Search

In the paper we present an approach to the automatic detection and identification of important elements in paper documents. This includes stamps, logos, printed text blocks, signatures and tables. Presented approach consists of two stages. The first one includes object detection by means of AdaBoost cascade of weak classifiers and Haar-like features. Resulting image blocks are, at the second stage, subjected to verification based on selected features calculated from recently proposed low-level descriptors combined with certain classifiers representing current machine-learning approaches. The training phase, for both stages, uses bootstrapping, i.e., integrative process, aiming at increasing the accuracy. Experiments performed on large set of digitized paper documents showed that adopted strategy is useful and efficient.

Gesture recognition is one of the most popular tasks in computer vision, where convolutional neural networks (CNNs) based method has obtained the state-of-the-art performance. However, it is generally acknowledged that CNNs need a large amount of data to achieve such performance. Event Camera is a kind of biologically inspired event-based camera, which can keep the information of moving objects and remove the redundant background data. In this paper, we set up a rock-paper-scissor hand gesture recognition system based on FlowNet and Event Camera. Event camera is used to acquire event data. Then we propose an algorithm for the proposed gesture recognition. Specifically, FlowNet2.0 is employed to extract the motion representation of the pre-processed visual data, and a CNN classification network is applied to recognize the symbols extracted according to the motion representation. As a comparison, we also apply the rock-paper-scissor gesture recognition algorithm on traditional camera. The experimental results show that the proposed system based on Event Camera gets better performance, and to some extent, weaken the dependence on the training data. The whole system achieves 94.0\(\%\) out-of-sample accuracy and allows computation at up to 30 fps.

Scientific papers are important for scholars to track trends in specific research areas. With the increase in the number of scientific papers, it is difficult for scholars to read all the papers to extract emerging or noteworthy knowledge. Paper modeling can help scholars master the key information in scientific papers, and relation classification (RC) between entity pairs is a major approach to paper modeling. To the best of our knowledge, most of the state-of-the-art RC methods are using entire sentence’s context information as input. However, long sentences have too much noise information, which is useless for classification. In this paper, a flexible context is selected as the input information for convolution neural network (CNN), which greatly reduces the noise. Moreover, we find that entity type is another important feature for RC. Based on these findings, we construct a typical CNN architecture to learn features from raw texts automatically, and use a softmax function to classify the entity pairs. Our experiment on SemEval-2018 task 7 dataset yields a macro-F1 value of 83.91%, ranking first among all participants.

The human visual system uses cast shadows to judge the three-dimensional layout of an object. The purpose of this installation is to demonstrate novel visual illusions of depth and transparency for paper materials, which is induced by the conventional light projection of cast shadow patterns. Thus, this installation focuses on perceptual rather than technical aspects. Illuminating a target object, the spatial vicinity of the object is darkened to produce the visual impression of a shadow of the object. By controlling the blurriness of the cast shadow and/or spatial distance between the object and its shadow, a perceptual change in the layout of the object is induced. The audience can interactively enjoy visual experiences wherein objects and letters on a paper perceptually float in the air. We also demonstrate that it is possible to edit the material appearance of a real object by manipulating the shape of the shadow; an opaque colored paper appear to be a transparent color film floating in the air.

E-commerce customers and general Web users should not believe that the products suggested by Recommender systems or results displayed by Web search engines are either complete or relevant to their search aspirations. The economic priority of Web related businesses requires a higher rank on Web snippets or product suggestions in order to receive additional customers; furthermore, Web search engines and recommender systems revenue is obtained from advertisements and pay-per-click. This survey paper presents a review of Web Search Engines, Ranking Algorithms, Citation Analysis and Recommender Systems. In addition, Neural networks and Deep Learning are also analyzed including their use in learning relevance and ranking. Finally, this survey paper also introduces the Random Neural Network with its practical applications.

