February 1, 2020

3359 words 16 mins read

Paper Group AWR 157

Context Attentive Document Ranking and Query Suggestion. Hyperpixel Flow: Semantic Correspondence with Multi-layer Neural Features. Remote Sensor Design for Visual Recognition with Convolutional Neural Networks. LeagueAI: Improving object detector performance and flexibility through automatically generated training data and domain randomization. St …

Context Attentive Document Ranking and Query Suggestion


Title	Context Attentive Document Ranking and Query Suggestion
Authors	Wasi Uddin Ahmad, Kai-Wei Chang, Hongning Wang
Abstract	We present a context-aware neural ranking model to exploit users’ on-task search activities and enhance retrieval performance. In particular, a two-level hierarchical recurrent neural network is introduced to learn search context representation of individual queries, search tasks, and corresponding dependency structure by jointly optimizing two companion retrieval tasks: document ranking and query suggestion. To identify the variable dependency structure between search context and users’ ongoing search activities, attention at both levels of recurrent states are introduced. Extensive experiment comparisons against a rich set of baseline methods and an in-depth ablation analysis confirm the value of our proposed approach for modeling search context buried in search tasks.
Tasks	Document Ranking
Published	2019-06-05
URL	https://arxiv.org/abs/1906.02329v1
PDF	https://arxiv.org/pdf/1906.02329v1.pdf
PWC	https://paperswithcode.com/paper/context-attentive-document-ranking-and-query
Repo	https://github.com/wasiahmad/mnsrf_ranking_suggestion
Framework	pytorch

Hyperpixel Flow: Semantic Correspondence with Multi-layer Neural Features


Title	Hyperpixel Flow: Semantic Correspondence with Multi-layer Neural Features
Authors	Juhong Min, Jongmin Lee, Jean Ponce, Minsu Cho
Abstract	Establishing visual correspondences under large intra-class variations requires analyzing images at different levels, from features linked to semantics and context to local patterns, while being invariant to instance-specific details. To tackle these challenges, we represent images by “hyperpixels” that leverage a small number of relevant features selected among early to late layers of a convolutional neural network. Taking advantage of the condensed features of hyperpixels, we develop an effective real-time matching algorithm based on Hough geometric voting. The proposed method, hyperpixel flow, sets a new state of the art on three standard benchmarks as well as a new dataset, SPair-71k, which contains a significantly larger number of image pairs than existing datasets, with more accurate and richer annotations for in-depth analysis.
Tasks
Published	2019-08-18
URL	https://arxiv.org/abs/1908.06537v1
PDF	https://arxiv.org/pdf/1908.06537v1.pdf
PWC	https://paperswithcode.com/paper/hyperpixel-flow-semantic-correspondence-with
Repo	https://github.com/juhongm999/HyperpixelFlow
Framework	pytorch

Remote Sensor Design for Visual Recognition with Convolutional Neural Networks


Title	Remote Sensor Design for Visual Recognition with Convolutional Neural Networks
Authors	Lucas Jaffe, Michael Zelinski, Wesam Sakla
Abstract	While deep learning technologies for computer vision have developed rapidly since 2012, modeling of remote sensing systems has remained focused around human vision. In particular, remote sensing systems are usually constructed to optimize sensing cost-quality trade-offs with respect to human image interpretability. While some recent studies have explored remote sensing system design as a function of simple computer vision algorithm performance, there has been little work relating this design to the state-of-the-art in computer vision: deep learning with convolutional neural networks. We develop experimental systems to conduct this analysis, showing results with modern deep learning algorithms and recent overhead image data. Our results are compared to standard image quality measurements based on human visual perception, and we conclude not only that machine and human interpretability differ significantly, but that computer vision performance is largely self-consistent across a range of disparate conditions. This research is presented as a cornerstone for a new generation of sensor design systems which focus on computer algorithm performance instead of human visual perception.
Tasks
Published	2019-06-24
URL	https://arxiv.org/abs/1906.09677v1
PDF	https://arxiv.org/pdf/1906.09677v1.pdf
PWC	https://paperswithcode.com/paper/remote-sensor-design-for-visual-recognition
Repo	https://github.com/LLNL/sepsense
Framework	pytorch

