October 19, 2019

3161 words 15 mins read

Paper Group ANR 217

Looking Beyond a Clever Narrative: Visual Context and Attention are Primary Drivers of Affect in Video Advertisements. Composite Binary Decomposition Networks. Adapting Semantic Segmentation Models for Changes in Illumination and Camera Perspective. Value Alignment, Fair Play, and the Rights of Service Robots. Quantitative Evaluation of Base and De …

Looking Beyond a Clever Narrative: Visual Context and Attention are Primary Drivers of Affect in Video Advertisements


Title	Looking Beyond a Clever Narrative: Visual Context and Attention are Primary Drivers of Affect in Video Advertisements
Authors	Abhinav Shukla, Harish Katti, Mohan Kankanhalli, Ramanathan Subramanian
Abstract	Emotion evoked by an advertisement plays a key role in influencing brand recall and eventual consumer choices. Automatic ad affect recognition has several useful applications. However, the use of content-based feature representations does not give insights into how affect is modulated by aspects such as the ad scene setting, salient object attributes and their interactions. Neither do such approaches inform us on how humans prioritize visual information for ad understanding. Our work addresses these lacunae by decomposing video content into detected objects, coarse scene structure, object statistics and actively attended objects identified via eye-gaze. We measure the importance of each of these information channels by systematically incorporating related information into ad affect prediction models. Contrary to the popular notion that ad affect hinges on the narrative and the clever use of linguistic and social cues, we find that actively attended objects and the coarse scene structure better encode affective information as compared to individual scene objects or conspicuous background elements.
Tasks
Published	2018-08-14
URL	http://arxiv.org/abs/1808.04610v1
PDF	http://arxiv.org/pdf/1808.04610v1.pdf
PWC	https://paperswithcode.com/paper/looking-beyond-a-clever-narrative-visual
Repo
Framework

Composite Binary Decomposition Networks


Title	Composite Binary Decomposition Networks
Authors	You Qiaoben, Zheng Wang, Jianguo Li, Yinpeng Dong, Yu-Gang Jiang, Jun Zhu
Abstract	Binary neural networks have great resource and computing efficiency, while suffer from long training procedure and non-negligible accuracy drops, when comparing to the full-precision counterparts. In this paper, we propose the composite binary decomposition networks (CBDNet), which first compose real-valued tensor of each layer with a limited number of binary tensors, and then decompose some conditioned binary tensors into two low-rank binary tensors, so that the number of parameters and operations are greatly reduced comparing to the original ones. Experiments demonstrate the effectiveness of the proposed method, as CBDNet can approximate image classification network ResNet-18 using 5.25 bits, VGG-16 using 5.47 bits, DenseNet-121 using 5.72 bits, object detection networks SSD300 using 4.38 bits, and semantic segmentation networks SegNet using 5.18 bits, all with minor accuracy drops.
Tasks	Image Classification, Object Detection, Semantic Segmentation
Published	2018-11-16
URL	http://arxiv.org/abs/1811.06668v1
PDF	http://arxiv.org/pdf/1811.06668v1.pdf
PWC	https://paperswithcode.com/paper/composite-binary-decomposition-networks
Repo
Framework

Adapting Semantic Segmentation Models for Changes in Illumination and Camera Perspective


Title	Adapting Semantic Segmentation Models for Changes in Illumination and Camera Perspective
Authors	Wei Zhou, Alex Zyner, Stewart Worrall, Eduardo Nebot
Abstract	Semantic segmentation using deep neural networks has been widely explored to generate high-level contextual information for autonomous vehicles. To acquire a complete $180^\circ$ semantic understanding of the forward surroundings, we propose to stitch semantic images from multiple cameras with varying orientations. However, previously trained semantic segmentation models showed unacceptable performance after significant changes to the camera orientations and the lighting conditions. To avoid time-consuming hand labeling, we explore and evaluate the use of data augmentation techniques, specifically skew and gamma correction, from a practical real-world standpoint to extend the existing model and provide more robust performance. The presented experimental results have shown significant improvements with varying illumination and camera perspective changes.
Tasks	Autonomous Vehicles, Data Augmentation, Semantic Segmentation
Published	2018-09-13
URL	http://arxiv.org/abs/1809.04730v1
PDF	http://arxiv.org/pdf/1809.04730v1.pdf
PWC	https://paperswithcode.com/paper/adapting-semantic-segmentation-models-for
Repo
Framework

