Paper Group NAWR 29
MFAS: Multimodal Fusion Architecture Search. Unsupervised Parallel Sentence Extraction with Parallel Segment Detection Helps Machine Translation. A Deep Learning-based Radar and Camera Sensor Fusion Architecture for Object Detection. Object Instance Annotation With Deep Extreme Level Set Evolution. GNE: a deep learning framework for gene network in …
MFAS: Multimodal Fusion Architecture Search
| Title | MFAS: Multimodal Fusion Architecture Search |
| Authors | Juan-Manuel Perez-Rua, Valentin Vielzeuf, Stephane Pateux, Moez Baccouche, Frederic Jurie |
| Abstract | We tackle the problem of finding good architectures for multimodal classification problems. We propose a novel and generic search space that spans a large number of possible fusion architectures. In order to find an optimal architecture for a given dataset in the proposed search space, we leverage an efficient sequential model-based exploration approach that is tailored for the problem. We demonstrate the value of posing multimodal fusion as a neural architecture search problem by extensive experimentation on a toy dataset and two other real multimodal datasets. We discover fusion architectures that exhibit state-of-the-art performance for problems with different domain and dataset size, including the \ntu dataset, the largest multimodal action recognition dataset available. |
| Tasks | Action Recognition In Videos, Neural Architecture Search, Temporal Action Localization |
| Published | 2019-06-01 |
| URL | http://openaccess.thecvf.com/content_CVPR_2019/html/Perez-Rua_MFAS_Multimodal_Fusion_Architecture_Search_CVPR_2019_paper.html |
| http://openaccess.thecvf.com/content_CVPR_2019/papers/Perez-Rua_MFAS_Multimodal_Fusion_Architecture_Search_CVPR_2019_paper.pdf | |
| PWC | https://paperswithcode.com/paper/mfas-multimodal-fusion-architecture-search-1 |
| Repo | https://github.com/juanmanpr/mfas |
| Framework | pytorch |
Unsupervised Parallel Sentence Extraction with Parallel Segment Detection Helps Machine Translation
| Title | Unsupervised Parallel Sentence Extraction with Parallel Segment Detection Helps Machine Translation |
| Authors | Viktor Hangya, Alex Fraser, er |
| Abstract | Mining parallel sentences from comparable corpora is important. Most previous work relies on supervised systems, which are trained on parallel data, thus their applicability is problematic in low-resource scenarios. Recent developments in building unsupervised bilingual word embeddings made it possible to mine parallel sentences based on cosine similarities of source and target language words. We show that relying only on this information is not enough, since sentences often have similar words but different meanings. We detect continuous parallel segments in sentence pair candidates and rely on them when mining parallel sentences. We show better mining accuracy on three language pairs in a standard shared task on artificial data. We also provide the first experiments showing that parallel sentences mined from real life sources improve unsupervised MT. Our code is available, we hope it will be used to support low-resource MT research. |
| Tasks | Machine Translation, Word Embeddings |
| Published | 2019-07-01 |
| URL | https://www.aclweb.org/anthology/P19-1118/ |
| https://www.aclweb.org/anthology/P19-1118 | |
| PWC | https://paperswithcode.com/paper/unsupervised-parallel-sentence-extraction |
| Repo | https://github.com/hangyav/UnsupPSE |
| Framework | pytorch |
A Deep Learning-based Radar and Camera Sensor Fusion Architecture for Object Detection
| Title | A Deep Learning-based Radar and Camera Sensor Fusion Architecture for Object Detection |
| Authors | Felix Nobis, Maximilian Geisslinger, Markus Weber, Johannes Betz, Markus Lienkamp |
