January 31, 2020

3153 words 15 mins read

Paper Group ANR 120

Declarative Data Analytics: a Survey. Supported-BinaryNet: Bitcell Array-based Weight Supports for Dynamic Accuracy-Latency Trade-offs in SRAM-based Binarized Neural Network. Neural Design Network: Graphic Layout Generation with Constraints. Bottleneck potentials in Markov Random Fields. Learning to Segment Brain Anatomy from 2D Ultrasound with Les …

Declarative Data Analytics: a Survey


Title	Declarative Data Analytics: a Survey
Authors	Nantia Makrynioti, Vasilis Vassalos
Abstract	The area of declarative data analytics explores the application of the declarative paradigm on data science and machine learning. It proposes declarative languages for expressing data analysis tasks and develops systems which optimize programs written in those languages. The execution engine can be either centralized or distributed, as the declarative paradigm advocates independence from particular physical implementations. The survey explores a wide range of declarative data analysis frameworks by examining both the programming model and the optimization techniques used, in order to provide conclusions on the current state of the art in the area and identify open challenges.
Tasks
Published	2019-02-04
URL	http://arxiv.org/abs/1902.01304v1
PDF	http://arxiv.org/pdf/1902.01304v1.pdf
PWC	https://paperswithcode.com/paper/declarative-data-analytics-a-survey
Repo
Framework

Supported-BinaryNet: Bitcell Array-based Weight Supports for Dynamic Accuracy-Latency Trade-offs in SRAM-based Binarized Neural Network


Title	Supported-BinaryNet: Bitcell Array-based Weight Supports for Dynamic Accuracy-Latency Trade-offs in SRAM-based Binarized Neural Network
Authors	Shamma Nasrin, Srikanth Ramakrishna, Theja Tulabandhula, Amit Ranjan Trivedi
Abstract	In this work, we introduce bitcell array-based support parameters to improve the prediction accuracy of SRAM-based binarized neural network (SRAM-BNN). Our approach enhances the training weight space of SRAM-BNN while requiring minimal overheads to a typical design. More flexibility of the weight space leads to higher prediction accuracy in our design. We adapt row digital-to-analog (DAC) converter, and computing flow in SRAM-BNN for bitcell array-based weight supports. Using the discussed interventions, our scheme also allows a dynamic trade-off of accuracy against latency to address dynamic latency constraints in typical real-time applications. We specifically discuss results on two training cases: (i) learning of support parameters on a pre-trained BNN and (ii) simultaneous learning of supports and weight binarization. In the former case, our approach reduces classification error in MNIST by 35.71% (error rate decreases from 1.4% to 0.91%). In the latter case, the error is reduced by 27.65% (error rate decreases from 1.4% to 1.13%). To reduce the power overheads, we propose a dynamic drop out a part of the support parameters. Our architecture can drop out 52% of the bitcell array-based support parameters without losing accuracy. We also characterize our design under varying degrees of process variability in the transistors.
Tasks
Published	2019-11-19
URL	https://arxiv.org/abs/1911.08518v2
PDF	https://arxiv.org/pdf/1911.08518v2.pdf
PWC	https://paperswithcode.com/paper/supported-binarynet-bitcell-array-based
Repo
Framework

Neural Design Network: Graphic Layout Generation with Constraints


Title	Neural Design Network: Graphic Layout Generation with Constraints
Authors	Hsin-Ying Lee, Weilong Yang, Lu Jiang, Madison Le, Irfan Essa, Haifeng Gong, Ming-Hsuan Yang
Abstract	Graphic design is essential for visual communication with layouts being fundamental to composing attractive designs. Layout generation differs from pixel-level image synthesis and is unique in terms of the requirement of mutual relations among the desired components. We propose a method for design layout generation that can satisfy user-specified constraints. The proposed neural design network (NDN) consists of three modules. The first module predicts a graph with complete relations from a graph with user-specified relations. The second module generates a layout from the predicted graph. Finally, the third module fine-tunes the predicted layout. Quantitative and qualitative experiments demonstrate that the generated layouts are visually similar to real design layouts. We also construct real designs based on predicted layouts for a better understanding of the visual quality. Finally, we demonstrate a practical application on layout recommendation.
Tasks	Image Generation
Published	2019-12-19
URL	https://arxiv.org/abs/1912.09421v1
PDF	https://arxiv.org/pdf/1912.09421v1.pdf
PWC	https://paperswithcode.com/paper/neural-design-network-graphic-layout
Repo
Framework

