October 20, 2019

2769 words 13 mins read

Paper Group AWR 327

Relay: A New IR for Machine Learning Frameworks. Off-line vs. On-line Evaluation of Recommender Systems in Small E-commerce. Cold-start recommendations in Collective Matrix Factorization. CornerNet: Detecting Objects as Paired Keypoints. Compact Policies for Fully-Observable Non-Deterministic Planning as SAT. SpiderCNN: Deep Learning on Point Sets …

Relay: A New IR for Machine Learning Frameworks


Title	Relay: A New IR for Machine Learning Frameworks
Authors	Jared Roesch, Steven Lyubomirsky, Logan Weber, Josh Pollock, Marisa Kirisame, Tianqi Chen, Zachary Tatlock
Abstract	Machine learning powers diverse services in industry including search, translation, recommendation systems, and security. The scale and importance of these models require that they be efficient, expressive, and portable across an array of heterogeneous hardware devices. These constraints are often at odds; in order to better accommodate them we propose a new high-level intermediate representation (IR) called Relay. Relay is being designed as a purely-functional, statically-typed language with the goal of balancing efficient compilation, expressiveness, and portability. We discuss the goals of Relay and highlight its important design constraints. Our prototype is part of the open source NNVM compiler framework, which powers Amazon’s deep learning framework MxNet.
Tasks	Recommendation Systems
Published	2018-09-26
URL	http://arxiv.org/abs/1810.00952v1
PDF	http://arxiv.org/pdf/1810.00952v1.pdf
PWC	https://paperswithcode.com/paper/181000952
Repo	https://github.com/sandeep-krishnamurthy/my-reading-list
Framework	mxnet

Off-line vs. On-line Evaluation of Recommender Systems in Small E-commerce


Title	Off-line vs. On-line Evaluation of Recommender Systems in Small E-commerce
Authors	Ladislav Peska, Peter Vojtas
Abstract	In this paper, we present our work towards comparing on-line and off-line evaluation metrics in the context of small e-commerce recommender systems. Recommending on small e-commerce enterprises is rather challenging due to the lower volume of interactions and low user loyalty, rarely extending beyond a single session. On the other hand, we usually have to deal with lower volumes of objects, which are easier to discover by users through various browsing/searching GUIs. The main goal of this paper is to determine applicability of off-line evaluation metrics in learning true usability of recommender systems (evaluated on-line in A/B testing). In total 800 variants of recommending algorithms were evaluated off-line w.r.t. 18 metrics covering rating-based, ranking-based, novelty and diversity evaluation. The off-line results were afterwards compared with on-line evaluation of 12 selected recommender variants and based on the results, we tried to learn and utilize an off-line to on-line results prediction model. Off-line results shown a great variance in performance w.r.t. different metrics with the Pareto front covering 68% of the approaches. Furthermore, we observed that on-line results are considerably affected by the novelty of users. On-line metrics correlates positively with ranking-based metrics (AUC, MRR, nDCG) for novice users, while too high values of diversity and novelty had a negative impact on the on-line results for them. For users with more visited items, however, the diversity became more important, while ranking-based metrics relevance gradually decrease.
Tasks	Recommendation Systems
Published	2018-09-10
URL	https://arxiv.org/abs/1809.03186v2
PDF	https://arxiv.org/pdf/1809.03186v2.pdf
PWC	https://paperswithcode.com/paper/off-line-vs-on-line-evaluation-of-recommender
Repo	https://github.com/lpeska/UMAP2020
Framework	none

Cold-start recommendations in Collective Matrix Factorization


Title	Cold-start recommendations in Collective Matrix Factorization
Authors	David Cortes
Abstract	This work explores the ability of collective matrix factorization models in recommender systems to make predictions about users and items for which there is side information available but no feedback or interactions data, and proposes a new formulation with a faster cold-start prediction formula that can be used in real-time systems. While these cold-start recommendations are not as good as warm-start ones, they were found to be of better quality than non-personalized recommendations, and predictions about new users were found to be more reliable than those about new items. The formulation proposed here resulted in improved cold-start recommendations in many scenarios, at the expense of worse warm-start ones.
Tasks	Recommendation Systems
Published	2018-09-02
URL	https://arxiv.org/abs/1809.00366v2
PDF	https://arxiv.org/pdf/1809.00366v2.pdf
PWC	https://paperswithcode.com/paper/cold-start-recommendations-in-collective
Repo	https://github.com/david-cortes/cmfrec
Framework	tf