Value Alignment, Fair Play, and the Rights of Service Robots


Title	Value Alignment, Fair Play, and the Rights of Service Robots
Authors	Daniel Estrada
Abstract	Ethics and safety research in artificial intelligence is increasingly framed in terms of “alignment” with human values and interests. I argue that Turing’s call for “fair play for machines” is an early and often overlooked contribution to the alignment literature. Turing’s appeal to fair play suggests a need to correct human behavior to accommodate our machines, a surprising inversion of how value alignment is treated today. Reflections on “fair play” motivate a novel interpretation of Turing’s notorious “imitation game” as a condition not of intelligence but instead of value alignment: a machine demonstrates a minimal degree of alignment (with the norms of conversation, for instance) when it can go undetected when interrogated by a human. I carefully distinguish this interpretation from the Moral Turing Test, which is not motivated by a principle of fair play, but instead depends on imitation of human moral behavior. Finally, I consider how the framework of fair play can be used to situate the debate over robot rights within the alignment literature. I argue that extending rights to service robots operating in public spaces is “fair” in precisely the sense that it encourages an alignment of interests between humans and machines.
Tasks
Published	2018-03-07
URL	http://arxiv.org/abs/1803.02852v1
PDF	http://arxiv.org/pdf/1803.02852v1.pdf
PWC	https://paperswithcode.com/paper/value-alignment-fair-play-and-the-rights-of
Repo
Framework

Quantitative Evaluation of Base and Detail Decomposition Filters Based on their Artifacts


Title	Quantitative Evaluation of Base and Detail Decomposition Filters Based on their Artifacts
Authors	Charles Hessel, Jean-Michel Morel
Abstract	This paper introduces a quantitative evaluation of filters that seek to separate an image into its large-scale variations, the base layer, and its fine-scale variations, the detail layer. Such methods have proliferated with the development of HDR imaging and the proposition of many new tone-mapping operators. We argue that an objective quality measurement for all methods can be based on their artifacts. To this aim, the four main recurrent artifacts are described and mathematically characterized. Among them two are classic, the luminance halo and the staircase effect, but we show the relevance of two more, the contrast halo and the compartmentalization effect. For each of these artifacts we design a test-pattern and its attached measurement formula. Then we fuse these measurements into a single quality mark, and obtain in that way a ranking method valid for all filters performing a base+detail decomposition. This synthetic ranking is applied to seven filters representative of the literature and shown to agree with expert artifact rejection criteria.
Tasks
Published	2018-08-28
URL	http://arxiv.org/abs/1808.09411v1
PDF	http://arxiv.org/pdf/1808.09411v1.pdf
PWC	https://paperswithcode.com/paper/quantitative-evaluation-of-base-and-detail
Repo
Framework

Simulation assisted machine learning


Title	Simulation assisted machine learning
Authors	Timo M. Deist, Andrew Patti, Zhaoqi Wang, David Krane, Taylor Sorenson, David Craft
Abstract	Motivation: In a predictive modeling setting, if sufficient details of the system behavior are known, one can build and use a simulation for making predictions. When sufficient system details are not known, one typically turns to machine learning, which builds a black-box model of the system using a large dataset of input sample features and outputs. We consider a setting which is between these two extremes: some details of the system mechanics are known but not enough for creating simulations that can be used to make high quality predictions. In this context we propose using approximate simulations to build a kernel for use in kernelized machine learning methods, such as support vector machines. The results of multiple simulations (under various uncertainty scenarios) are used to compute similarity measures between every pair of samples: sample pairs are given a high similarity score if they behave similarly under a wide range of simulation parameters. These similarity values, rather than the original high dimensional feature data, are used to build the kernel. Results: We demonstrate and explore the simulation based kernel (SimKern) concept using four synthetic complex systems–three biologically inspired models and one network flow optimization model. We show that, when the number of training samples is small compared to the number of features, the SimKern approach dominates over no-prior-knowledge methods. This approach should be applicable in all disciplines where predictive models are sought and informative yet approximate simulations are available. Availability: The Python SimKern software, the demonstration models (in MATLAB, R), and the datasets are available at https://github.com/davidcraft/SimKern.
Tasks
Published	2018-02-15
URL	https://arxiv.org/abs/1802.05688v4
PDF	https://arxiv.org/pdf/1802.05688v4.pdf
PWC	https://paperswithcode.com/paper/simulation-assisted-machine-learning
Repo
Framework

