January 25, 2020

2990 words 15 mins read

Paper Group ANR 1708

Meta-Learning for Contextual Bandit Exploration. CUNI System for the Building Educational Applications 2019 Shared Task: Grammatical Error Correction. A Comparison of Architectures and Pretraining Methods for Contextualized Multilingual Word Embeddings. Heterogeneous causal effects with imperfect compliance: a novel Bayesian machine learning approa …

Meta-Learning for Contextual Bandit Exploration


Title	Meta-Learning for Contextual Bandit Exploration
Authors	Amr Sharaf, Hal Daumé III
Abstract	We describe MELEE, a meta-learning algorithm for learning a good exploration policy in the interactive contextual bandit setting. Here, an algorithm must take actions based on contexts, and learn based only on a reward signal from the action taken, thereby generating an exploration/exploitation trade-off. MELEE addresses this trade-off by learning a good exploration strategy for offline tasks based on synthetic data, on which it can simulate the contextual bandit setting. Based on these simulations, MELEE uses an imitation learning strategy to learn a good exploration policy that can then be applied to true contextual bandit tasks at test time. We compare MELEE to seven strong baseline contextual bandit algorithms on a set of three hundred real-world datasets, on which it outperforms alternatives in most settings, especially when differences in rewards are large. Finally, we demonstrate the importance of having a rich feature representation for learning how to explore.
Tasks	Imitation Learning, Meta-Learning
Published	2019-01-23
URL	http://arxiv.org/abs/1901.08159v1
PDF	http://arxiv.org/pdf/1901.08159v1.pdf
PWC	https://paperswithcode.com/paper/meta-learning-for-contextual-bandit
Repo
Framework

CUNI System for the Building Educational Applications 2019 Shared Task: Grammatical Error Correction


Title	CUNI System for the Building Educational Applications 2019 Shared Task: Grammatical Error Correction
Authors	Jakub Náplava, Milan Straka
Abstract	In this paper, we describe our systems submitted to the Building Educational Applications (BEA) 2019 Shared Task (Bryant et al., 2019). We participated in all three tracks. Our models are NMT systems based on the Transformer model, which we improve by incorporating several enhancements: applying dropout to whole source and target words, weighting target subwords, averaging model checkpoints, and using the trained model iteratively for correcting the intermediate translations. The system in the Restricted Track is trained on the provided corpora with oversampled “cleaner” sentences and reaches 59.39 F0.5 score on the test set. The system in the Low-Resource Track is trained from Wikipedia revision histories and reaches 44.13 F0.5 score. Finally, we finetune the system from the Low-Resource Track on restricted data and achieve 64.55 F0.5 score, placing third in the Unrestricted Track.
Tasks	Grammatical Error Correction
Published	2019-09-12
URL	https://arxiv.org/abs/1909.05553v1
PDF	https://arxiv.org/pdf/1909.05553v1.pdf
PWC	https://paperswithcode.com/paper/cuni-system-for-the-building-educational-1
Repo
Framework

A Comparison of Architectures and Pretraining Methods for Contextualized Multilingual Word Embeddings


Title	A Comparison of Architectures and Pretraining Methods for Contextualized Multilingual Word Embeddings
Authors	Niels van der Heijden, Samira Abnar, Ekaterina Shutova
Abstract	The lack of annotated data in many languages is a well-known challenge within the field of multilingual natural language processing (NLP). Therefore, many recent studies focus on zero-shot transfer learning and joint training across languages to overcome data scarcity for low-resource languages. In this work we (i) perform a comprehensive comparison of state-ofthe-art multilingual word and sentence encoders on the tasks of named entity recognition (NER) and part of speech (POS) tagging; and (ii) propose a new method for creating multilingual contextualized word embeddings, compare it to multiple baselines and show that it performs at or above state-of-theart level in zero-shot transfer settings. Finally, we show that our method allows for better knowledge sharing across languages in a joint training setting.
Tasks	Multilingual Word Embeddings, Named Entity Recognition, Part-Of-Speech Tagging, Transfer Learning, Word Embeddings
Published	2019-12-15
URL	https://arxiv.org/abs/1912.10169v1
PDF	https://arxiv.org/pdf/1912.10169v1.pdf
PWC	https://paperswithcode.com/paper/a-comparison-of-architectures-and-pretraining
Repo
Framework