Near infrared (NIR) spectroscopy is used in many applications to gather information about chemical composition of materials. For paper waste sorting, a small number of scores computed from NIR-spectra and assuming more or less unimodal clustered data, a pixel classifier can still be crafted by hand using knowledge about chemical properties and a reasonable amount of intuition. Additional information can be gained by visual data (VIS). However, it is not obvious what features, e.g. based on color, saturation, textured areas, are finally important for successfully separating the paper classes in feature space. Hence, a rigorous feature analysis becomes inevitable. We have chosen a generic machine-learning approach to successfully fuse NIR and VIS information. By exploiting a classification tree and a variety of additional visual features, we could increase the recognition rate to 78% for 11 classes, compared to 63% only using NIR scores. A modified feature ranking measure, which takes redundancies of features into account, allows us to analyze the importance of features and reduce them effectively. While some visual features like color saturation and hue showed to be important, some NIR scores could even be dropped. Finally, we generalize this approach to analyze raw NIR-spectra instead of score values and apply it to plastic waste sorting.

International audience; The work presented in this paper deals with the use of ontologies to compare scientific texts. It particularly deals with scientific papers, specifically their abstracts, short texts that are relatively well structured and normally provide enough knowledge to allow a community of readers to assess the content of the associated scientific papers. The problem is, therefore, to determine how to assess the semantic proximity/similarity of two papers by examining their respective abstracts. Given that a domain ontology provides a useful way to represent knowledge relative to a given domain, this work considers ontologies relative to scientific domains. Our process begins by defining the relevant domain for an abstract through an automatic classification that makes it possible to associate this abstract to its relevant scientific domain, chosen from several candidate domains. The content of an abstract is represented in the form of a conceptual graph which is enriched to construct its semantic perimeter. As presented below, this notion of semantic perimeter usefully allows us to assess the similarity between the texts by matching their graphs. Detecting plagiarism is the main application field addressed in this paper, among the many possible application fields of our approach.

Among the tasks to be carried out by conference organizers is the one of assigning reviewers to papers. That problem is known in the literature as the Conference Paper Assignment Problem (CPAP). In this paper we approach the solution of a reasonably rich variant of the CPAP by means of Answer Set Programming (ASP). ASP is an established logic-based programming paradigm which has been successfully applied for solving complex problems arising in Artificial Intelligence. We show how the CPAP can be elegantly encoded by means of an ASP program, and we analyze the results of an experiment, conducted on real-world data, that outlines the viability of our solution.

International audience; This book constitutes the thoroughly refereed proceedings of the First International Conference on Machine Learning for Networking, MLN 2018, held in Paris, France, in November 2018. The 22 revised full papers included in the volume were carefully reviewed and selected from 48 submissions. They present new trends in the following topics: Deep and reinforcement learning; Pattern recognition and classification for networks; Machine learning for network slicing optimization, 5G system, user behavior prediction, multimedia, IoT, security and protection; Optimization and new innovative machine learning methods; Performance analysis of machine learning algorithms; Experimental evaluations of machine learning; Data mining in heterogeneous networks; Distributed and decentralized machine learning algorithms; Intelligent cloud-support communications, resource allocation, energy-aware/green communications, software defined networks, cooperative networks, positioning and navigation systems, wireless communications, wireless sensor networks, underwater sensor networks.

Unsupervised random-walk keyphrase extraction models mainly rely on global structural information of the word graph, with nodes representing candidate words and edges capturing the co-occurrence information between candidate words. However, integrating different types of useful information into the representation learning process to help better extract keyphrases is relatively unexplored. In this paper, we propose a random-walk method to extract keyphrases using word embeddings. Specifically, we first design a new word embedding learning model to integrate local context information of the word graph (i.e., the local word collocation patterns) with some crucial features of candidate words and edges. Then, a novel random-walk ranking model is designed to extract keyphrases by leveraging such word embeddings. Experimental results show that our approach outperforms 8 state-of-the-art unsupervised methods on two real datasets consistently for keyphrase extraction.

The SemEval-2019 Task 12 is toponym resolution in scientific papers. We focus on Subtask 1: Toponym Detection which is the identification of spans of text for place names mentioned in a document. We propose two methods: 1) sliding window convolutional neural network using ELMo embeddings (cnn-elmo), and 2) sliding window multi-Layer perceptron using ELMo embeddings (mlp-elmo). We also submit Bi-lateral LSTM with Conditional Random Fields (bi-LSTM) as a strong baseline given its state-of-art performance in Named Entity Recognition (NER) task. Our best performing model is cnn-elmo with a F1 of 0.844 which was below bi-LSTM F1 of 0.862 when evaluated on overlap macro detection. Eight teams participated in this subtask with a total of 21 submissions.

