Paper Group ANR 110
Set2Model Networks: Learning Discriminatively To Learn Generative Models. Short Communication on QUIST: A Quick Clustering Algorithm. A Minimalistic Approach to Sum-Product Network Learning for Real Applications. Analysis of proposed PDE-based underwater image enhancement algorithms. Web Spam Detection Using Multiple Kernels in Twin Support Vector …
Set2Model Networks: Learning Discriminatively To Learn Generative Models
| Title | Set2Model Networks: Learning Discriminatively To Learn Generative Models |
| Authors | A. Vakhitov, A. Kuzmin, V. Lempitsky |
| Abstract | We present a new “learning-to-learn”-type approach that enables rapid learning of concepts from small-to-medium sized training sets and is primarily designed for web-initialized image retrieval. At the core of our approach is a deep architecture (a Set2Model network) that maps sets of examples to simple generative probabilistic models such as Gaussians or mixtures of Gaussians in the space of high-dimensional descriptors. The parameters of the embedding into the descriptor space are trained in the end-to-end fashion in the meta-learning stage using a set of training learning problems. The main technical novelty of our approach is the derivation of the backprop process through the mixture model fitting, which makes the likelihood of the resulting models differentiable with respect to the positions of the input descriptors. While the meta-learning process for a Set2Model network is discriminative, a trained Set2Model network performs generative learning of generative models in the descriptor space, which facilitates learning in the cases when no negative examples are available, and whenever the concept being learned is polysemous or represented by noisy training sets. Among other experiments, we demonstrate that these properties allow Set2Model networks to pick visual concepts from the raw outputs of Internet image search engines better than a set of strong baselines. |
| Tasks | Image Retrieval, Meta-Learning |
| Published | 2016-12-22 |
| URL | http://arxiv.org/abs/1612.07697v2 |
| http://arxiv.org/pdf/1612.07697v2.pdf | |
| PWC | https://paperswithcode.com/paper/set2model-networks-learning-discriminatively |
| Repo | |
| Framework | |
Short Communication on QUIST: A Quick Clustering Algorithm
| Title | Short Communication on QUIST: A Quick Clustering Algorithm |
| Authors | Sherenaz W. Al-Haj Baddar |
| Abstract | In this short communication we introduce the quick clustering algorithm (QUIST), an efficient hierarchical clustering algorithm based on sorting. QUIST is a poly-logarithmic divisive clustering algorithm that does not assume the number of clusters, and/or the cluster size to be known ahead of time. It is also insensitive to the original ordering of the input. |
| Tasks | |
| Published | 2016-06-01 |
| URL | http://arxiv.org/abs/1606.00398v1 |
| http://arxiv.org/pdf/1606.00398v1.pdf | |
| PWC | https://paperswithcode.com/paper/short-communication-on-quist-a-quick |
| Repo | |
| Framework | |
A Minimalistic Approach to Sum-Product Network Learning for Real Applications
| Title | A Minimalistic Approach to Sum-Product Network Learning for Real Applications |
| Authors | Viktoriya Krakovna, Moshe Looks |
| Abstract | Sum-Product Networks (SPNs) are a class of expressive yet tractable hierarchical graphical models. LearnSPN is a structure learning algorithm for SPNs that uses hierarchical co-clustering to simultaneously identifying similar entities and similar features. The original LearnSPN algorithm assumes that all the variables are discrete and there is no missing data. We introduce a practical, simplified version of LearnSPN, MiniSPN, that runs faster and can handle missing data and heterogeneous features common in real applications. We demonstrate the performance of MiniSPN on standard benchmark datasets and on two datasets from Google’s Knowledge Graph exhibiting high missingness rates and a mix of discrete and continuous features. |
| Tasks | |
| Published | 2016-02-12 |
| URL | http://arxiv.org/abs/1602.04259v3 |
| http://arxiv.org/pdf/1602.04259v3.pdf | |
| PWC | https://paperswithcode.com/paper/a-minimalistic-approach-to-sum-product |
| Repo | |
| Framework | |
Analysis of proposed PDE-based underwater image enhancement algorithms
| Title | Analysis of proposed PDE-based underwater image enhancement algorithms |
| Authors | U. A. Nnolim |