Heterogeneous causal effects with imperfect compliance: a novel Bayesian machine learning approach


Title	Heterogeneous causal effects with imperfect compliance: a novel Bayesian machine learning approach
Authors	Falco J. Bargagli-Stoffi, Kristof De-Witte, Giorgio Gnecco
Abstract	This paper introduces an innovative Bayesian machine learning algorithm to draw inference on heterogeneous causal effects in the presence of imperfect compliance (e.g., under an irregular assignment mechanism). We show, through Monte Carlo simulations, that the proposed Bayesian Causal Forest with Instrumental Variable (BCF-IV) algorithm outperforms other machine learning techniques tailored for causal inference (namely, Generalized Random Forest and Causal Trees with Instrumental Variable) in estimating the causal effects. Moreover, we show that it converges to an optimal asymptotic performance in discovering the drivers of heterogeneity in a simulated scenario. BCF-IV sheds a light on the heterogeneity of causal effects in instrumental variable scenarios and, in turn, provides the policy-makers with a relevant tool for targeted policies. Its empirical application evaluates the effects of additional funding on students’ performances. The results indicate that BCF-IV could be used to enhance the effectiveness of school funding on students’ performance by 3.2 to 3.5 times.
Tasks	Causal Inference
Published	2019-05-29
URL	https://arxiv.org/abs/1905.12707v2
PDF	https://arxiv.org/pdf/1905.12707v2.pdf
PWC	https://paperswithcode.com/paper/heterogeneous-causal-effects-with-imperfect
Repo
Framework

Sample Complexity of Sample Average Approximation for Conditional Stochastic Optimization


Title	Sample Complexity of Sample Average Approximation for Conditional Stochastic Optimization
Authors	Yifan Hu, Xin Chen, Niao He
Abstract	In this paper, we study a class of stochastic optimization problems, referred to as the \emph{Conditional Stochastic Optimization} (CSO), in the form of $\min_{x \in \mathcal{X}} \EE_{\xi}f_\xi\Big({\EE_{\eta\xi}[g_\eta(x,\xi)]}\Big)$, which finds a wide spectrum of applications including portfolio selection, reinforcement learning, robust learning, causal inference and so on. Assuming availability of samples from the distribution $\PP(\xi)$ and samples from the conditional distribution $\PP(\eta\xi)$, we establish the sample complexity of the sample average approximation (SAA) for CSO, under a variety of structural assumptions, such as Lipschitz continuity, smoothness, and error bound conditions. We show that the total sample complexity improves from $\cO(d/\eps^4)$ to $\cO(d/\eps^3)$ when assuming smoothness of the outer function, and further to $\cO(1/\eps^2)$ when the empirical function satisfies the quadratic growth condition. We also establish the sample complexity of a modified SAA, when $\xi$ and $\eta$ are independent. Several numerical experiments further support our theoretical findings. Keywords: stochastic optimization, sample average approximation, large deviations theory
Tasks	Causal Inference, Stochastic Optimization
Published	2019-05-28
URL	https://arxiv.org/abs/1905.11957v2
PDF	https://arxiv.org/pdf/1905.11957v2.pdf
PWC	https://paperswithcode.com/paper/sample-complexity-of-sample-average
Repo
Framework

Epistemic Uncertainty Sampling


Title	Epistemic Uncertainty Sampling
Authors	Vu-Linh Nguyen, Sébastien Destercke, Eyke Hüllermeier
Abstract	Various strategies for active learning have been proposed in the machine learning literature. In uncertainty sampling, which is among the most popular approaches, the active learner sequentially queries the label of those instances for which its current prediction is maximally uncertain. The predictions as well as the measures used to quantify the degree of uncertainty, such as entropy, are almost exclusively of a probabilistic nature. In this paper, we advocate a distinction between two different types of uncertainty, referred to as epistemic and aleatoric, in the context of active learning. Roughly speaking, these notions capture the reducible and the irreducible part of the total uncertainty in a prediction, respectively. We conjecture that, in uncertainty sampling, the usefulness of an instance is better reflected by its epistemic than by its aleatoric uncertainty. This leads us to suggest the principle of “epistemic uncertainty sampling”, which we instantiate by means of a concrete approach for measuring epistemic and aleatoric uncertainty. In experimental studies, epistemic uncertainty sampling does indeed show promising performance.
Tasks	Active Learning
Published	2019-08-31
URL	https://arxiv.org/abs/1909.00218v1
PDF	https://arxiv.org/pdf/1909.00218v1.pdf
PWC	https://paperswithcode.com/paper/epistemic-uncertainty-sampling
Repo
Framework