| Abstract | Object detection in camera images, using deep learning has been proven successfully in recent years. Rising detection rates and computationally efficient network structures are pushing this technique towards application in production vehicles. Nevertheless, the sensor quality of the camera is limited in severe weather conditions and through increased sensor noise in sparsely lit areas and at night. Our approach enhances current 2D object detection networks by fusing camera data and projected sparse radar data in the network layers. The proposed CameraRadarFusion Net (CRF-Net) automatically learns at which level the fusion of the sensor data is most beneficial for the detection result. Additionally, we introduce BlackIn, a training strategy inspired by Dropout, which focuses the learning on a specific sensor type. We show that the fusion network is able to outperform a state-of-the-art image-only network for two different datasets. The code for this research will be made available to the public at: https://github.com/TUMFTM/CameraRadarFusionNet. |
| Tasks | Object Detection, Sensor Fusion |
| Published | 2019-10-15 |
| URL | https://ieeexplore.ieee.org/document/8916629 |
| https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8916629 | |
| PWC | https://paperswithcode.com/paper/a-deep-learning-based-radar-and-camera-sensor |
| Repo | https://github.com/TUMFTM/CameraRadarFusionNet |
| Framework | none |
Object Instance Annotation With Deep Extreme Level Set Evolution
| Title | Object Instance Annotation With Deep Extreme Level Set Evolution |
| Authors | Zian Wang, David Acuna, Huan Ling, Amlan Kar, Sanja Fidler |
| Abstract | In this paper, we tackle the task of interactive object segmentation. We revive the old ideas on level set segmentation which framed object annotation as curve evolution. Carefully designed energy functions ensured that the curve was well aligned with image boundaries, and generally “well behaved”. The Level Set Method can handle objects with complex shapes and topological changes such as merging and splitting, thus able to deal with occluded objects and objects with holes. We propose Deep Extreme Level Set Evolution that combines powerful CNN models with level set optimization in an end-to-end fashion. Our method learns to predict evolution parameters conditioned on the image and evolves the predicted initial contour to produce the final result. We make our model interactive by incorporating user clicks on the extreme boundary points, following DEXTR. We show that our approach significantly outperforms DEXTR on the static Cityscapes dataset and the video segmentation benchmark DAVIS, and performs on par on PASCAL and SBD. |
| Tasks | Semantic Segmentation, Video Semantic Segmentation |
| Published | 2019-06-01 |
| URL | http://openaccess.thecvf.com/content_CVPR_2019/html/Wang_Object_Instance_Annotation_With_Deep_Extreme_Level_Set_Evolution_CVPR_2019_paper.html |
| http://openaccess.thecvf.com/content_CVPR_2019/papers/Wang_Object_Instance_Annotation_With_Deep_Extreme_Level_Set_Evolution_CVPR_2019_paper.pdf | |
| PWC | https://paperswithcode.com/paper/object-instance-annotation-with-deep-extreme |
| Repo | https://github.com/fidler-lab/delse |
| Framework | pytorch |
GNE: a deep learning framework for gene network inference by aggregating biological information
| Title | GNE: a deep learning framework for gene network inference by aggregating biological information |
| Authors | Kishan KC, Rui Li, Feng Cui, Qi Yu, Anne R. Haake |
| Abstract | The topological landscape of gene interaction networks provides a rich source of information for inferring functional patterns of genes or proteins. However, it is still a challenging task to aggregate heterogeneous biological information such as gene expression and gene interactions to achieve more accurate inference for prediction and discovery of new gene interactions. In particular, how to generate a unified vector representation to integrate diverse input data is a key challenge addressed here. We propose a scalable and robust deep learning framework to learn embedded representations to unify known gene interactions and gene expression for gene interaction predictions. These low- dimensional embeddings derive deeper insights into the structure of rapidly accumulating and diverse gene interaction networks and greatly simplify downstream modeling. We compare the predictive power of our deep embeddings to the strong baselines. The results suggest that our deep embeddings achieve significantly more accurate predictions. Moreover, a set of novel gene interaction predictions are validated by up-to-date literature-based database entries. The proposed model demonstrates the importance of integrating heterogeneous information about genes for gene network inference. GNE is freely available under the GNU General Public License and can be downloaded from GitHub (https://github.com/kckishan/GNE). |