LeagueAI: Improving object detector performance and flexibility through automatically generated training data and domain randomization


Title	LeagueAI: Improving object detector performance and flexibility through automatically generated training data and domain randomization
Authors	Oliver Struckmeier
Abstract	In this technical report I present my method for automatic synthetic dataset generation for object detection and demonstrate it on the video game League of Legends. This report furthermore serves as a handbook on how to automatically generate datasets and as an introduction on the dataset generation part of the LeagueAI framework. The LeagueAI framework is a software framework that provides detailed information about the game League of Legends based on the same input a human player would have, namely vision. The framework allows researchers and enthusiasts to develop their own intelligent agents or to extract detailed information about the state of the game. A big problem of machine vision applications usually is the laborious work of gathering large amounts of hand labeled data. Thus, a crucial part of the vision pipeline of the LeagueAI framework, the dataset generation, is presented in this report. The method involves extracting image raw data from the game’s 3D models and combining them with the game background to create game-like synthetic images and to generate the corresponding labels automatically. In an experiment I compared a model trained on synthetic data to a model trained on hand labeled data and a model trained on a combined dataset. The model trained on the synthetic data showed higher detection precision on more classes and more reliable tracking performance of the player character. The model trained on the combined dataset did not perform better because of the different formats of the older hand labeled dataset and the synthetic data.
Tasks	League of Legends, Object Detection
Published	2019-05-28
URL	https://arxiv.org/abs/1905.13546v1
PDF	https://arxiv.org/pdf/1905.13546v1.pdf
PWC	https://paperswithcode.com/paper/190513546
Repo	https://github.com/Oleffa/LeagueAI
Framework	pytorch

Stochastic In-Face Frank-Wolfe Methods for Non-Convex Optimization and Sparse Neural Network Training


Title	Stochastic In-Face Frank-Wolfe Methods for Non-Convex Optimization and Sparse Neural Network Training
Authors	Paul Grigas, Alfonso Lobos, Nathan Vermeersch
Abstract	The Frank-Wolfe method and its extensions are well-suited for delivering solutions with desirable structural properties, such as sparsity or low-rank structure. We introduce a new variant of the Frank-Wolfe method that combines Frank-Wolfe steps and steepest descent steps, as well as a novel modification of the “Frank-Wolfe gap” to measure convergence in the non-convex case. We further extend this method to incorporate in-face directions for preserving structured solutions as well as block coordinate steps, and we demonstrate computational guarantees in terms of the modified Frank-Wolfe gap for all of these variants. We are particularly motivated by the application of this methodology to the training of neural networks with sparse properties, and we apply our block coordinate method to the problem of $\ell_1$ regularized neural network training. We present the results of several numerical experiments on both artificial and real datasets demonstrating significant improvements of our method in training sparse neural networks.
Tasks
Published	2019-06-09
URL	https://arxiv.org/abs/1906.03580v1
PDF	https://arxiv.org/pdf/1906.03580v1.pdf
PWC	https://paperswithcode.com/paper/stochastic-in-face-frank-wolfe-methods-for
Repo	https://github.com/AlfLobos/Stochastic-In-Face-Frank-Wolfe-v1.0
Framework	pytorch