Ordinal Distance Metric Learning with MDS for Image Ranking


Title	Ordinal Distance Metric Learning with MDS for Image Ranking
Authors	Panpan Yu, Qingna Li
Abstract	Image ranking is to rank images based on some known ranked images. In this paper, we propose an improved linear ordinal distance metric learning approach based on the linear distance metric learning model. By decomposing the distance metric $A$ as $L^TL$, the problem can be cast as looking for a linear map between two sets of points in different spaces, meanwhile maintaining some data structures. The ordinal relation of the labels can be maintained via classical multidimensional scaling, a popular tool for dimension reduction in statistics. A least squares fitting term is then introduced to the cost function, which can also maintain the local data structure. The resulting model is an unconstrained problem, and can better fit the data structure. Extensive numerical results demonstrate the improvement of the new approach over the linear distance metric learning model both in speed and ranking performance.
Tasks	Dimensionality Reduction, Metric Learning
Published	2019-02-27
URL	http://arxiv.org/abs/1902.10284v1
PDF	http://arxiv.org/pdf/1902.10284v1.pdf
PWC	https://paperswithcode.com/paper/ordinal-distance-metric-learning-with-mds-for
Repo
Framework

Temporal Coherence for Active Learning in Videos


Title	Temporal Coherence for Active Learning in Videos
Authors	Javad Zolfaghari Bengar, Abel Gonzalez-Garcia, Gabriel Villalonga, Bogdan Raducanu, Hamed H. Aghdam, Mikhail Mozerov, Antonio M. Lopez, Joost van de Weijer
Abstract	Autonomous driving systems require huge amounts of data to train. Manual annotation of this data is time-consuming and prohibitively expensive since it involves human resources. Therefore, active learning emerged as an alternative to ease this effort and to make data annotation more manageable. In this paper, we introduce a novel active learning approach for object detection in videos by exploiting temporal coherence. Our active learning criterion is based on the estimated number of errors in terms of false positives and false negatives. The detections obtained by the object detector are used to define the nodes of a graph and tracked forward and backward to temporally link the nodes. Minimizing an energy function defined on this graphical model provides estimates of both false positives and false negatives. Additionally, we introduce a synthetic video dataset, called SYNTHIA-AL, specially designed to evaluate active learning for video object detection in road scenes. Finally, we show that our approach outperforms active learning baselines tested on two datasets.
Tasks	Active Learning, Autonomous Driving, Object Detection, Video Object Detection
Published	2019-08-30
URL	https://arxiv.org/abs/1908.11757v1
PDF	https://arxiv.org/pdf/1908.11757v1.pdf
PWC	https://paperswithcode.com/paper/temporal-coherence-for-active-learning-in
Repo
Framework

Active Learning for Domain Classification in a Commercial Spoken Personal Assistant


Title	Active Learning for Domain Classification in a Commercial Spoken Personal Assistant
Authors	Xi C. Chen, Adithya Sagar, Justine T. Kao, Tony Y. Li, Christopher Klein, Stephen Pulman, Ashish Garg, Jason D. Williams
Abstract	We describe a method for selecting relevant new training data for the LSTM-based domain selection component of our personal assistant system. Adding more annotated training data for any ML system typically improves accuracy, but only if it provides examples not already adequately covered in the existing data. However, obtaining, selecting, and labeling relevant data is expensive. This work presents a simple technique that automatically identifies new helpful examples suitable for human annotation. Our experimental results show that the proposed method, compared with random-selection and entropy-based methods, leads to higher accuracy improvements given a fixed annotation budget. Although developed and tested in the setting of a commercial intelligent assistant, the technique is of wider applicability.
Tasks	Active Learning
Published	2019-08-29
URL	https://arxiv.org/abs/1908.11404v1
PDF	https://arxiv.org/pdf/1908.11404v1.pdf
PWC	https://paperswithcode.com/paper/active-learning-for-domain-classification-in
Repo
Framework