Bottleneck potentials in Markov Random Fields


Title	Bottleneck potentials in Markov Random Fields
Authors	Ahmed Abbas, Paul Swoboda
Abstract	We consider general discrete Markov Random Fields(MRFs) with additional bottleneck potentials which penalize the maximum (instead of the sum) over local potential value taken by the MRF-assignment. Bottleneck potentials or analogous constructions have been considered in (i) combinatorial optimization (e.g. bottleneck shortest path problem, the minimum bottleneck spanning tree problem, bottleneck function minimization in greedoids), (ii) inverse problems with $L_{\infty}$-norm regularization, and (iii) valued constraint satisfaction on the $(\min,\max)$-pre-semirings. Bottleneck potentials for general discrete MRFs are a natural generalization of the above direction of modeling work to Maximum-A-Posteriori (MAP) inference in MRFs. To this end, we propose MRFs whose objective consists of two parts: terms that factorize according to (i) $(\min,+)$, i.e. potentials as in plain MRFs, and (ii) $(\min,\max)$, i.e. bottleneck potentials. To solve the ensuing inference problem, we propose high-quality relaxations and efficient algorithms for solving them. We empirically show efficacy of our approach on large scale seismic horizon tracking problems.
Tasks	Combinatorial Optimization
Published	2019-04-17
URL	https://arxiv.org/abs/1904.08080v2
PDF	https://arxiv.org/pdf/1904.08080v2.pdf
PWC	https://paperswithcode.com/paper/bottleneck-potentials-in-markov-random-fields
Repo
Framework

Learning to Segment Brain Anatomy from 2D Ultrasound with Less Data


Title	Learning to Segment Brain Anatomy from 2D Ultrasound with Less Data
Authors	Jeya Maria Jose V., Rajeev Yasarla, Puyang Wang, Ilker Hacihaliloglu, Vishal M. Patel
Abstract	Automatic segmentation of anatomical landmarks from ultrasound (US) plays an important role in the management of preterm neonates with a very low birth weight due to the increased risk of developing intraventricular hemorrhage (IVH) or other complications. One major problem in developing an automatic segmentation method for this task is the limited availability of annotated data. To tackle this issue, we propose a novel image synthesis method using multi-scale self attention generator to synthesize US images from various segmentation masks. We show that our method can synthesize high-quality US images for every manipulated segmentation label with qualitative and quantitative improvements over the recent state-of-the-art synthesis methods. Furthermore, for the segmentation task, we propose a novel method, called Confidence-guided Brain Anatomy Segmentation (CBAS) network, where segmentation and corresponding confidence maps are estimated at different scales. In addition, we introduce a technique which guides CBAS to learn the weights based on the confidence measure about the estimate. Extensive experiments demonstrate that the proposed method for both synthesis and segmentation tasks achieve significant improvements over the recent state-of-the-art methods. In particular, we show that the new synthesis framework can be used to generate realistic US images which can be used to improve the performance of a segmentation algorithm.
Tasks	Image Generation
Published	2019-12-18
URL	https://arxiv.org/abs/1912.08364v1
PDF	https://arxiv.org/pdf/1912.08364v1.pdf
PWC	https://paperswithcode.com/paper/learning-to-segment-brain-anatomy-from-2d
Repo
Framework

Bootstrapping deep music separation from primitive auditory grouping principles


Title	Bootstrapping deep music separation from primitive auditory grouping principles
Authors	Prem Seetharaman, Gordon Wichern, Jonathan Le Roux, Bryan Pardo
Abstract	Separating an audio scene such as a cocktail party into constituent, meaningful components is a core task in computer audition. Deep networks are the state-of-the-art approach. They are trained on synthetic mixtures of audio made from isolated sound source recordings so that ground truth for the separation is known. However, the vast majority of available audio is not isolated. The brain uses primitive cues that are independent of the characteristics of any particular sound source to perform an initial segmentation of the audio scene. We present a method for bootstrapping a deep model for music source separation without ground truth by using multiple primitive cues. We apply our method to train a network on a large set of unlabeled music recordings from YouTube to separate vocals from accompaniment without the need for ground truth isolated sources or artificial training mixtures.
Tasks	Music Source Separation
Published	2019-10-23
URL	https://arxiv.org/abs/1910.11133v1
PDF	https://arxiv.org/pdf/1910.11133v1.pdf
PWC	https://paperswithcode.com/paper/bootstrapping-deep-music-separation-from
Repo
Framework