Model-Free Optimization Using Eagle Perching Optimizer


Title	Model-Free Optimization Using Eagle Perching Optimizer
Authors	Ameer Tamoor Khan, Shuai Li Senior, Predrag S. Stanimirovic, Yinyan Zhang
Abstract	The paper proposes a novel nature-inspired technique of optimization. It mimics the perching nature of eagles and uses mathematical formulations to introduce a new addition to metaheuristic algorithms. The nature of the proposed algorithm is based on exploration and exploitation. The proposed algorithm is developed into two versions with some modifications. In the first phase, it undergoes a rigorous analysis to find out their performance. In the second phase it is benchmarked using ten functions of two categories; uni-modal functions and multi-modal functions. In the third phase, we conducted a detailed analysis of the algorithm by exploiting its controlling units or variables. In the fourth and last phase, we consider real world optimization problems with constraints. Both versions of the algorithm show an appreciable performance, but analysis puts more weight to the modified version. The competitive analysis shows that the proposed algorithm outperforms the other tested metaheuristic algorithms. The proposed method has better robustness and computational efficiency.
Tasks
Published	2018-07-08
URL	http://arxiv.org/abs/1807.02754v1
PDF	http://arxiv.org/pdf/1807.02754v1.pdf
PWC	https://paperswithcode.com/paper/model-free-optimization-using-eagle-perching
Repo
Framework

Monocular 3D Pose Recovery via Nonconvex Sparsity with Theoretical Analysis


Title	Monocular 3D Pose Recovery via Nonconvex Sparsity with Theoretical Analysis
Authors	Jianqiao Wangni, Dahua Lin, Ji Liu, Kostas Daniilidis, Jianbo Shi
Abstract	For recovering 3D object poses from 2D images, a prevalent method is to pre-train an over-complete dictionary $\mathcal D={B_i}_i^D$ of 3D basis poses. During testing, the detected 2D pose $Y$ is matched to dictionary by $Y \approx \sum_i M_i B_i$ where ${M_i}i^D={c_i \Pi R_i}$, by estimating the rotation $R_i$, projection $\Pi$ and sparse combination coefficients $c \in \mathbb R{+}^D$. In this paper, we propose non-convex regularization $H(c)$ to learn coefficients $c$, including novel leaky capped $\ell_1$-norm regularization (LCNR), \begin{align} H(c)=\alpha \sum_{i } \min(c_i,\tau)+ \beta \sum_{i } \max( c_i,\tau), \end{align} where $0\leq \beta \leq \alpha$ and $0<\tau$ is a certain threshold, so the invalid components smaller than $\tau$ are composed with larger regularization and other valid components with smaller regularization. We propose a multi-stage optimizer with convex relaxation and ADMM. We prove that the estimation error $\mathcal L(l)$ decays w.r.t. the stages $l$, \begin{align} Pr\left(\mathcal L(l) < \rho^{l-1} \mathcal L(0) + \delta \right) \geq 1- \epsilon, \end{align} where $0< \rho <1, 0<\delta, 0<\epsilon \ll 1$. Experiments on large 3D human datasets like H36M are conducted to support our improvement upon previous approaches. To the best of our knowledge, this is the first theoretical analysis in this line of research, to understand how the recovery error is affected by fundamental factors, e.g. dictionary size, observation noises, optimization times. We characterize the trade-off between speed and accuracy towards real-time inference in applications.
Tasks
Published	2018-12-29
URL	http://arxiv.org/abs/1812.11295v1
PDF	http://arxiv.org/pdf/1812.11295v1.pdf
PWC	https://paperswithcode.com/paper/monocular-3d-pose-recovery-via-nonconvex
Repo
Framework