| Abstract | This report describes the experimental analysis of proposed underwater image enhancement algorithms based on partial differential equations (PDEs). The algorithms perform simultaneous smoothing and enhancement due to the combination of both processes within the PDE-formulation. The framework enables the incorporation of suitable colour and contrast enhancement algorithms within one unified functional. Additional modification of the formulation includes the combination of the popular Contrast Limited Adaptive Histogram Equalization (CLAHE) with the proposed approach. This modification enables the hybrid algorithm to provide both local enhancement (due to the CLAHE) and global enhancement (due to the proposed contrast term). Additionally, the CLAHE clip limit parameter is computed dynamically in each iteration and used to gauge the amount of local enhancement performed by the CLAHE within the formulation. This enables the algorithm to reduce or prevent the enhancement of noisy artifacts, which if present, are also smoothed out by the anisotropic diffusion term within the PDE formulation. In other words, the modified algorithm combines the strength of the CLAHE, AD and the contrast term while minimizing their weaknesses. Ultimately, the system is optimized using image data metrics for automated enhancement and compromise between visual and quantitative results. Experiments indicate that the proposed algorithms perform a series of functions such as illumination correction, colour enhancement correction and restoration, contrast enhancement and noise suppression. Moreover, the proposed approaches surpass most other conventional algorithms found in the literature. |
| Tasks | Image Enhancement |
| Published | 2016-12-14 |
| URL | http://arxiv.org/abs/1612.04447v1 |
| http://arxiv.org/pdf/1612.04447v1.pdf | |
| PWC | https://paperswithcode.com/paper/analysis-of-proposed-pde-based-underwater |
| Repo | |
| Framework | |
Web Spam Detection Using Multiple Kernels in Twin Support Vector Machine
| Title | Web Spam Detection Using Multiple Kernels in Twin Support Vector Machine |
| Authors | Seyed Hamid Reza Mohammadi, Mohammad Ali Zare Chahooki |
| Abstract | Search engines are the most important tools for web data acquisition. Web pages are crawled and indexed by search Engines. Users typically locate useful web pages by querying a search engine. One of the challenges in search engines administration is spam pages which waste search engine resources. These pages by deception of search engine ranking algorithms try to be showed in the first page of results. There are many approaches to web spam pages detection such as measurement of HTML code style similarity, pages linguistic pattern analysis and machine learning algorithm on page content features. One of the famous algorithms has been used in machine learning approach is Support Vector Machine (SVM) classifier. Recently basic structure of SVM has been changed by new extensions to increase robustness and classification accuracy. In this paper we improved accuracy of web spam detection by using two nonlinear kernels into Twin SVM (TSVM) as an improved extension of SVM. The classifier ability to data separation has been increased by using two separated kernels for each class of data. Effectiveness of new proposed method has been experimented with two publicly used spam datasets called UK-2007 and UK-2006. Results show the effectiveness of proposed kernelized version of TSVM in web spam page detection. |
| Tasks | |
| Published | 2016-05-10 |
| URL | http://arxiv.org/abs/1605.02917v1 |
| http://arxiv.org/pdf/1605.02917v1.pdf | |
| PWC | https://paperswithcode.com/paper/web-spam-detection-using-multiple-kernels-in |
| Repo | |
| Framework | |
Hierarchical compositional feature learning
| Title | Hierarchical compositional feature learning |
| Authors | Miguel Lázaro-Gredilla, Yi Liu, D. Scott Phoenix, Dileep George |