Stereo R-CNN based 3D Object Detection for Autonomous Driving


Title	Stereo R-CNN based 3D Object Detection for Autonomous Driving
Authors	Peiliang Li, Xiaozhi Chen, Shaojie Shen
Abstract	We propose a 3D object detection method for autonomous driving by fully exploiting the sparse and dense, semantic and geometry information in stereo imagery. Our method, called Stereo R-CNN, extends Faster R-CNN for stereo inputs to simultaneously detect and associate object in left and right images. We add extra branches after stereo Region Proposal Network (RPN) to predict sparse keypoints, viewpoints, and object dimensions, which are combined with 2D left-right boxes to calculate a coarse 3D object bounding box. We then recover the accurate 3D bounding box by a region-based photometric alignment using left and right RoIs. Our method does not require depth input and 3D position supervision, however, outperforms all existing fully supervised image-based methods. Experiments on the challenging KITTI dataset show that our method outperforms the state-of-the-art stereo-based method by around 30% AP on both 3D detection and 3D localization tasks. Code has been released at https://github.com/HKUST-Aerial-Robotics/Stereo-RCNN.
Tasks	3D Object Detection, Autonomous Driving, Object Detection
Published	2019-02-26
URL	http://arxiv.org/abs/1902.09738v2
PDF	http://arxiv.org/pdf/1902.09738v2.pdf
PWC	https://paperswithcode.com/paper/stereo-r-cnn-based-3d-object-detection-for
Repo	https://github.com/srinu6/Stereo-3D-Object-Detection-for-Autonomous-Driving
Framework	pytorch

Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learning


Title	Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learning
Authors	Xue Bin Peng, Aviral Kumar, Grace Zhang, Sergey Levine
Abstract	In this paper, we aim to develop a simple and scalable reinforcement learning algorithm that uses standard supervised learning methods as subroutines. Our goal is an algorithm that utilizes only simple and convergent maximum likelihood loss functions, while also being able to leverage off-policy data. Our proposed approach, which we refer to as advantage-weighted regression (AWR), consists of two standard supervised learning steps: one to regress onto target values for a value function, and another to regress onto weighted target actions for the policy. The method is simple and general, can accommodate continuous and discrete actions, and can be implemented in just a few lines of code on top of standard supervised learning methods. We provide a theoretical motivation for AWR and analyze its properties when incorporating off-policy data from experience replay. We evaluate AWR on a suite of standard OpenAI Gym benchmark tasks, and show that it achieves competitive performance compared to a number of well-established state-of-the-art RL algorithms. AWR is also able to acquire more effective policies than most off-policy algorithms when learning from purely static datasets with no additional environmental interactions. Furthermore, we demonstrate our algorithm on challenging continuous control tasks with highly complex simulated characters.
Tasks	Continuous Control
Published	2019-10-01
URL	https://arxiv.org/abs/1910.00177v3
PDF	https://arxiv.org/pdf/1910.00177v3.pdf
PWC	https://paperswithcode.com/paper/advantage-weighted-regression-simple-and
Repo	https://github.com/peisuke/awr
Framework	none

FlauBERT: Unsupervised Language Model Pre-training for French


Title	FlauBERT: Unsupervised Language Model Pre-training for French
Authors	Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab
Abstract	Language models have become a key step to achieve state-of-the art results in many different Natural Language Processing (NLP) tasks. Leveraging the huge amount of unlabeled texts nowadays available, they provide an efficient way to pre-train continuous word representations that can be fine-tuned for a downstream task, along with their contextualization at the sentence level. This has been widely demonstrated for English using contextualized representations (Dai and Le, 2015; Peters et al., 2018; Howard and Ruder, 2018; Radford et al., 2018; Devlin et al., 2019; Yang et al., 2019b). In this paper, we introduce and share FlauBERT, a model learned on a very large and heterogeneous French corpus. Models of different sizes are trained using the new CNRS (French National Centre for Scientific Research) Jean Zay supercomputer. We apply our French language models to diverse NLP tasks (text classification, paraphrasing, natural language inference, parsing, word sense disambiguation) and show that most of the time they outperform other pre-training approaches. Different versions of FlauBERT as well as a unified evaluation protocol for the downstream tasks, called FLUE (French Language Understanding Evaluation), are shared to the research community for further reproducible experiments in French NLP.
Tasks	Language Modelling, Natural Language Inference, Text Classification, Word Sense Disambiguation
Published	2019-12-11
URL	https://arxiv.org/abs/1912.05372v4
PDF	https://arxiv.org/pdf/1912.05372v4.pdf
PWC	https://paperswithcode.com/paper/flaubert-unsupervised-language-model-pre
Repo	https://github.com/getalp/disambiguate
Framework	pytorch