Learning to Groove with Inverse Sequence Transformations


Title	Learning to Groove with Inverse Sequence Transformations
Authors	Jon Gillick, Adam Roberts, Jesse Engel, Douglas Eck, David Bamman
Abstract	We explore models for translating abstract musical ideas (scores, rhythms) into expressive performances using Seq2Seq and recurrent Variational Information Bottleneck (VIB) models. Though Seq2Seq models usually require painstakingly aligned corpora, we show that it is possible to adapt an approach from the Generative Adversarial Network (GAN) literature (e.g. Pix2Pix (Isola et al., 2017) and Vid2Vid (Wang et al. 2018a)) to sequences, creating large volumes of paired data by performing simple transformations and training generative models to plausibly invert these transformations. Music, and drumming in particular, provides a strong test case for this approach because many common transformations (quantization, removing voices) have clear semantics, and models for learning to invert them have real-world applications. Focusing on the case of drum set players, we create and release a new dataset for this purpose, containing over 13 hours of recordings by professional drummers aligned with fine-grained timing and dynamics information. We also explore some of the creative potential of these models, including demonstrating improvements on state-of-the-art methods for Humanization (instantiating a performance from a musical score).
Tasks	Quantization
Published	2019-05-14
URL	https://arxiv.org/abs/1905.06118v2
PDF	https://arxiv.org/pdf/1905.06118v2.pdf
PWC	https://paperswithcode.com/paper/learning-to-groove-with-inverse-sequence
Repo
Framework

Application of Machine Learning in Forecasting International Trade Trends


Title	Application of Machine Learning in Forecasting International Trade Trends
Authors	Feras Batarseh, Munisamy Gopinath, Ganesh Nalluru, Jayson Beckman
Abstract	International trade policies have recently garnered attention for limiting cross-border exchange of essential goods (e.g. steel, aluminum, soybeans, and beef). Since trade critically affects employment and wages, predicting future patterns of trade is a high-priority for policy makers around the world. While traditional economic models aim to be reliable predictors, we consider the possibility that Machine Learning (ML) techniques allow for better predictions to inform policy decisions. Open-government data provide the fuel to power the algorithms that can explain and forecast trade flows to inform policies. Data collected in this article describe international trade transactions and commonly associated economic factors. Machine learning (ML) models deployed include: ARIMA, GBoosting, XGBoosting, and LightGBM for predicting future trade patterns, and K-Means clustering of countries according to economic factors. Unlike short-term and subjective (straight-line) projections and medium-term (aggre-gated) projections, ML methods provide a range of data-driven and interpretable projections for individual commodities. Models, their results, and policies are introduced and evaluated for prediction quality.
Tasks
Published	2019-10-07
URL	https://arxiv.org/abs/1910.03112v1
PDF	https://arxiv.org/pdf/1910.03112v1.pdf
PWC	https://paperswithcode.com/paper/application-of-machine-learning-in
Repo
Framework

Learning to Adaptively Scale Recurrent Neural Networks


Title	Learning to Adaptively Scale Recurrent Neural Networks
Authors	Hao Hu, Liqiang Wang, Guo-Jun Qi
Abstract	Recent advancements in recurrent neural network (RNN) research have demonstrated the superiority of utilizing multiscale structures in learning temporal representations of time series. Currently, most of multiscale RNNs use fixed scales, which do not comply with the nature of dynamical temporal patterns among sequences. In this paper, we propose Adaptively Scaled Recurrent Neural Networks (ASRNN), a simple but efficient way to handle this problem. Instead of using predefined scales, ASRNNs are able to learn and adjust scales based on different temporal contexts, making them more flexible in modeling multiscale patterns. Compared with other multiscale RNNs, ASRNNs are bestowed upon dynamical scaling capabilities with much simpler structures, and are easy to be integrated with various RNN cells. The experiments on multiple sequence modeling tasks indicate ASRNNs can efficiently adapt scales based on different sequence contexts and yield better performances than baselines without dynamical scaling abilities.
Tasks	Time Series
Published	2019-02-15
URL	http://arxiv.org/abs/1902.05696v1
PDF	http://arxiv.org/pdf/1902.05696v1.pdf
PWC	https://paperswithcode.com/paper/learning-to-adaptively-scale-recurrent-neural
Repo
Framework