| Abstract | We introduce the hierarchical compositional network (HCN), a directed generative model able to discover and disentangle, without supervision, the building blocks of a set of binary images. The building blocks are binary features defined hierarchically as a composition of some of the features in the layer immediately below, arranged in a particular manner. At a high level, HCN is similar to a sigmoid belief network with pooling. Inference and learning in HCN are very challenging and existing variational approximations do not work satisfactorily. A main contribution of this work is to show that both can be addressed using max-product message passing (MPMP) with a particular schedule (no EM required). Also, using MPMP as an inference engine for HCN makes new tasks simple: adding supervision information, classifying images, or performing inpainting all correspond to clamping some variables of the model to their known values and running MPMP on the rest. When used for classification, fast inference with HCN has exactly the same functional form as a convolutional neural network (CNN) with linear activations and binary weights. However, HCN’s features are qualitatively very different. |
| Tasks | |
| Published | 2016-11-07 |
| URL | http://arxiv.org/abs/1611.02252v2 |
| http://arxiv.org/pdf/1611.02252v2.pdf | |
| PWC | https://paperswithcode.com/paper/hierarchical-compositional-feature-learning |
| Repo | |
| Framework | |
Ranking academic institutions on potential paper acceptance in upcoming conferences
| Title | Ranking academic institutions on potential paper acceptance in upcoming conferences |
| Authors | Jobin Wilson, Ram Mohan, Muhammad Arif, Santanu Chaudhury, Brejesh Lall |
| Abstract | The crux of the problem in KDD Cup 2016 involves developing data mining techniques to rank research institutions based on publications. Rank importance of research institutions are derived from predictions on the number of full research papers that would potentially get accepted in upcoming top-tier conferences, utilizing public information on the web. This paper describes our solution to KDD Cup 2016. We used a two step approach in which we first identify full research papers corresponding to each conference of interest and then train two variants of exponential smoothing models to make predictions. Our solution achieves an overall score of 0.7508, while the winning submission scored 0.7656 in the overall results. |
| Tasks | |
| Published | 2016-10-10 |
| URL | http://arxiv.org/abs/1610.02828v1 |
| http://arxiv.org/pdf/1610.02828v1.pdf | |
| PWC | https://paperswithcode.com/paper/ranking-academic-institutions-on-potential |
| Repo | |
| Framework | |
Tie-breaker: Using language models to quantify gender bias in sports journalism
| Title | Tie-breaker: Using language models to quantify gender bias in sports journalism |
| Authors | Liye Fu, Cristian Danescu-Niculescu-Mizil, Lillian Lee |
| Abstract | Gender bias is an increasingly important issue in sports journalism. In this work, we propose a language-model-based approach to quantify differences in questions posed to female vs. male athletes, and apply it to tennis post-match interviews. We find that journalists ask male players questions that are generally more focused on the game when compared with the questions they ask their female counterparts. We also provide a fine-grained analysis of the extent to which the salience of this bias depends on various factors, such as question type, game outcome or player rank. |
| Tasks | Language Modelling |
| Published | 2016-07-13 |
| URL | http://arxiv.org/abs/1607.03895v1 |
| http://arxiv.org/pdf/1607.03895v1.pdf | |
| PWC | https://paperswithcode.com/paper/tie-breaker-using-language-models-to-quantify |
| Repo | |
| Framework | |
Procedural Generation of Angry Birds Levels using Building Constructive Grammar with Chinese-Style and/or Japanese-Style Models
| Title | Procedural Generation of Angry Birds Levels using Building Constructive Grammar with Chinese-Style and/or Japanese-Style Models |
| Authors | YuXuan Jiang, Misaki Kaidan, Chun Yin Chu, Tomohiro Harada, Ruck Thawonmas |
| Abstract | This paper presents a procedural generation method that creates visually attractive levels for the Angry Birds game. Besides being an immensely popular mobile game, Angry Birds has recently become a test bed for various artificial intelligence technologies. We propose a new approach for procedurally generating Angry Birds levels using Chinese style and Japanese style building structures. A conducted experiment confirms the effectiveness of our approach with statistical significance. |
| Tasks | |
| Published | 2016-04-27 |
| URL | http://arxiv.org/abs/1604.07906v1 |
| http://arxiv.org/pdf/1604.07906v1.pdf | |
| PWC | https://paperswithcode.com/paper/procedural-generation-of-angry-birds-levels |
| Repo | |
| Framework | |
Identity Matters in Deep Learning
| Title | Identity Matters in Deep Learning |