Dual Sequential Monte Carlo: Tunneling Filtering and Planning in Continuous POMDPs


Title	Dual Sequential Monte Carlo: Tunneling Filtering and Planning in Continuous POMDPs
Authors	Yunbo Wang, Bo Liu, Jiajun Wu, Yuke Zhu, Simon S. Du, Li Fei-Fei, Joshua B. Tenenbaum
Abstract	We present the DualSMC network that solves continuous POMDPs by learning belief representations and then leveraging them for planning. It is based on the fact that filtering, i.e. state estimation, and planning can be viewed as two related sequential Monte Carlo processes, with one in the belief space and the other in the future planning trajectory space. In particular, we first introduce a novel particle filter network that makes better use of the adversarial relationship between the proposer model and the observation model. We then introduce a new planning algorithm over the belief representations, which learns uncertainty-dependent policies. We allow these two parts to be trained jointly with each other. We testify the effectiveness of our approach on three continuous control and planning tasks: the floor positioning, the 3D light-dark navigation, and a modified Reacher task.
Tasks	Continuous Control
Published	2019-09-28
URL	https://arxiv.org/abs/1909.13003v1
PDF	https://arxiv.org/pdf/1909.13003v1.pdf
PWC	https://paperswithcode.com/paper/dual-sequential-monte-carlo-tunneling
Repo	https://github.com/Cranial-XIX/DualSMC
Framework	pytorch

Learning Stabilizable Nonlinear Dynamics with Contraction-Based Regularization


Title	Learning Stabilizable Nonlinear Dynamics with Contraction-Based Regularization
Authors	Sumeet Singh, Spencer M. Richards, Vikas Sindhwani, Jean-Jacques E. Slotine, Marco Pavone
Abstract	We propose a novel framework for learning stabilizable nonlinear dynamical systems for continuous control tasks in robotics. The key contribution is a control-theoretic regularizer for dynamics fitting rooted in the notion of stabilizability, a constraint which guarantees the existence of robust tracking controllers for arbitrary open-loop trajectories generated with the learned system. Leveraging tools from contraction theory and statistical learning in Reproducing Kernel Hilbert Spaces, we formulate stabilizable dynamics learning as a functional optimization with convex objective and bi-convex functional constraints. Under a mild structural assumption and relaxation of the functional constraints to sampling-based constraints, we derive the optimal solution with a modified Representer theorem. Finally, we utilize random matrix feature approximations to reduce the dimensionality of the search parameters and formulate an iterative convex optimization algorithm that jointly fits the dynamics functions and searches for a certificate of stabilizability. We validate the proposed algorithm in simulation for a planar quadrotor, and on a quadrotor hardware testbed emulating planar dynamics. We verify, both in simulation and on hardware, significantly improved trajectory generation and tracking performance with the control-theoretic regularized model over models learned using traditional regression techniques, especially when learning from small supervised datasets. The results support the conjecture that the use of stabilizability constraints as a form of regularization can help prune the hypothesis space in a manner that is tailored to the downstream task of trajectory generation and feedback control, resulting in models that are not only dramatically better conditioned, but also data efficient.
Tasks	Continuous Control
Published	2019-07-29
URL	https://arxiv.org/abs/1907.13122v1
PDF	https://arxiv.org/pdf/1907.13122v1.pdf
PWC	https://paperswithcode.com/paper/learning-stabilizable-nonlinear-dynamics-with
Repo	https://github.com/StanfordASL/SNDL
Framework	none