Sarcasm is often expressed through several verbal and non-verbal cues, e.g., a change of tone, overemphasis in a word, a drawn-out syllable, or a straight looking face. Most of the recent work in sarcasm detection has been carried out on textual data. In this paper, we argue that incorporating multimodal cues can improve the automatic classification of sarcasm. As a first step towards enabling the development of multimodal approaches for sarcasm detection, we propose a new sarcasm dataset, Multimodal Sarcasm Detection Dataset (MUStARD), compiled from popular TV shows. MUStARD consists of audiovisual utterances annotated with sarcasm labels. Each utterance is accompanied by its context of historical utterances in the dialogue, which provides additional information on the scenario where the utterance occurs. Our initial results show that the use of multimodal information can reduce the relative error rate of sarcasm detection by up to 12.9% in F-score when compared to the use of individual modalities. The full dataset is publicly available for use at https://github.com/soujanyaporia/MUStARD, Comment: Accepted at ACL 2019

Multi-view data that contains the data represented in many types of features has received much attention recently. The class of method utilising non-negative matrix factorization (NMF) and manifold learning to seek the meaningful latent structure of data has been popularly used for both traditional data and multi-view data. The NMF and manifold based multi-view clustering methods focus on dealing with the challenges of manifold learning and applying manifold learning on the NMF framework. This paper provides a comprehensive review of this important class of methods on multi-view data. We conduct an extensive experiments on several datasets and raise many open problems that can be dealt with in the future so a higher clustering performance can be achieved.

Paraphrase extraction is the method of collecting words, phrases or sentences that have same meaning from free text or structured documents. Candidate monolingual pararel corpus was build from scientific papers by clustering techniques. The corpus contain couples of citation sentences that mention same index. The corpus evaluted using weighting techniques and proximity measurements based on the density of each sentence cluster. The results of this reserach are expected to simplify paraphrase labeling by language and computer science experts.

Fully convolutional neural networks have shown remarkable success in performing semantic segmentation. The use of convolutional layers for the entire architecture and skip connections to combine different resolution features or predictions have been adopted in successful networks, such as U-Net and DenseNet. However, these models employ several max-pooling layers that cause the network to lose spatial information and require them to mimic an autoencoder architecture to perform semantic segmentation at the original input resolution. In this paper, we propose a network that extracts features automatically with convolutional layers, like the fully convolutional neural network, but retains the spatial information of each of the extracted features. It then utilises the extracted features to make predictions with an efficient upsampling method. We evaluate the network performance on a liver segmentation task where it performs with comparable accuracy to other state-of-the-art networks while being much smaller in terms of the number of parameters as well as faster in computation time.

Agriculture plays an important role in farmer’s life. Sometimes manual identification of disease is time consuming and need of labor is more. One of the most important facts that reduce the growth of plants is disease attack. Overall study about agriculture shows that quality and quantity of agricultural products may be reduced due to various factors of plant diseases. These diseases can be more easily identified by using machine learning approach as compared to manual method. Hence machine learning method can be used to identify the affected leaf images.Images captured by camera will be processed using different image processing technique. These techniques will help in identifying plant diseases thereby increasing the yield of plants. This survey paper describes plant disease identification using Machine Learning Approach and study in detail about various techniques for disease identification and classification is also done.

A code clone is a pair of code fragments, within or between software systems that are similar. Since code clones often negatively impact the maintainability of a software system, a great many numbers of code clone detection techniques and tools have been proposed and studied over the last decade. To detect all possible similar source code patterns in general, the clone detection tools work on syntax level (such as texts, tokens, AST and so on) while lacking user-specific preferences. This often means the reported clones must be manually validated prior to any analysis in order to filter out the true positive clones from task or user-specific considerations. This manual clone validation effort is very time-consuming and often error-prone, in particular for large-scale clone detection. In this paper, we propose a machine learning based approach for automating the validation process. In an experiment with clones detected by several clone detectors in several different software systems, we found our approach has an accuracy of up to 87.4% when compared against the manual validation by multiple expert judges. The proposed method shows promising results in several comparative studies with the existing related approaches for automatic code clone validation. We also present our experimental results in terms of different code clone detection tools, machine learning algorithms and open source software systems.