| Tasks | Gene Interaction Prediction, Link Prediction |
| Published | 2019-04-05 |
| URL | https://bmcsystbiol.biomedcentral.com/articles/10.1186/s12918-019-0694-y |
| https://bmcsystbiol.biomedcentral.com/track/pdf/10.1186/s12918-019-0694-y | |
| PWC | https://paperswithcode.com/paper/gne-a-deep-learning-framework-for-gene |
| Repo | https://github.com/kckishan/GNE |
| Framework | tf |
Demystifying Black-box Models with Symbolic Metamodels
| Title | Demystifying Black-box Models with Symbolic Metamodels |
| Authors | Ahmed M. Alaa, Mihaela Van Der Schaar |
| Abstract | Understanding the predictions of a machine learning model can be as crucial as the model’s accuracy in many application domains. However, the black-box nature of most highly-accurate (complex) models is a major hindrance to their interpretability. To address this issue, we introduce the symbolic metamodeling framework — a general methodology for interpreting predictions by converting “black-box” models into “white-box” functions that are understandable to human subjects. A symbolic metamodel is a model of a model, i.e., a surrogate model of a trained (machine learning) model expressed through a succinct symbolic expression that comprises familiar mathematical functions and can be subjected to symbolic manipulation. We parameterize symbolic metamodels using Meijer G-functions — a class of complex-valued contour integrals that depend on scalar parameters, and whose solutions reduce to familiar elementary, algebraic, analytic and closed-form functions for different parameter settings. This parameterization enables efficient optimization of metamodels via gradient descent, and allows discovering the functional forms learned by a machine learning model with minimal a priori assumptions. We show that symbolic metamodeling provides an all-encompassing framework for model interpretation — all common forms of global and local explanations of a model can be analytically derived from its symbolic metamodel. |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/9308-demystifying-black-box-models-with-symbolic-metamodels |
| http://papers.nips.cc/paper/9308-demystifying-black-box-models-with-symbolic-metamodels.pdf | |
| PWC | https://paperswithcode.com/paper/demystifying-black-box-models-with-symbolic |
| Repo | https://github.com/ahmedmalaa/Symbolic-Metamodeling |
| Framework | none |
Multivariate Distributionally Robust Convex Regression under Absolute Error Loss
| Title | Multivariate Distributionally Robust Convex Regression under Absolute Error Loss |
| Authors | Jose Blanchet, Peter W. Glynn, Jun Yan, Zhengqing Zhou |
| Abstract | This paper proposes a novel non-parametric multidimensional convex regression estimator which is designed to be robust to adversarial perturbations in the empirical measure. We minimize over convex functions the maximum (over Wasserstein perturbations of the empirical measure) of the absolute regression errors. The inner maximization is solved in closed form resulting in a regularization penalty involves the norm of the gradient. We show consistency of our estimator and a rate of convergence of order $ \widetilde{O}\left( n^{-1/d}\right) $, matching the bounds of alternative estimators based on square-loss minimization. Contrary to all of the existing results, our convergence rates hold without imposing compactness on the underlying domain and with no a priori bounds on the underlying convex function or its gradient norm. |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/9352-multivariate-distributionally-robust-convex-regression-under-absolute-error-loss |
| http://papers.nips.cc/paper/9352-multivariate-distributionally-robust-convex-regression-under-absolute-error-loss.pdf | |
| PWC | https://paperswithcode.com/paper/multivariate-distributionally-robust-convex |
| Repo | https://github.com/JunYan65/DRCR_NIPS2019_Code |
| Framework | none |
Average Case Column Subset Selection for Entrywise \ell_1-Norm Loss