Parabolic Approximation Line Search: An efficient and effective line search approach for DNNs


Title	Parabolic Approximation Line Search: An efficient and effective line search approach for DNNs
Authors	Maximus Mutschler, Andreas Zell
Abstract	This work shows that line searches performed only over the loss function defined by the current batch can successfully compete with common optimization methods on state-of-the-art architectures in wall clock time. In other words, our approach is performing competitively, despite for the most part ignoring the noise originating from batch sampling of the loss function. Furthermore, we empirically show that local minima on lines in direction of the negative gradient can be estimated almost perfectly by a parabolic approximation. This suggests that the loss function is at least locally convex, which mitigates the common perception of a highly non convex loss landscape. Our approach combines well-known methods such as parabolic approximation, line search and conjugate gradient, to perform an efficient line search. To evaluate general performance as well as the hyper parameter sensitivity of our optimizer, we performed multiple comprehensive hyperparameter grid searches on several datasets and architectures. In addition, we provide a convergence prove on a simplified scenario. PyTorch and Tensorflow implementations are provided at https://github.com/cogsys-tuebingen/PAL.
Tasks
Published	2019-03-28
URL	https://arxiv.org/abs/1903.11991v2
PDF	https://arxiv.org/pdf/1903.11991v2.pdf
PWC	https://paperswithcode.com/paper/pal-a-fast-dnn-optimization-method-based-on
Repo	https://github.com/lessw2020/Best-Deep-Learning-Optimizers
Framework	pytorch

Continual Unsupervised Representation Learning


Title	Continual Unsupervised Representation Learning
Authors	Dushyant Rao, Francesco Visin, Andrei A. Rusu, Yee Whye Teh, Razvan Pascanu, Raia Hadsell
Abstract	Continual learning aims to improve the ability of modern learning systems to deal with non-stationary distributions, typically by attempting to learn a series of tasks sequentially. Prior art in the field has largely considered supervised or reinforcement learning tasks, and often assumes full knowledge of task labels and boundaries. In this work, we propose an approach (CURL) to tackle a more general problem that we will refer to as unsupervised continual learning. The focus is on learning representations without any knowledge about task identity, and we explore scenarios when there are abrupt changes between tasks, smooth transitions from one task to another, or even when the data is shuffled. The proposed approach performs task inference directly within the model, is able to dynamically expand to capture new concepts over its lifetime, and incorporates additional rehearsal-based techniques to deal with catastrophic forgetting. We demonstrate the efficacy of CURL in an unsupervised learning setting with MNIST and Omniglot, where the lack of labels ensures no information is leaked about the task. Further, we demonstrate strong performance compared to prior art in an i.i.d setting, or when adapting the technique to supervised tasks such as incremental class learning.
Tasks	Continual Learning, Omniglot, Representation Learning, Unsupervised Representation Learning
Published	2019-10-31
URL	https://arxiv.org/abs/1910.14481v1
PDF	https://arxiv.org/pdf/1910.14481v1.pdf
PWC	https://paperswithcode.com/paper/continual-unsupervised-representation
Repo	https://github.com/deepmind/deepmind-research
Framework	tf

GCDT: A Global Context Enhanced Deep Transition Architecture for Sequence Labeling