| Authors | Moritz Hardt, Tengyu Ma |
| Abstract | An emerging design principle in deep learning is that each layer of a deep artificial neural network should be able to easily express the identity transformation. This idea not only motivated various normalization techniques, such as \emph{batch normalization}, but was also key to the immense success of \emph{residual networks}. In this work, we put the principle of \emph{identity parameterization} on a more solid theoretical footing alongside further empirical progress. We first give a strikingly simple proof that arbitrarily deep linear residual networks have no spurious local optima. The same result for linear feed-forward networks in their standard parameterization is substantially more delicate. Second, we show that residual networks with ReLu activations have universal finite-sample expressivity in the sense that the network can represent any function of its sample provided that the model has more parameters than the sample size. Directly inspired by our theory, we experiment with a radically simple residual architecture consisting of only residual convolutional layers and ReLu activations, but no batch normalization, dropout, or max pool. Our model improves significantly on previous all-convolutional networks on the CIFAR10, CIFAR100, and ImageNet classification benchmarks. |
| Tasks | |
| Published | 2016-11-14 |
| URL | http://arxiv.org/abs/1611.04231v3 |
| http://arxiv.org/pdf/1611.04231v3.pdf | |
| PWC | https://paperswithcode.com/paper/identity-matters-in-deep-learning |
| Repo | |
| Framework | |
Tunable Online MUS/MSS Enumeration
| Title | Tunable Online MUS/MSS Enumeration |
| Authors | Jaroslav Bendik, Nikola Benes, Ivana Cerna, Jiri Barnat |
| Abstract | In various areas of computer science, the problem of dealing with a set of constraints arises. If the set of constraints is unsatisfiable, one may ask for a minimal description of the reason for this unsatisifi- ability. Minimal unsatisifable subsets (MUSes) and maximal satisifiable subsets (MSSes) are two kinds of such minimal descriptions. The goal of this work is the enumeration of MUSes and MSSes for a given constraint system. As such full enumeration may be intractable in general, we focus on building an online algorithm, which produces MUSes/MSSes in an on-the-fly manner as soon as they are discovered. The problem has been studied before even in its online version. However, our algorithm uses a novel approach that is able to outperform current state-of-the art algorithms for online MUS/MSS enumeration. Moreover, the performance of our algorithm can be adjusted using tunable parameters. We evaluate the algorithm on a set of benchmarks. |
| Tasks | |
| Published | 2016-06-10 |
| URL | http://arxiv.org/abs/1606.03289v1 |
| http://arxiv.org/pdf/1606.03289v1.pdf | |
| PWC | https://paperswithcode.com/paper/tunable-online-musmss-enumeration |
| Repo | |
| Framework | |
Finding Low-Rank Solutions via Non-Convex Matrix Factorization, Efficiently and Provably
| Title | Finding Low-Rank Solutions via Non-Convex Matrix Factorization, Efficiently and Provably |
| Authors | Dohyung Park, Anastasios Kyrillidis, Constantine Caramanis, Sujay Sanghavi |
| Abstract | A rank-$r$ matrix $X \in \mathbb{R}^{m \times n}$ can be written as a product $U V^\top$, where $U \in \mathbb{R}^{m \times r}$ and $V \in \mathbb{R}^{n \times r}$. One could exploit this observation in optimization: e.g., consider the minimization of a convex function $f(X)$ over rank-$r$ matrices, where the set of rank-$r$ matrices is modeled via the factorization $UV^\top$. Though such parameterization reduces the number of variables, and is more computationally efficient (of particular interest is the case $r \ll \min{m, n}$), it comes at a cost: $f(UV^\top)$ becomes a non-convex function w.r.t. $U$ and $V$. We study such parameterization for optimization of generic convex objectives $f$, and focus on first-order, gradient descent algorithmic solutions. We propose the Bi-Factored Gradient Descent (BFGD) algorithm, an efficient first-order method that operates on the $U, V$ factors. We show that when $f$ is (restricted) smooth, BFGD has local sublinear convergence, and linear convergence when $f$ is both (restricted) smooth and (restricted) strongly convex. For several key applications, we provide simple and efficient initialization schemes that provide approximate solutions good enough for the above convergence results to hold. |