| Title | Average Case Column Subset Selection for Entrywise \ell_1-Norm Loss |
| Authors | Zhao Song, David Woodruff, Peilin Zhong |
| Abstract | We study the column subset selection problem with respect to the entrywise $\ell_1$-norm loss. It is known that in the worst case, to obtain a good rank-$k$ approximation to a matrix, one needs an arbitrarily large $n^{\Omega(1)}$ number of columns to obtain a $(1+\epsilon)$-approximation to an $n \times n$ matrix. Nevertheless, we show that under certain minimal and realistic distributional settings, it is possible to obtain a $(1+\epsilon)$-approximation with a nearly linear running time and poly$(k/\epsilon)+O(k\log n)$ columns. Namely, we show that if the input matrix $A$ has the form $A = B + E$, where $B$ is an arbitrary rank-$k$ matrix, and $E$ is a matrix with i.i.d. entries drawn from any distribution $\mu$ for which the $(1+\gamma)$-th moment exists, for an arbitrarily small constant $\gamma > 0$, then it is possible to obtain a $(1+\epsilon)$-approximate column subset selection to the entrywise $\ell_1$-norm in nearly linear time. Conversely we show that if the first moment does not exist, then it is not possible to obtain a $(1+\epsilon)$-approximate subset selection algorithm even if one chooses any $n^{o(1)}$ columns. This is the first algorithm of any kind for achieving a $(1+\epsilon)$-approximation for entrywise $\ell_1$-norm loss low rank approximation. |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/9201-average-case-column-subset-selection-for-entrywise-ell_1-norm-loss |
| http://papers.nips.cc/paper/9201-average-case-column-subset-selection-for-entrywise-ell_1-norm-loss.pdf | |
| PWC | https://paperswithcode.com/paper/average-case-column-subset-selection-for |
| Repo | https://github.com/zpl7840/noise_l1_low_rank_approximation |
| Framework | none |
Communication trade-offs for Local-SGD with large step size
| Title | Communication trade-offs for Local-SGD with large step size |
| Authors | Aymeric Dieuleveut, Kumar Kshitij Patel |
| Abstract | Synchronous mini-batch SGD is state-of-the-art for large-scale distributed machine learning. However, in practice, its convergence is bottlenecked by slow communication rounds between worker nodes. A natural solution to reduce communication is to use the \emph{``local-SGD’'} model in which the workers train their model independently and synchronize every once in a while. This algorithm improves the computation-communication trade-off but its convergence is not understood very well. We propose a non-asymptotic error analysis, which enables comparison to \emph{one-shot averaging} i.e., a single communication round among independent workers, and \emph{mini-batch averaging} i.e., communicating at every step. We also provide adaptive lower bounds on the communication frequency for large step-sizes ($ t^{-\alpha} $, $ \alpha\in (1/2 , 1 ) $) and show that \emph{Local-SGD} reduces communication by a factor of $O\Big(\frac{\sqrt{T}}{P^{3/2}}\Big)$, with $T$ the total number of gradients and $P$ machines. | |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/9512-communication-trade-offs-for-local-sgd-with-large-step-size |
| http://papers.nips.cc/paper/9512-communication-trade-offs-for-local-sgd-with-large-step-size.pdf | |
| PWC | https://paperswithcode.com/paper/communication-trade-offs-for-local-sgd-with |
| Repo | https://github.com/kishinmh/Local-SGD |
| Framework | none |
Personalizing Many Decisions with High-Dimensional Covariates
| Title | Personalizing Many Decisions with High-Dimensional Covariates |
| Authors | Nima Hamidi, Mohsen Bayati, Kapil Gupta |
| Abstract | We consider the k-armed stochastic contextual bandit problem with d dimensional features, when both k and d can be large. To the best of our knowledge, all existing algorithm for this problem have a regret bound that scale as polynomials of degree at least two in k and d. The main contribution of this paper is to introduce and theoretically analyze a new algorithm (REAL Bandit) with a regret that scales by r^2(k+d) when r is rank of the k by d matrix of unknown parameters. REAL Bandit relies on ideas from low-rank matrix estimation literature and a new row-enhancement subroutine that yields sharper bounds for estimating each row of the parameter matrix that may be of independent interest. |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/9323-personalizing-many-decisions-with-high-dimensional-covariates |