Efficient sparse semismooth Newton methods for the clustered lasso problem


Title	Efficient sparse semismooth Newton methods for the clustered lasso problem
Authors	Meixia Lin, Yong-Jin Liu, Defeng Sun, Kim-Chuan Toh
Abstract	We focus on solving the clustered lasso problem, which is a least squares problem with the $\ell_1$-type penalties imposed on both the coefficients and their pairwise differences to learn the group structure of the regression parameters. Here we first reformulate the clustered lasso regularizer as a weighted ordered-lasso regularizer, which is essential in reducing the computational cost from $O(n^2)$ to $O(n\log (n))$. We then propose an inexact semismooth Newton augmented Lagrangian ({\sc Ssnal}) algorithm to solve the clustered lasso problem or its dual via this equivalent formulation, depending on whether the sample size is larger than the dimension of the features. An essential component of the {\sc Ssnal} algorithm is the computation of the generalized Jacobian of the proximal mapping of the clustered lasso regularizer. Based on the new formulation, we derive an efficient procedure for its computation. Comprehensive results on the global convergence and local linear convergence of the {\sc Ssnal} algorithm are established. For the purpose of exposition and comparison, we also summarize/design several first-order methods that can be used to solve the problem under consideration, but with the key improvement from the new formulation of the clustered lasso regularizer. As a demonstration of the applicability of our algorithms, numerical experiments on the clustered lasso problem are performed. The experiments show that the {\sc Ssnal} algorithm substantially outperforms the best alternative algorithm for the clustered lasso problem.
Tasks
Published	2018-08-22
URL	http://arxiv.org/abs/1808.07181v3
PDF	http://arxiv.org/pdf/1808.07181v3.pdf
PWC	https://paperswithcode.com/paper/efficient-sparse-hessian-based-algorithms-for
Repo
Framework

Post-training 4-bit quantization of convolution networks for rapid-deployment


Title	Post-training 4-bit quantization of convolution networks for rapid-deployment
Authors	Ron Banner, Yury Nahshan, Elad Hoffer, Daniel Soudry
Abstract	Convolutional neural networks require significant memory bandwidth and storage for intermediate computations, apart from substantial computing resources. Neural network quantization has significant benefits in reducing the amount of intermediate results, but it often requires the full datasets and time-consuming fine tuning to recover the accuracy lost after quantization. This paper introduces the first practical 4-bit post training quantization approach: it does not involve training the quantized model (fine-tuning), nor it requires the availability of the full dataset. We target the quantization of both activations and weights and suggest three complementary methods for minimizing quantization error at the tensor level, two of whom obtain a closed-form analytical solution. Combining these methods, our approach achieves accuracy that is just a few percents less the state-of-the-art baseline across a wide range of convolutional models. The source code to replicate all experiments is available on GitHub: \url{https://github.com/submission2019/cnn-quantization}.
Tasks	Quantization
Published	2018-10-02
URL	https://arxiv.org/abs/1810.05723v3
PDF	https://arxiv.org/pdf/1810.05723v3.pdf
PWC	https://paperswithcode.com/paper/post-training-4-bit-quantization-of
Repo
Framework

Generative Spatiotemporal Modeling Of Neutrophil Behavior


Title	Generative Spatiotemporal Modeling Of Neutrophil Behavior
Authors	Narita Pandhe, Balazs Rada, Shannon Quinn
Abstract	Cell motion and appearance have a strong correlation with cell cycle and disease progression. Many contemporary efforts in machine learning utilize spatio-temporal models to predict a cell’s physical state and, consequently, the advancement of disease. Alternatively, generative models learn the underlying distribution of the data, creating holistic representations that can be used in learning. In this work, we propose an aggregate model that combine Generative Adversarial Networks (GANs) and Autoregressive (AR) models to predict cell motion and appearance in human neutrophils imaged by differential interference contrast (DIC) microscopy. We bifurcate the task of learning cell statistics by leveraging GANs for the spatial component and AR models for the temporal component. The aggregate model learned results offer a promising computational environment for studying changes in organellar shape, quantity, and spatial distribution over large sequences.
Tasks
Published	2018-04-02
URL	http://arxiv.org/abs/1804.00393v1
PDF	http://arxiv.org/pdf/1804.00393v1.pdf
PWC	https://paperswithcode.com/paper/generative-spatiotemporal-modeling-of
Repo
Framework

Fast and Globally Optimal Rigid Registration of 3D Point Sets by Transformation Decomposition