CornerNet: Detecting Objects as Paired Keypoints


Title	CornerNet: Detecting Objects as Paired Keypoints
Authors	Hei Law, Jia Deng
Abstract	We propose CornerNet, a new approach to object detection where we detect an object bounding box as a pair of keypoints, the top-left corner and the bottom-right corner, using a single convolution neural network. By detecting objects as paired keypoints, we eliminate the need for designing a set of anchor boxes commonly used in prior single-stage detectors. In addition to our novel formulation, we introduce corner pooling, a new type of pooling layer that helps the network better localize corners. Experiments show that CornerNet achieves a 42.2% AP on MS COCO, outperforming all existing one-stage detectors.
Tasks	Object Detection
Published	2018-08-03
URL	http://arxiv.org/abs/1808.01244v2
PDF	http://arxiv.org/pdf/1808.01244v2.pdf
PWC	https://paperswithcode.com/paper/cornernet-detecting-objects-as-paired
Repo	https://github.com/jiangzhubo/CornetNet_PPT
Framework	none

Compact Policies for Fully-Observable Non-Deterministic Planning as SAT


Title	Compact Policies for Fully-Observable Non-Deterministic Planning as SAT
Authors	Tomas Geffner, Hector Geffner
Abstract	Fully observable non-deterministic (FOND) planning is becoming increasingly important as an approach for computing proper policies in probabilistic planning, extended temporal plans in LTL planning, and general plans in generalized planning. In this work, we introduce a SAT encoding for FOND planning that is compact and can produce compact strong cyclic policies. Simple variations of the encodings are also introduced for strong planning and for what we call, dual FOND planning, where some non-deterministic actions are assumed to be fair (e.g., probabilistic) and others unfair (e.g., adversarial). The resulting FOND planners are compared empirically with existing planners over existing and new benchmarks. The notion of “probabilistic interesting problems” is also revisited to yield a more comprehensive picture of the strengths and limitations of current FOND planners and the proposed SAT approach.
Tasks
Published	2018-06-25
URL	http://arxiv.org/abs/1806.09455v1
PDF	http://arxiv.org/pdf/1806.09455v1.pdf
PWC	https://paperswithcode.com/paper/compact-policies-for-fully-observable-non
Repo	https://github.com/tomsons22/FOND-SAT
Framework	none

SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters


Title	SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters
Authors	Yifan Xu, Tianqi Fan, Mingye Xu, Long Zeng, Yu Qiao
Abstract	Deep neural networks have enjoyed remarkable success for various vision tasks, however it remains challenging to apply CNNs to domains lacking a regular underlying structures such as 3D point clouds. Towards this we propose a novel convolutional architecture, termed SpiderCNN, to efficiently extract geometric features from point clouds. SpiderCNN is comprised of units called SpiderConv, which extend convolutional operations from regular grids to irregular point sets that can be embedded in R^n, by parametrizing a family of convolutional filters. We design the filter as a product of a simple step function that captures local geodesic information and a Taylor polynomial that ensures the expressiveness. SpiderCNN inherits the multi-scale hierarchical architecture from classical CNNs, which allows it to extract semantic deep features. Experiments on ModelNet40 demonstrate that SpiderCNN achieves state-of-the-art accuracy 92.4% on standard benchmarks, and shows competitive performance on segmentation task.
Tasks	3D Part Segmentation
Published	2018-03-30
URL	http://arxiv.org/abs/1803.11527v3
PDF	http://arxiv.org/pdf/1803.11527v3.pdf
PWC	https://paperswithcode.com/paper/spidercnn-deep-learning-on-point-sets-with
Repo	https://github.com/xyf513/SpiderCNN
Framework	tf