Efficient Decision Making and Belief Space Planning using Sparse Approximations


Title	Efficient Decision Making and Belief Space Planning using Sparse Approximations
Authors	Khen Elimelech, Vadim Indelman
Abstract	In this work, we introduce a new approach for the efficient solution of autonomous decision and planning problems, with a special focus on decision making under uncertainty and belief space planning (BSP) in high-dimensional state spaces. Usually, to solve the decision problem, we identify the optimal action, according to some objective function. Instead, we claim that we can sometimes generate and solve an analogous yet simplified decision problem, which can be solved more efficiently. Furthermore, a wise simplification method can lead to the same action selection, or one for which the maximal loss can be guaranteed. This simplification is separated from the state inference, and does not compromise its accuracy, as the selected action would finally be applied on the original state. At first, we develop the concept for general decision problems, and provide a theoretical framework of definitions to allow a coherent discussion. We then practically apply these ideas to BSP problems, in which the problem is simplified by considering a sparse approximation of the initial belief. The scalable sparsification algorithm we provide is able to yield solutions which are guaranteed to be consistent with the original problem. We demonstrate the benefits of the approach in the solution of a highly realistic active-SLAM problem, and manage to significantly reduce computation time, with practically no loss in the quality of solution. This rigorous and fundamental work is conceptually novel, and holds numerous possible extensions.
Tasks	Decision Making, Decision Making Under Uncertainty
Published	2019-09-02
URL	https://arxiv.org/abs/1909.00885v2
PDF	https://arxiv.org/pdf/1909.00885v2.pdf
PWC	https://paperswithcode.com/paper/efficient-decision-making-and-belief-space
Repo
Framework

Accurate Localization in Wireless Sensor Networks in the Presence of Cross Technology Interference


Title	Accurate Localization in Wireless Sensor Networks in the Presence of Cross Technology Interference
Authors	Usman Nazir
Abstract	Localization of mobile nodes in a wireless sensor networks (WSNs) is an active area of research. In this paper, we present a novel RSSI based localization algorithm for 802.15.4 (ZigBee) based WSNs. We propose and implement a novel range based localization algorithm to minimize cross technology interference operating in the same band. The goal is to minimize the mean square error of the localization algorithm. Hardware implementation of the algorithm is in agreement with ideal (no interference) simulation results where an accuracy of less than 0.5m has been achieved.
Tasks
Published	2019-05-25
URL	https://arxiv.org/abs/1906.00753v1
PDF	https://arxiv.org/pdf/1906.00753v1.pdf
PWC	https://paperswithcode.com/paper/190600753
Repo
Framework

Fair Multi-party Machine Learning – a Game Theoretic approach


Title	Fair Multi-party Machine Learning – a Game Theoretic approach
Authors	Zhiliang Chen
Abstract	High performance machine learning models have become highly dependent on the availability of large quantity and quality of training data. To achieve this, various central agencies such as the government have suggested for different data providers to pool their data together to learn a unified predictive model, which performs better. However, these providers are usually profit-driven and would only agree to participate inthe data sharing process if the process is deemed both profitable and fair for themselves. Due to the lack of existing literature, it is unclear whether a fair and stable outcome is possible in such data sharing processes. Hence, we wish to investigate the outcomes surrounding these scenarios and study if data providers would even agree to collaborate in the first place. Tapping on cooperative game concepts in Game Theory, we introduce the data sharing process between a group of agents as a new class of cooperative games with modified definition of stability and fairness. Using these new definitions, we then theoretically study the optimal and suboptimal outcomes of such data sharing processes and their sensitivity to perturbation.Through experiments, we present intuitive insights regarding theoretical results analysed in this paper and discuss various ways in which data can be valued reasonably.
Tasks
Published	2019-11-22
URL	https://arxiv.org/abs/1911.11555v1
PDF	https://arxiv.org/pdf/1911.11555v1.pdf
PWC	https://paperswithcode.com/paper/fair-multi-party-machine-learning-a-game
Repo
Framework

STEP: Spatial Temporal Graph Convolutional Networks for Emotion Perception from Gaits