Title	GCDT: A Global Context Enhanced Deep Transition Architecture for Sequence Labeling
Authors	Yijin Liu, Fandong Meng, Jinchao Zhang, Jinan Xu, Yufeng Chen, Jie Zhou
Abstract	Current state-of-the-art systems for sequence labeling are typically based on the family of Recurrent Neural Networks (RNNs). However, the shallow connections between consecutive hidden states of RNNs and insufficient modeling of global information restrict the potential performance of those models. In this paper, we try to address these issues, and thus propose a Global Context enhanced Deep Transition architecture for sequence labeling named GCDT. We deepen the state transition path at each position in a sentence, and further assign every token with a global representation learned from the entire sentence. Experiments on two standard sequence labeling tasks show that, given only training data and the ubiquitous word embeddings (Glove), our GCDT achieves 91.96 F1 on the CoNLL03 NER task and 95.43 F1 on the CoNLL2000 Chunking task, which outperforms the best reported results under the same settings. Furthermore, by leveraging BERT as an additional resource, we establish new state-of-the-art results with 93.47 F1 on NER and 97.30 F1 on Chunking.
Tasks	Chunking, Named Entity Recognition, Word Embeddings
Published	2019-06-06
URL	https://arxiv.org/abs/1906.02437v1
PDF	https://arxiv.org/pdf/1906.02437v1.pdf
PWC	https://paperswithcode.com/paper/gcdt-a-global-context-enhanced-deep
Repo	https://github.com/Adaxry/GCDT
Framework	tf

E3: Entailment-driven Extracting and Editing for Conversational Machine Reading


Title	E3: Entailment-driven Extracting and Editing for Conversational Machine Reading
Authors	Victor Zhong, Luke Zettlemoyer
Abstract	Conversational machine reading systems help users answer high-level questions (e.g. determine if they qualify for particular government benefits) when they do not know the exact rules by which the determination is made(e.g. whether they need certain income levels or veteran status). The key challenge is that these rules are only provided in the form of a procedural text (e.g. guidelines from government website) which the system must read to figure out what to ask the user. We present a new conversational machine reading model that jointly extracts a set of decision rules from the procedural text while reasoning about which are entailed by the conversational history and which still need to be edited to create questions for the user. On the recently introduced ShARC conversational machine reading dataset, our Entailment-driven Extract and Edit network (E3) achieves a new state-of-the-art, outperforming existing systems as well as a new BERT-based baseline. In addition, by explicitly highlighting which information still needs to be gathered, E3 provides a more explainable alternative to prior work. We release source code for our models and experiments at https://github.com/vzhong/e3.
Tasks	Reading Comprehension
Published	2019-06-12
URL	https://arxiv.org/abs/1906.05373v2
PDF	https://arxiv.org/pdf/1906.05373v2.pdf
PWC	https://paperswithcode.com/paper/e3-entailment-driven-extracting-and-editing
Repo	https://github.com/vzhong/e3
Framework	pytorch

Scalable Fair Clustering


Title	Scalable Fair Clustering
Authors	Arturs Backurs, Piotr Indyk, Krzysztof Onak, Baruch Schieber, Ali Vakilian, Tal Wagner
Abstract	We study the fair variant of the classic $k$-median problem introduced by Chierichetti et al. [2017]. In the standard $k$-median problem, given an input pointset $P$, the goal is to find $k$ centers $C$ and assign each input point to one of the centers in $C$ such that the average distance of points to their cluster center is minimized. In the fair variant of $k$-median, the points are colored, and the goal is to minimize the same average distance objective while ensuring that all clusters have an “approximately equal” number of points of each color. Chierichetti et al. proposed a two-phase algorithm for fair $k$-clustering. In the first step, the pointset is partitioned into subsets called fairlets that satisfy the fairness requirement and approximately preserve the $k$-median objective. In the second step, fairlets are merged into $k$ clusters by one of the existing $k$-median algorithms. The running time of this algorithm is dominated by the first step, which takes super-quadratic time. In this paper, we present a practical approximate fairlet decomposition algorithm that runs in nearly linear time. Our algorithm additionally allows for finer control over the balance of resulting clusters than the original work. We complement our theoretical bounds with empirical evaluation.
Tasks
Published	2019-02-10
URL	https://arxiv.org/abs/1902.03519v2
PDF	https://arxiv.org/pdf/1902.03519v2.pdf
PWC	https://paperswithcode.com/paper/scalable-fair-clustering
Repo	https://github.com/talwagner/fair_clustering
Framework	none