Cubic LSTMs for Video Prediction


Title	Cubic LSTMs for Video Prediction
Authors	Hehe Fan, Linchao Zhu, Yi Yang
Abstract	Predicting future frames in videos has become a promising direction of research for both computer vision and robot learning communities. The core of this problem involves moving object capture and future motion prediction. While object capture specifies which objects are moving in videos, motion prediction describes their future dynamics. Motivated by this analysis, we propose a Cubic Long Short-Term Memory (CubicLSTM) unit for video prediction. CubicLSTM consists of three branches, i.e., a spatial branch for capturing moving objects, a temporal branch for processing motions, and an output branch for combining the first two branches to generate predicted frames. Stacking multiple CubicLSTM units along the spatial branch and output branch, and then evolving along the temporal branch can form a cubic recurrent neural network (CubicRNN). Experiment shows that CubicRNN produces more accurate video predictions than prior methods on both synthetic and real-world datasets.
Tasks	motion prediction, Video Prediction
Published	2019-04-20
URL	http://arxiv.org/abs/1904.09412v1
PDF	http://arxiv.org/pdf/1904.09412v1.pdf
PWC	https://paperswithcode.com/paper/190409412
Repo
Framework

KeyIn: Discovering Subgoal Structure with Keyframe-based Video Prediction


Title	KeyIn: Discovering Subgoal Structure with Keyframe-based Video Prediction
Authors	Karl Pertsch, Oleh Rybkin, Jingyun Yang, Kosta Derpanis, Joseph Lim, Kostas Daniilidis, Andrew Jaegle
Abstract	Real-world image sequences can often be naturally decomposed into a small number of frames depicting interesting, highly stochastic moments (its $\textit{keyframes}$) and the low-variance frames in between them. In image sequences depicting trajectories to a goal, keyframes can be seen as capturing the $\textit{subgoals}$ of the sequence as they depict the high-variance moments of interest that ultimately led to the goal. In this paper, we introduce a video prediction model that discovers the keyframe structure of image sequences in an unsupervised fashion. We do so using a hierarchical Keyframe-Intermediate model (KeyIn) that stochastically predicts keyframes and their offsets in time and then uses these predictions to deterministically predict the intermediate frames. We propose a differentiable formulation of this problem that allows us to train the full hierarchical model using a sequence reconstruction loss. We show that our model is able to find meaningful keyframe structure in a simulated dataset of robotic demonstrations and that these keyframes can serve as subgoals for planning. Our model outperforms other hierarchical prediction approaches for planning on a simulated pushing task.
Tasks	Video Prediction
Published	2019-04-11
URL	http://arxiv.org/abs/1904.05869v1
PDF	http://arxiv.org/pdf/1904.05869v1.pdf
PWC	https://paperswithcode.com/paper/keyin-discovering-subgoal-structure-with
Repo
Framework

The Journey is the Reward: Unsupervised Learning of Influential Trajectories


Title	The Journey is the Reward: Unsupervised Learning of Influential Trajectories
Authors	Jonathan Binas, Sherjil Ozair, Yoshua Bengio
Abstract	Unsupervised exploration and representation learning become increasingly important when learning in diverse and sparse environments. The information-theoretic principle of empowerment formalizes an unsupervised exploration objective through an agent trying to maximize its influence on the future states of its environment. Previous approaches carry certain limitations in that they either do not employ closed-loop feedback or do not have an internal state. As a consequence, a privileged final state is taken as an influence measure, rather than the full trajectory. We provide a model-free method which takes into account the whole trajectory while still offering the benefits of option-based approaches. We successfully apply our approach to settings with large action spaces, where discovery of meaningful action sequences is particularly difficult.
Tasks	Representation Learning
Published	2019-05-22
URL	https://arxiv.org/abs/1905.09334v1
PDF	https://arxiv.org/pdf/1905.09334v1.pdf
PWC	https://paperswithcode.com/paper/the-journey-is-the-reward-unsupervised
Repo
Framework

Large Scale Joint Semantic Re-Localisation and Scene Understanding via Globally Unique Instance Coordinate Regression