Adversarial Framing for Image and Video Classification


Title	Adversarial Framing for Image and Video Classification
Authors	Konrad Zolna, Michal Zajac, Negar Rostamzadeh, Pedro O. Pinheiro
Abstract	Neural networks are prone to adversarial attacks. In general, such attacks deteriorate the quality of the input by either slightly modifying most of its pixels, or by occluding it with a patch. In this paper, we propose a method that keeps the image unchanged and only adds an adversarial framing on the border of the image. We show empirically that our method is able to successfully attack state-of-the-art methods on both image and video classification problems. Notably, the proposed method results in a universal attack which is very fast at test time. Source code can be found at https://github.com/zajaczajac/adv_framing .
Tasks	Video Classification
Published	2018-12-11
URL	https://arxiv.org/abs/1812.04599v3
PDF	https://arxiv.org/pdf/1812.04599v3.pdf
PWC	https://paperswithcode.com/paper/adversarial-framing-for-image-and-video
Repo	https://github.com/zajaczajac/adv_framing
Framework	pytorch

3D Point Capsule Networks


Title	3D Point Capsule Networks
Authors	Yongheng Zhao, Tolga Birdal, Haowen Deng, Federico Tombari
Abstract	In this paper, we propose 3D point-capsule networks, an auto-encoder designed to process sparse 3D point clouds while preserving spatial arrangements of the input data. 3D capsule networks arise as a direct consequence of our novel unified 3D auto-encoder formulation. Their dynamic routing scheme and the peculiar 2D latent space deployed by our approach bring in improvements for several common point cloud-related tasks, such as object classification, object reconstruction and part segmentation as substantiated by our extensive evaluations. Moreover, it enables new applications such as part interpolation and replacement.
Tasks	3D Feature Matching, 3D Geometry Perception, 3D Object Classification, 3D Object Reconstruction, 3D Part Segmentation, 3D Point Cloud Matching, 3D Shape Generation, 3D Shape Representation
Published	2018-12-27
URL	https://arxiv.org/abs/1812.10775v2
PDF	https://arxiv.org/pdf/1812.10775v2.pdf
PWC	https://paperswithcode.com/paper/3d-point-capsule-networks
Repo	https://github.com/CPUFronz/CapsVoxGAN
Framework	pytorch

HGR-Net: A Fusion Network for Hand Gesture Segmentation and Recognition


Title	HGR-Net: A Fusion Network for Hand Gesture Segmentation and Recognition
Authors	Amirhossein Dadashzadeh, Alireza Tavakoli Targhi, Maryam Tahmasbi, Majid Mirmehdi
Abstract	We propose a two-stage convolutional neural network (CNN) architecture for robust recognition of hand gestures, called HGR-Net, where the first stage performs accurate semantic segmentation to determine hand regions, and the second stage identifies the gesture. The segmentation stage architecture is based on the combination of fully convolutional residual network and atrous spatial pyramid pooling. Although the segmentation sub-network is trained without depth information, it is particularly robust against challenges such as illumination variations and complex backgrounds. The recognition stage deploys a two-stream CNN, which fuses the information from the red-green-blue and segmented images by combining their deep representations in a fully connected layer before classification. Extensive experiments on public datasets show that our architecture achieves almost as good as state-of-the-art performance in segmentation and recognition of static hand gestures, at a fraction of training time, run time, and model size. Our method can operate at an average of 23 ms per frame.
Tasks	Gesture Recognition, Hand Gesture Recognition, Hand-Gesture Recognition, Hand Segmentation, Semantic Segmentation
Published	2018-06-14
URL	https://arxiv.org/abs/1806.05653v3
PDF	https://arxiv.org/pdf/1806.05653v3.pdf
PWC	https://paperswithcode.com/paper/hgr-net-a-fusion-network-for-hand-gesture
Repo	https://github.com/Plrbear/HGR-Net
Framework	tf