Title	STEP: Spatial Temporal Graph Convolutional Networks for Emotion Perception from Gaits
Authors	Uttaran Bhattacharya, Trisha Mittal, Rohan Chandra, Tanmay Randhavane, Aniket Bera, Dinesh Manocha
Abstract	We present a novel classifier network called STEP, to classify perceived human emotion from gaits, based on a Spatial Temporal Graph Convolutional Network (ST-GCN) architecture. Given an RGB video of an individual walking, our formulation implicitly exploits the gait features to classify the emotional state of the human into one of four emotions: happy, sad, angry, or neutral. We use hundreds of annotated real-world gait videos and augment them with thousands of annotated synthetic gaits generated using a novel generative network called STEP-Gen, built on an ST-GCN based Conditional Variational Autoencoder (CVAE). We incorporate a novel push-pull regularization loss in the CVAE formulation of STEP-Gen to generate realistic gaits and improve the classification accuracy of STEP. We also release a novel dataset (E-Gait), which consists of $2,177$ human gaits annotated with perceived emotions along with thousands of synthetic gaits. In practice, STEP can learn the affective features and exhibits classification accuracy of 89% on E-Gait, which is 14 - 30% more accurate over prior methods.
Tasks
Published	2019-10-28
URL	https://arxiv.org/abs/1910.12906v1
PDF	https://arxiv.org/pdf/1910.12906v1.pdf
PWC	https://paperswithcode.com/paper/step-spatial-temporal-graph-convolutional
Repo
Framework

Differentially Private Distributed Data Summarization under Covariate Shift


Title	Differentially Private Distributed Data Summarization under Covariate Shift
Authors	Kanthi Sarpatwar, Karthikeyan Shanmugam, Venkata Sitaramagiridharganesh Ganapavarapu, Ashish Jagmohan, Roman Vaculin
Abstract	We envision AI marketplaces to be platforms where consumers, with very less data for a target task, can obtain a relevant model by accessing many private data sources with vast number of data samples. One of the key challenges is to construct a training dataset that matches a target task without compromising on privacy of the data sources. To this end, we consider the following distributed data summarizataion problem. Given K private source datasets denoted by $[D_i]{i\in [K]}$ and a small target validation set $D_v$, which may involve a considerable covariate shift with respect to the sources, compute a summary dataset $D_s\subseteq \bigcup{i\in [K]} D_i$ such that its statistical distance from the validation dataset $D_v$ is minimized. We use the popular Maximum Mean Discrepancy as the measure of statistical distance. The non-private problem has received considerable attention in prior art, for example in prototype selection (Kim et al., NIPS 2016). Our work is the first to obtain strong differential privacy guarantees while ensuring the quality guarantees of the non-private version. We study this problem in a Parsimonious Curator Privacy Model, where a trusted curator coordinates the summarization process while minimizing the amount of private information accessed. Our central result is a novel protocol that (a) ensures the curator accesses at most $O(K^{\frac{1}{3}}D_s + D_v)$ points (b) has formal privacy guarantees on the leakage of information between the data owners and (c) closely matches the best known non-private greedy algorithm. Our protocol uses two hash functions, one inspired by the Rahimi-Recht random features method and the second leverages state of the art differential privacy mechanisms. We introduce a novel “noiseless” differentially private auctioning protocol for winner notification and demonstrate the efficacy of our protocol using real-world datasets.
Tasks	Data Summarization
Published	2019-10-28
URL	https://arxiv.org/abs/1910.12832v2
PDF	https://arxiv.org/pdf/1910.12832v2.pdf
PWC	https://paperswithcode.com/paper/differentially-private-distributed-data
Repo
Framework

A Tunable Loss Function for Binary Classification


Title	A Tunable Loss Function for Binary Classification
Authors	Tyler Sypherd, Mario Diaz, Lalitha Sankar, Peter Kairouz
Abstract	We present $\alpha$-loss, $\alpha \in [1,\infty]$, a tunable loss function for binary classification that bridges log-loss ($\alpha=1$) and $0$-$1$ loss ($\alpha = \infty$). We prove that $\alpha$-loss has an equivalent margin-based form and is classification-calibrated, two desirable properties for a good surrogate loss function for the ideal yet intractable $0$-$1$ loss. For logistic regression-based classification, we provide an upper bound on the difference between the empirical and expected risk at the empirical risk minimizers for $\alpha$-loss by exploiting its Lipschitzianity along with recent results on the landscape features of empirical risk functions. Finally, we show that $\alpha$-loss with $\alpha = 2$ performs better than log-loss on MNIST for logistic regression.
Tasks
Published	2019-02-12
URL	http://arxiv.org/abs/1902.04639v2
PDF	http://arxiv.org/pdf/1902.04639v2.pdf
PWC	https://paperswithcode.com/paper/a-tunable-loss-function-for-binary
Repo
Framework