Title	Large Scale Joint Semantic Re-Localisation and Scene Understanding via Globally Unique Instance Coordinate Regression
Authors	Ignas Budvytis, Marvin Teichmann, Tomas Vojir, Roberto Cipolla
Abstract	In this work we present a novel approach to joint semantic localisation and scene understanding. Our work is motivated by the need for localisation algorithms which not only predict 6-DoF camera pose but also simultaneously recognise surrounding objects and estimate 3D geometry. Such capabilities are crucial for computer vision guided systems which interact with the environment: autonomous driving, augmented reality and robotics. In particular, we propose a two step procedure. During the first step we train a convolutional neural network to jointly predict per-pixel globally unique instance labels and corresponding local coordinates for each instance of a static object (e.g. a building). During the second step we obtain scene coordinates by combining object center coordinates and local coordinates and use them to perform 6-DoF camera pose estimation. We evaluate our approach on real world (CamVid-360) and artificial (SceneCity) autonomous driving datasets. We obtain smaller mean distance and angular errors than state-of-the-art 6-DoF pose estimation algorithms based on direct pose regression and pose estimation from scene coordinates on all datasets. Our contributions include: (i) a novel formulation of scene coordinate regression as two separate tasks of object instance recognition and local coordinate regression and a demonstration that our proposed solution allows to predict accurate 3D geometry of static objects and estimate 6-DoF pose of camera on (ii) maps larger by several orders of magnitude than previously attempted by scene coordinate regression methods, as well as on (iii) lightweight, approximate 3D maps built from 3D primitives such as building-aligned cuboids.
Tasks	Autonomous Driving, Pose Estimation, Scene Understanding
Published	2019-09-23
URL	https://arxiv.org/abs/1909.10239v1
PDF	https://arxiv.org/pdf/1909.10239v1.pdf
PWC	https://paperswithcode.com/paper/190910239
Repo
Framework

Stress-Plus-X (SPX) Graph Layout


Title	Stress-Plus-X (SPX) Graph Layout
Authors	Sabin Devkota, Reyan Ahmed, Felice De Luca, Katherine E. Isaacs, Stephen Kobourov
Abstract	Stress, edge crossings, and crossing angles play an important role in the quality and readability of graph drawings. Most standard graph drawing algorithms optimize one of these criteria which may lead to layouts that are deficient in other criteria. We introduce an optimization framework, Stress-Plus-X (SPX), that simultaneously optimizes stress together with several other criteria: edge crossings, minimum crossing angle, and upwardness (for directed acyclic graphs). SPX achieves results that are close to the state-of-the-art algorithms that optimize these metrics individually. SPX is flexible and extensible and can optimize a subset or all of these criteria simultaneously. Our experimental analysis shows that our joint optimization approach is successful in drawing graphs with good performance across readability criteria.
Tasks
Published	2019-08-04
URL	https://arxiv.org/abs/1908.01769v5
PDF	https://arxiv.org/pdf/1908.01769v5.pdf
PWC	https://paperswithcode.com/paper/stress-plus-x-spx-graph-layout
Repo
Framework

Synthetic Data for Deep Learning


Title	Synthetic Data for Deep Learning
Authors	Sergey I. Nikolenko
Abstract	Synthetic data is an increasingly popular tool for training deep learning models, especially in computer vision but also in other areas. In this work, we attempt to provide a comprehensive survey of the various directions in the development and application of synthetic data. First, we discuss synthetic datasets for basic computer vision problems, both low-level (e.g., optical flow estimation) and high-level (e.g., semantic segmentation), synthetic environments and datasets for outdoor and urban scenes (autonomous driving), indoor scenes (indoor navigation), aerial navigation, simulation environments for robotics, applications of synthetic data outside computer vision (in neural programming, bioinformatics, NLP, and more); we also survey the work on improving synthetic data development and alternative ways to produce it such as GANs. Second, we discuss in detail the synthetic-to-real domain adaptation problem that inevitably arises in applications of synthetic data, including synthetic-to-real refinement with GAN-based models and domain adaptation at the feature/model level without explicit data transformations. Third, we turn to privacy-related applications of synthetic data and review the work on generating synthetic datasets with differential privacy guarantees. We conclude by highlighting the most promising directions for further work in synthetic data studies.
Tasks	Autonomous Driving, Domain Adaptation, Optical Flow Estimation, Semantic Segmentation
Published	2019-09-25
URL	https://arxiv.org/abs/1909.11512v1
PDF	https://arxiv.org/pdf/1909.11512v1.pdf
PWC	https://paperswithcode.com/paper/synthetic-data-for-deep-learning
Repo
Framework