Neural Document Summarization by Jointly Learning to Score and Select Sentences


Title	Neural Document Summarization by Jointly Learning to Score and Select Sentences
Authors	Qingyu Zhou, Nan Yang, Furu Wei, Shaohan Huang, Ming Zhou, Tiejun Zhao
Abstract	Sentence scoring and sentence selection are two main steps in extractive document summarization systems. However, previous works treat them as two separated subtasks. In this paper, we present a novel end-to-end neural network framework for extractive document summarization by jointly learning to score and select sentences. It first reads the document sentences with a hierarchical encoder to obtain the representation of sentences. Then it builds the output summary by extracting sentences one by one. Different from previous methods, our approach integrates the selection strategy into the scoring model, which directly predicts the relative importance given previously selected sentences. Experiments on the CNN/Daily Mail dataset show that the proposed framework significantly outperforms the state-of-the-art extractive summarization models.
Tasks	Document Summarization, Extractive Document Summarization
Published	2018-07-06
URL	http://arxiv.org/abs/1807.02305v1
PDF	http://arxiv.org/pdf/1807.02305v1.pdf
PWC	https://paperswithcode.com/paper/neural-document-summarization-by-jointly
Repo	https://github.com/magic282/NeuSum
Framework	pytorch

Transfer Learning for Illustration Classification


Title	Transfer Learning for Illustration Classification
Authors	Manuel Lagunas, Elena Garces
Abstract	The field of image classification has shown an outstanding success thanks to the development of deep learning techniques. Despite the great performance obtained, most of the work has focused on natural images ignoring other domains like artistic depictions. In this paper, we use transfer learning techniques to propose a new classification network with better performance in illustration images. Starting from the deep convolutional network VGG19, pre-trained with natural images, we propose two novel models which learn object representations in the new domain. Our optimized network will learn new low-level features of the images (colours, edges, textures) while keeping the knowledge of the objects and shapes that it already learned from the ImageNet dataset. Thus, requiring much less data for the training. We propose a novel dataset of illustration images labelled by content where our optimized architecture achieves $\textbf{86.61%}$ of top-1 and $\textbf{97.21%}$ of top-5 precision. We additionally demonstrate that our model is still able to recognize objects in photographs.
Tasks	Image Classification, Transfer Learning
Published	2018-05-23
URL	http://arxiv.org/abs/1806.02682v1
PDF	http://arxiv.org/pdf/1806.02682v1.pdf
PWC	https://paperswithcode.com/paper/transfer-learning-for-illustration
Repo	https://github.com/MathiasStensrud/capstone-2
Framework	none

Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation


Title	Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
Authors	Pengyuan Lyu, Cong Yao, Wenhao Wu, Shuicheng Yan, Xiang Bai
Abstract	Previous deep learning based state-of-the-art scene text detection methods can be roughly classified into two categories. The first category treats scene text as a type of general objects and follows general object detection paradigm to localize scene text by regressing the text box locations, but troubled by the arbitrary-orientation and large aspect ratios of scene text. The second one segments text regions directly, but mostly needs complex post processing. In this paper, we present a method that combines the ideas of the two types of methods while avoiding their shortcomings. We propose to detect scene text by localizing corner points of text bounding boxes and segmenting text regions in relative positions. In inference stage, candidate boxes are generated by sampling and grouping corner points, which are further scored by segmentation maps and suppressed by NMS. Compared with previous methods, our method can handle long oriented text naturally and doesn’t need complex post processing. The experiments on ICDAR2013, ICDAR2015, MSRA-TD500, MLT and COCO-Text demonstrate that the proposed algorithm achieves better or comparable results in both accuracy and efficiency. Based on VGG16, it achieves an F-measure of 84.3% on ICDAR2015 and 81.5% on MSRA-TD500.
Tasks	Multi-Oriented Scene Text Detection, Object Detection, Scene Text Detection
Published	2018-02-25
URL	http://arxiv.org/abs/1802.08948v2
PDF	http://arxiv.org/pdf/1802.08948v2.pdf
PWC	https://paperswithcode.com/paper/multi-oriented-scene-text-detection-via
Repo	https://github.com/lvpengyuan/corner
Framework	pytorch