Title	Fast and Globally Optimal Rigid Registration of 3D Point Sets by Transformation Decomposition
Authors	Xuechen Li, Yinlong Liu, Yiru Wang, Chen Wang, Manning Wang, Zhijian Song
Abstract	The rigid registration of two 3D point sets is a fundamental problem in computer vision. The current trend is to solve this problem globally using the BnB optimization framework. However, the existing global methods are slow for two main reasons: the computational complexity of BnB is exponential to the problem dimensionality (which is six for 3D rigid registration), and the bound evaluation used in BnB is inefficient. In this paper, we propose two techniques to address these problems. First, we introduce the idea of translation invariant vectors, which allows us to decompose the search of a 6D rigid transformation into a search of 3D rotation followed by a search of 3D translation, each of which is solved by a separate BnB algorithm. This transformation decomposition reduces the problem dimensionality of BnB algorithms and substantially improves its efficiency. Then, we propose a new data structure, named 3D Integral Volume, to accelerate the bound evaluation in both BnB algorithms. By combining these two techniques, we implement an efficient algorithm for rigid registration of 3D point sets. Extensive experiments on both synthetic and real data show that the proposed algorithm is three orders of magnitude faster than the existing state-of-the-art global methods.
Tasks
Published	2018-12-29
URL	http://arxiv.org/abs/1812.11307v3
PDF	http://arxiv.org/pdf/1812.11307v3.pdf
PWC	https://paperswithcode.com/paper/fast-and-globally-optimal-rigid-registration
Repo
Framework

Sparse Group Inductive Matrix Completion


Title	Sparse Group Inductive Matrix Completion
Authors	Ivan Nazarov, Boris Shirokikh, Maria Burkina, Gennady Fedonin, Maxim Panov
Abstract	We consider the problem of matrix completion with side information (\textit{inductive matrix completion}). In real-world applications many side-channel features are typically non-informative making feature selection an important part of the problem. We incorporate feature selection into inductive matrix completion by proposing a matrix factorization framework with group-lasso regularization on side feature parameter matrices. We demonstrate, that the theoretical sample complexity for the proposed method is much lower compared to its competitors in sparse problems, and propose an efficient optimization algorithm for the resulting low-rank matrix completion problem with sparsifying regularizers. Experiments on synthetic and real-world datasets show that the proposed approach outperforms other methods.
Tasks	Feature Selection, Low-Rank Matrix Completion, Matrix Completion
Published	2018-04-27
URL	http://arxiv.org/abs/1804.10653v2
PDF	http://arxiv.org/pdf/1804.10653v2.pdf
PWC	https://paperswithcode.com/paper/sparse-group-inductive-matrix-completion
Repo
Framework

Sampling Clustering


Title	Sampling Clustering
Authors	Ching Tarn, Yinan Zhang, Ye Feng
Abstract	We propose an efficient linear-time graph-based divisive cluster analysis approach called Sampling Clustering. It constructs a lite informative dendrogram by recursively dividing a graph into subgraphs. In each recursive call, a graph is sampled first with a set of vertices being removed to disconnect latent clusters, then condensed by adding edges to the remaining vertices to avoid graph fragmentation caused by vertex removals. We also present some sampling and condensing methods and discuss the effectiveness in this paper. Our implementations run in linear time and achieve outstanding performance on various types of datasets. Experimental results show that they outperform state-of-the-art clustering algorithms with significantly less computing resource requirements.
Tasks
Published	2018-06-21
URL	https://arxiv.org/abs/1806.08245v2
PDF	https://arxiv.org/pdf/1806.08245v2.pdf
PWC	https://paperswithcode.com/paper/sampling-clustering
Repo
Framework

Multi-feature Fusion for Image Retrieval Using Constrained Dominant Sets


Title	Multi-feature Fusion for Image Retrieval Using Constrained Dominant Sets
Authors	Leulseged Tesfaye Alemu, Marcello Pelillo
Abstract	Aggregating different image features for image retrieval has recently shown its effectiveness. While highly effective, though, the question of how to uplift the impact of the best features for a specific query image persists as an open computer vision problem. In this paper, we propose a computationally efficient approach to fuse several hand-crafted and deep features, based on the probabilistic distribution of a given membership score of a constrained cluster in an unsupervised manner. First, we introduce an incremental nearest neighbor (NN) selection method, whereby we dynamically select k-NN to the query. We then build several graphs from the obtained NN sets and employ constrained dominant sets (CDS) on each graph G to assign edge weights which consider the intrinsic manifold structure of the graph, and detect false matches to the query. Finally, we elaborate the computation of feature positive-impact weight (PIW) based on the dispersive degree of the characteristics vector. To this end, we exploit the entropy of a cluster membership-score distribution. In addition, the final NN set bypasses a heuristic voting scheme. Experiments on several retrieval benchmark datasets show that our method can improve the state-of-the-art result.
Tasks	Image Retrieval
Published	2018-08-15
URL	http://arxiv.org/abs/1808.05075v1
PDF	http://arxiv.org/pdf/1808.05075v1.pdf
PWC	https://paperswithcode.com/paper/multi-feature-fusion-for-image-retrieval
Repo
Framework