| http://papers.nips.cc/paper/9323-personalizing-many-decisions-with-high-dimensional-covariates.pdf | |
| PWC | https://paperswithcode.com/paper/personalizing-many-decisions-with-high |
| Repo | https://github.com/nimily/real-bandit |
| Framework | none |
Learning Local Search Heuristics for Boolean Satisfiability
| Title | Learning Local Search Heuristics for Boolean Satisfiability |
| Authors | Emre Yolcu, Barnabas Poczos |
| Abstract | We present an approach to learn SAT solver heuristics from scratch through deep reinforcement learning with a curriculum. In particular, we incorporate a graph neural network in a stochastic local search algorithm to act as the variable selection heuristic. We consider Boolean satisfiability problems from different classes and learn specialized heuristics for each class. Although we do not aim to compete with the state-of-the-art SAT solvers in run time, we demonstrate that the learned heuristics allow us to find satisfying assignments in fewer steps compared to a generic heuristic, and we provide analysis of our results through experiments. |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/9012-learning-local-search-heuristics-for-boolean-satisfiability |
| http://papers.nips.cc/paper/9012-learning-local-search-heuristics-for-boolean-satisfiability.pdf | |
| PWC | https://paperswithcode.com/paper/learning-local-search-heuristics-for-boolean |
| Repo | https://github.com/emreyolcu/sat |
| Framework | pytorch |
Deep learning for minimum mean-square error approaches to speech enhancement
| Title | Deep learning for minimum mean-square error approaches to speech enhancement |
| Authors | Aaron Nicolson, Kuldip K.Paliwal |
| Abstract | Recently, the focus of speech enhancement research has shifted from minimum mean-square error (MMSE) approaches, like the MMSE short-time spectral amplitude (MMSE-STSA) estimator, to state-of-the-art masking- and mapping-based deep learning approaches. We aim to bridge the gap between these two differing speech enhancement approaches. Deep learning methods for MMSE approaches are investigated in this work, with the objective of producing intelligible enhanced speech at a high quality. Since the speech enhancement performance of an MMSE approach improves with the accuracy of the used a priori signal-to-noise ratio (SNR) estimator, a residual long short-term memory (ResLSTM) network is utilised here to accurately estimate the a priori SNR. MMSE approaches utilising the ResLSTM a priori SNR estimator are evaluated using subjective and objective measures of speech quality and intelligibility. The tested conditions include real-world non-stationary and coloured noise sources at multiple SNR levels. MMSE approaches utilising the proposed a priori SNR estimator are able to achieve higher enhanced speech quality and intelligibility scores than recent masking- and mapping-based deep learning approaches. The results presented in this work show that the performance of an MMSE approach to speech enhancement significantly increases when utilising deep learning. Availability: The proposed a priori SNR estimator is available at: https://github.com/anicolson/DeepXi. |
| Tasks | Speech Enhancement |
| Published | 2019-08-01 |
| URL | https://www.sciencedirect.com/science/article/pii/S0167639318304308 |
| https://www.sciencedirect.com/science/article/pii/S0167639318304308/pdfft?md5=12787b8406c4af704b416a0dbd51e40c&pid=1-s2.0-S0167639318304308-main.pdf | |
| PWC | https://paperswithcode.com/paper/deep-learning-for-minimum-mean-square-error |
| Repo | https://github.com/anicolson/DeepXi |
| Framework | tf |
PTB Graph Parsing with Tree Approximation
| Title | PTB Graph Parsing with Tree Approximation |
| Authors | Yoshihide Kato, Shigeki Matsubara |