Mean Field Multi-Agent Reinforcement Learning


Title	Mean Field Multi-Agent Reinforcement Learning
Authors	Yaodong Yang, Rui Luo, Minne Li, Ming Zhou, Weinan Zhang, Jun Wang
Abstract	Existing multi-agent reinforcement learning methods are limited typically to a small number of agents. When the agent number increases largely, the learning becomes intractable due to the curse of the dimensionality and the exponential growth of agent interactions. In this paper, we present Mean Field Reinforcement Learning where the interactions within the population of agents are approximated by those between a single agent and the average effect from the overall population or neighboring agents; the interplay between the two entities is mutually reinforced: the learning of the individual agent’s optimal policy depends on the dynamics of the population, while the dynamics of the population change according to the collective patterns of the individual policies. We develop practical mean field Q-learning and mean field Actor-Critic algorithms and analyze the convergence of the solution to Nash equilibrium. Experiments on Gaussian squeeze, Ising model, and battle games justify the learning effectiveness of our mean field approaches. In addition, we report the first result to solve the Ising model via model-free reinforcement learning methods.
Tasks	Multi-agent Reinforcement Learning, Q-Learning
Published	2018-02-15
URL	http://arxiv.org/abs/1802.05438v4
PDF	http://arxiv.org/pdf/1802.05438v4.pdf
PWC	https://paperswithcode.com/paper/mean-field-multi-agent-reinforcement-learning
Repo	https://github.com/mlii/mfrl
Framework	none

Model Reduction with Memory and the Machine Learning of Dynamical Systems


Title	Model Reduction with Memory and the Machine Learning of Dynamical Systems
Authors	Chao Ma, Jianchun Wang, Weinan E
Abstract	The well-known Mori-Zwanzig theory tells us that model reduction leads to memory effect. For a long time, modeling the memory effect accurately and efficiently has been an important but nearly impossible task in developing a good reduced model. In this work, we explore a natural analogy between recurrent neural networks and the Mori-Zwanzig formalism to establish a systematic approach for developing reduced models with memory. Two training models-a direct training model and a dynamically coupled training model-are proposed and compared. We apply these methods to the Kuramoto-Sivashinsky equation and the Navier-Stokes equation. Numerical experiments show that the proposed method can produce reduced model with good performance on both short-term prediction and long-term statistical properties.
Tasks
Published	2018-08-10
URL	http://arxiv.org/abs/1808.04258v1
PDF	http://arxiv.org/pdf/1808.04258v1.pdf
PWC	https://paperswithcode.com/paper/model-reduction-with-memory-and-the-machine
Repo	https://github.com/tccnchsu/PhysicNeuralNetworks
Framework	none

Drop-Activation: Implicit Parameter Reduction and Harmonic Regularization


Title	Drop-Activation: Implicit Parameter Reduction and Harmonic Regularization
Authors	Senwei Liang, Yuehaw Khoo, Haizhao Yang
Abstract	Overfitting frequently occurs in deep learning. In this paper, we propose a novel regularization method called Drop-Activation to reduce overfitting and improve generalization. The key idea is to drop nonlinear activation functions by setting them to be identity functions randomly during training time. During testing, we use a deterministic network with a new activation function to encode the average effect of dropping activations randomly. Our theoretical analyses support the regularization effect of Drop-Activation as implicit parameter reduction and verify its capability to be used together with Batch Normalization (Ioffe and Szegedy 2015). The experimental results on CIFAR-10, CIFAR-100, SVHN, EMNIST, and ImageNet show that Drop-Activation generally improves the performance of popular neural network architectures for the image classification task. Furthermore, as a regularizer Drop-Activation can be used in harmony with standard training and regularization techniques such as Batch Normalization and Auto Augment (Cubuk et al. 2019). The code is available at \url{https://github.com/LeungSamWai/Drop-Activation}.
Tasks	Image Classification
Published	2018-11-14
URL	https://arxiv.org/abs/1811.05850v5
PDF	https://arxiv.org/pdf/1811.05850v5.pdf
PWC	https://paperswithcode.com/paper/drop-activation-implicit-parameter-reduction
Repo	https://github.com/LeungSamWai/Drop-Activation
Framework	pytorch