Determining the programming language of a source code file has been considered in the research community; it has been shown that Machine Learning (ML) and Natural Language Processing (NLP) algorithms can be effective in identifying the programming language of source code files. However, determining the programming language of a code snippet or a few lines of source code is still a challenging task. Online forums such as Stack Overflow and code repositories such as GitHub contain a large number of code snippets. In this paper, we describe Source Code Classification (SCC), a classifier that can identify the programming language of code snippets written in 21 different programming languages. A Multinomial Naive Bayes (MNB) classifier is employed which is trained using Stack Overflow posts. It is shown to achieve an accuracy of 75% which is higher than that with Programming Languages Identification (PLI-a proprietary online classifier of snippets) whose accuracy is only 55.5%. The average score for precision, recall and the F1 score with the proposed tool are 0.76, 0.75 and 0.75, respectively. In addition, it can distinguish between code snippets from a family of programming languages such as C, C++ and C#, and can also identify the programming language version such as C# 3.0, C# 4.0 and C# 5.0.

In today's open source era, developers look forsimilar software applications in source code repositories for anumber of reasons, including, exploring alternative implementations, reusing source code, or looking for a better application. However, while there are a great many studies for finding similarapplications written in the same programming language, there isa marked lack of studies for finding similar software applicationswritten in different languages. In this paper, we fill the gapby proposing a novel modelCroLSimwhich is able to detectsimilar software applications across different programming lan-guages. In our approach, we use the API documentation tofind relationships among the API calls used by the differentprogramming languages. We adopt a deep learning based word-vector learning method to identify semantic relationships amongthe API documentation which we then use to detect cross-language similar software applications. For evaluating CroLSim, we formed a repository consisting of 8,956 Java, 7,658 C#, and 10,232 Python applications collected from GitHub. Weobserved thatCroLSimcan successfully detect similar softwareapplications across different programming languages with a meanaverage precision rate of 0.65, an average confidence rate of3.6 (out of 5) with 75% high rated successful queries, whichoutperforms all related existing approaches with a significantperformance improvement.

Many robotic tasks in small manufacturing sites are quite simple. For example, a pick and place task requires only a few common commands. Unfortunately, the standard languages and programming environments for industrial robots are complex, making even these simple tasks nearly impossible for novices. To enable novices to program simple tasks we created a block-based programming language and environment focused on usability, learnability, and understandability and embedded its programming environment in a state-of-the-art robot simulator. By using this high-fidelity prototype over the course of a year in a case study, a user study, and for countless demonstrations we have gained many concrete insights. In this paper we discuss the details of the language, the design of its programming environment, and concrete insights gained via longitudinal usage.

The problem of classifying a research article into one or more fields of science is of particular importance for the academic search engines and digital libraries. A robust classification algorithm offers the users a wide variety of useful tools, such as the refinement of their search results, the browsing of articles by category, the recommendation of other similar articles, etc. In the current literature we encounter approaches which attempt to address this problem without taking into consideration important parameters such as the previous history of the authors and the categorization of the scientific journals which publish the articles. In addition, the existing works overlook the huge volume of the involved academic data. In this paper, we expand an existing effective algorithm for research articles classification, and we parallelize it on Apache Spark –a parallelization framework which is capable of sharing large amounts of data into the main memory of the nodes of a cluster– to enable the processing of large academic datasets. Furthermore, we present data manipulation methodologies which are useful not only for this particular problem, but also for most parallel machine learning approaches. In our experimental evaluation, we demonstrate that our proposed algorithm is considerably more accurate than the supervised learning approaches implemented within the machine learning library of Spark, whereas it outperforms them in terms of execution speed by a significant margin.