| Abstract | The Penn Treebank (PTB) represents syntactic structures as graphs due to nonlocal dependencies. This paper proposes a method that approximates PTB graph-structured representations by trees. By our approximation method, we can reduce nonlocal dependency identification and constituency parsing into single tree-based parsing. An experimental result demonstrates that our approximation method with an off-the-shelf tree-based constituency parser significantly outperforms the previous methods in nonlocal dependency identification. |
| Tasks | Constituency Parsing |
| Published | 2019-07-01 |
| URL | https://www.aclweb.org/anthology/P19-1530/ |
| https://www.aclweb.org/anthology/P19-1530 | |
| PWC | https://paperswithcode.com/paper/ptb-graph-parsing-with-tree-approximation |
| Repo | https://github.com/yosihide/ptb2cf |
| Framework | none |
Consistency-based Semi-supervised Learning for Object detection
| Title | Consistency-based Semi-supervised Learning for Object detection |
| Authors | Jisoo Jeong, Seungeui Lee, Jeesoo Kim, Nojun Kwak |
| Abstract | Making a precise annotation in a large dataset is crucial to the performance of object detection. While the object detection task requires a huge number of annotated samples to guarantee its performance, placing bounding boxes for every object in each sample is time-consuming and costs a lot. To alleviate this problem, we propose a Consistency-based Semi-supervised learning method for object Detection (CSD), which is a way of using consistency constraints as a tool for enhancing detection performance by making full use of available unlabeled data. Specifically, the consistency constraint is applied not only for object classification but also for the localization. We also proposed Background Elimination (BE) to avoid the negative effect of the predominant backgrounds on the detection performance. We have evaluated the proposed CSD both in single-stage and two-stage detectors and the results show the effectiveness of our method. |
| Tasks | Object Classification, Object Detection |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/9259-consistency-based-semi-supervised-learning-for-object-detection |
| http://papers.nips.cc/paper/9259-consistency-based-semi-supervised-learning-for-object-detection.pdf | |
| PWC | https://paperswithcode.com/paper/consistency-based-semi-supervised-learning |
| Repo | https://github.com/soo89/CSD-RFCN |
| Framework | pytorch |
An Accelerated Decentralized Stochastic Proximal Algorithm for Finite Sums
| Title | An Accelerated Decentralized Stochastic Proximal Algorithm for Finite Sums |
| Authors | Hadrien Hendrikx, Francis Bach, Laurent Massoulié |
| Abstract | Modern large-scale finite-sum optimization relies on two key aspects: distribution and stochastic updates. For smooth and strongly convex problems, existing decentralized algorithms are slower than modern accelerated variance-reduced stochastic algorithms when run on a single machine, and are therefore not efficient. Centralized algorithms are fast, but their scaling is limited by global aggregation steps that result in communication bottlenecks. In this work, we propose an efficient \textbf{A}ccelerated \textbf{D}ecentralized stochastic algorithm for \textbf{F}inite \textbf{S}ums named ADFS, which uses local stochastic proximal updates and randomized pairwise communications between nodes. On $n$ machines, ADFS learns from $nm$ samples in the same time it takes optimal algorithms to learn from $m$ samples on one machine. This scaling holds until a critical network size is reached, which depends on communication delays, on the number of samples $m$, and on the network topology. We provide a theoretical analysis based on a novel augmented graph approach combined with a precise evaluation of synchronization times and an extension of the accelerated proximal coordinate gradient algorithm to arbitrary sampling. We illustrate the improvement of ADFS over state-of-the-art decentralized approaches with experiments. |
| Tasks | |
| Published | 2019-12-01 |
| URL | http://papers.nips.cc/paper/8381-an-accelerated-decentralized-stochastic-proximal-algorithm-for-finite-sums |
| http://papers.nips.cc/paper/8381-an-accelerated-decentralized-stochastic-proximal-algorithm-for-finite-sums.pdf | |
| PWC | https://paperswithcode.com/paper/an-accelerated-decentralized-stochastic |
| Repo | https://github.com/HadrienHx/ADFS_NeurIPS |
| Framework | none |