| Tasks | |
| Published | 2016-06-10 |
| URL | http://arxiv.org/abs/1606.03168v3 |
| http://arxiv.org/pdf/1606.03168v3.pdf | |
| PWC | https://paperswithcode.com/paper/finding-low-rank-solutions-via-non-convex |
| Repo | |
| Framework | |
Track and Transfer: Watching Videos to Simulate Strong Human Supervision for Weakly-Supervised Object Detection
| Title | Track and Transfer: Watching Videos to Simulate Strong Human Supervision for Weakly-Supervised Object Detection |
| Authors | Krishna Kumar Singh, Fanyi Xiao, Yong Jae Lee |
| Abstract | The status quo approach to training object detectors requires expensive bounding box annotations. Our framework takes a markedly different direction: we transfer tracked object boxes from weakly-labeled videos to weakly-labeled images to automatically generate pseudo ground-truth boxes, which replace manually annotated bounding boxes. We first mine discriminative regions in the weakly-labeled image collection that frequently/rarely appear in the positive/negative images. We then match those regions to videos and retrieve the corresponding tracked object boxes. Finally, we design a hough transform algorithm to vote for the best box to serve as the pseudo GT for each image, and use them to train an object detector. Together, these lead to state-of-the-art weakly-supervised detection results on the PASCAL 2007 and 2010 datasets. |
| Tasks | Object Detection, Weakly Supervised Object Detection |
| Published | 2016-04-19 |
| URL | http://arxiv.org/abs/1604.05766v1 |
| http://arxiv.org/pdf/1604.05766v1.pdf | |
| PWC | https://paperswithcode.com/paper/track-and-transfer-watching-videos-to |
| Repo | |
| Framework | |
A statistical learning strategy for closed-loop control of fluid flows
| Title | A statistical learning strategy for closed-loop control of fluid flows |
| Authors | Florimond Guéniat, Lionel Mathelin, M. Yousuff Hussaini |
| Abstract | This work discusses a closed-loop control strategy for complex systems utilizing scarce and streaming data. A discrete embedding space is first built using hash functions applied to the sensor measurements from which a Markov process model is derived, approximating the complex system’s dynamics. A control strategy is then learned using reinforcement learning once rewards relevant with respect to the control objective are identified. This method is designed for experimental configurations, requiring no computations nor prior knowledge of the system, and enjoys intrinsic robustness. It is illustrated on two systems: the control of the transitions of a Lorenz 63 dynamical system, and the control of the drag of a cylinder flow. The method is shown to perform well. |
| Tasks | |
| Published | 2016-04-11 |
| URL | http://arxiv.org/abs/1604.03392v1 |
| http://arxiv.org/pdf/1604.03392v1.pdf | |
| PWC | https://paperswithcode.com/paper/a-statistical-learning-strategy-for-closed |
| Repo | |
| Framework | |
On the crucial impact of the coupling projector-backprojector in iterative tomographic reconstruction
| Title | On the crucial impact of the coupling projector-backprojector in iterative tomographic reconstruction |
| Authors | Filippo Arcadu, Marco Stampanoni, Federica Marone |
| Abstract | The performance of an iterative reconstruction algorithm for X-ray tomography is strongly determined by the features of the used forward and backprojector. For this reason, a large number of studies has focused on the to design of projectors with increasingly higher accuracy and speed. To what extent the accuracy of an iterative algorithm is affected by the mathematical affinity and the similarity between the actual implementation of the forward and backprojection, referred here as “coupling projector-backprojector”, has been an overlooked aspect so far. The experimental study presented here shows that the reconstruction quality and the convergence of an iterative algorithm greatly rely on a good matching between the implementation of the tomographic operators. In comparison, other aspects like the accuracy of the standalone operators, the usage of physical constraints or the choice of stopping criteria may even play a less relevant role. |
| Tasks | |
| Published | 2016-12-16 |
| URL | http://arxiv.org/abs/1612.05515v1 |
| http://arxiv.org/pdf/1612.05515v1.pdf | |
| PWC | https://paperswithcode.com/paper/on-the-crucial-impact-of-the-coupling |
| Repo | |
| Framework | |