Prior work on texture analysis of historic, photographic papers has focused primarily on measures of texture similarity. However, automated grouping or clustering of photographic paper textures in a way that is meaningful to art conservators remains an open problem. In this work a deep learning approach to automated classification is presented, for clusters derived from a human sorting experiment conducted by 19 art conservators and paper experts and subsequently extended through crowd-sourcing. The proposed approach uses a deep convolutional neural network, and results are presented on the performance in automatically classifying images when compared to human experts.

Machine learning is widely used in malware detection systems as a core component. However, machine learning algorithm is based on the assumption that the underlying malware concept is stable for training and testing. The assumption is vulnerable to well-crafted concept drift attacks, such as mimicry attacks, gradient descent attacks, poisoning attacks and so on. This paper proposes an ensemble learning system which combines vertical and horizontal correlation learning models. The significant diversity among vertical and horizontal correlation models increases the difficulty of concept drift attacks. And average p-value assessment is applied to fortify the system to be sensitive to hidden concept drift. The experiment results show that the hybrid system could actively recognize the concept drift among different Miuref variants.

We present the SemEval-2019 Task 12 which focuses on toponym resolution in scientific articles. Given an article from PubMed, the task consists of detecting mentions of names of places, or toponyms, and mapping the mentions to their corresponding entries in GeoNames.org, a database of geospatial locations. We proposed three subtasks. In Subtask 1, we asked participants to detect all toponyms in an article. In Subtask 2, given toponym mentions as input, we asked participants to disambiguate them by linking them to entries in GeoNames. In Subtask 3, we asked participants to perform both the detection and the disambiguation steps for all toponyms. A total of 29 teams registered, and 8 teams submitted a system run. We summarize the corpus and the tools created for the challenge. They are freely available at https://competitions.codalab.org/competitions/19948. We also analyze the methods, the results and the errors made by the competing systems with a focus on toponym disambiguation.

With today’s eye-tracking technologies, it is possible to analyze cognitive processes such as reading. There are many studies on that topic. We have studied some of these studies that analyze eye movements recording during reading by various machine learning algorithms.

Any text can be presented as an infinite number of topic models.Any of these topics would not have any attributes that would make it possible to break them up into classes. The Author has suggested additive regularization when creating models to single out topic clusters from the Probabilistic Latent Semantic Analysis, PLSA.The method proposed by the Author allows singling out topic classes based on their density in the document-topic space (Matrix Θ) for a selected collection of documents.Such segmentation is similar to hierarchical variant LDA (HLDA) but has no such drawback as that the LDA models have.

Recycling wastewater of the pulping and paper-making industry are widely considered for clean production, which heavily rely on the timely and accurate monitoring in paper-making wastewater treatment processes. A novel predicting model based on genetic-deep belief network algorithm was proposed to improve the predictive accuracy and reliability for process monitoring. Considering the deep belief networks (DBN) as a deep learning model is aiming to describe the relationship among variables in a complex process modeling, genetic algorithm (GA) was employed to reduce the input variables dimensionality, simplify the network structure and overcome the dynamic characteristic difficulties of process data in monitoring. Compared with DBN and back propagation neural network (BPNN), the GA-DBN effectively achieved a better predictive accuracy than other tests models in complex wastewater treatment processes. The value of the coefficient of determination of GA-DBN model is increased by 1.71–1.86% and 5.21–9.32%, respectively. The GA-DBN model can be structured with fewer variables or samples to describe the complex paper-making wastewater treatment process, obtaining the better model performance and predictive accuracy.

Video content is used extensively in many fields. However, in some fields, video manipulation techniques are required to improve the human-friendliness of such content. In this paper, we propose a method that automatically generates animations in the style of colored paper mosaics, to create human-friendly, artistic imagery. To enhance temporal coherence while maintaining the characteristics of colored paper mosaics, we also propose a particle video-based method that determines coherent locations for tiles in animations. The proposed method generates evenly distributed particles, which are used to produce animated tiles via our tile modeling process.

