes. These cumulative attention maps, in turn, serve as the pixel-level supervision, which can further assist the network in discovering more integral object regions. Our method (OAA) can be plugged into any classification network and progressively accumulate the discriminative regions into integral objects as the training process goes. Despite its simplicity, when applying the resulting attention maps to the weakly-supervised semantic segmentation task, our approach improves the existing state-of-the-art methods on the PASCAL VOC 2012 segmentation benchmark, achieving a mIoU score of 66.4% on the test set. Code is available at https://mmcheng.net/oaa/.
Tasks Semantic Segmentation, Weakly-Supervised Semantic Segmentation
Published 2019-10-01
URL http://openaccess.thecvf.com/content_ICCV_2019/html/Jiang_Integral_Object_Mining_via_Online_Attention_Accumulation_ICCV_2019_paper.html
PDF http://openaccess.thecvf.com/content_ICCV_2019/papers/Jiang_Integral_Object_Mining_via_Online_Attention_Accumulation_ICCV_2019_paper.pdf
PWC https://paperswithcode.com/paper/integral-object-mining-via-online-attention
Repo https://github.com/PengtaoJiang/OAA-PyTorch
Framework pytorch
comments powered by Disqus