International audience; The task of tagging papers with semantic metadata in order to analyze their relatedness represents a good foundation for a paper recommender system. The analysis from this paper extends from previous research in order to create a graph of papers from a specific domain with the purpose of determining each article's importance within the considered corpus of papers. Moreover, as non-latent representations are powerful when used in conjunction with latent ones, our system retrieves semantically close words, not present in the paper, in order to improve the retrieval of papers. Our previous analyses used the semantic representation of papers in different semantic models with the purpose of creating visual graphs based on the semantic relatedness links between the abstracts. The current analysis takes a step forward by proposing a model that can suggest which papers are of the highest relevance, share similar concepts, and are semantically related with the initial query. Our study is performed using paper abstracts in the field of information technology extracted from the Web of Science citation index. The research includes a use case and its corresponding results by using interactive and exploratory network graph representations.

Image inpainting has been widely investigated to remove undesired parts of images. One of the effective approaches is exemplar-based inpainting, which uses texture patterns in an image as exemplars for filling in missing regions. As one of exemplar-based methods, an inpainting method based on automatic perspective correction using vanishing points was recently proposed. However, the target scene is limited to artificial one in which vanishing points are easily detectable. Although some other methods for automatic perspective correction have also been proposed, the effect on image inpainting has not been evaluated in detail so far. This paper analyzes the effect of multiple methods for automatic perspective correction on image inpainting by developing a method that combines a variety of automatic perspective correction methods and image inpainting. Specifically, we examine the influence of the amount of perspective distortion and characteristics of textures on image inpainting results by using images distorted by simulation. We also examine the effect using real images. In addition, we demonstrate the advantage of employing multiple criteria for perspective correction over the conventional method from the analyzed results.

This paper provides an extensive study on the availability of image representations based on convolutional networks (ConvNets) for the task of visual instance retrieval. Besides the choice of convo ...

Paper manufacturing is energy demanding and improved modelling of the pulp bleach process is the main non-invasive means of reducing energy costs. In this paper, time it takes to bleach paper pulp to desired brightness is examined. The model currently used is analysed and benchmarked against two machine learning models Random Forest and TreeBoost. Results suggests that the current model can be superseded by the machine learning models and it does not use the optimal compact subset of features. Despite the differences between the machine learning models, a feature ranking correlation has been observed for the new models. One novel, yet unused, feature that both machine learning models found to be important is the concentration of bleach agent.

As more and more academic papers are being submitted to conferences and journals, evaluating all these papers by professionals is time-consuming and can cause inequality due to the personal factors of the reviewers. In this paper, in order to assist professionals in evaluating academic papers, we propose a novel task: automatic academic paper rating (AAPR), which automatically determine whether to accept academic papers. We build a new dataset for this task and propose a novel modularized hierarchical convolutional neural network to achieve automatic academic paper rating. Evaluation results show that the proposed model outperforms the baselines by a large margin. The dataset and code are available at \url{https://github.com/lancopku/AAPR}, Accepted by ACL2018

We present a printed paper actuator as a low cost, reversible and electrical actuation and sensing method. This is a novel but easily accessible enabling technology that expands upon the library of actuation-sensing materials in HCI. By integrating three physical phenomena, including the bilayer bending actuation, the shape memory effect of the thermoplastic and the current-driven joule heating via conductive printing filament, we developed the actuator by simply printing a single layer conductive Polylactide (PLA) on a piece of copy paper via a desktop fused deposition modeling (FDM) 3D printer. This paper describes the fabrication process, the material mechanism, and the transformation primitives, followed by the electronic sensing and control methods. A software tool that assists the design, simulation and printing toolpath generation is introduced. Finally, we explored applications under four contexts: robotics, interactive art, entertainment and home environment.

Image mosaicing is one of the most important subjects of research in computer vision at current. Image mocaicing requires the integration of direct techniques and feature based techniques. Direct techniques are found to be very useful for mosaicing large overlapping regions, small translations and rotations while feature based techniques are useful for small overlapping regions. Feature based image mosaicing is a combination of corner detection, corner matching, motion parameters estimation and image stitching. Furthermore, image mosaicing is considered the process of obtaining a wider field-of-view of a scene from a sequence of partial views, which has been an attractive research area because of its wide range of applications, including motion detection, resolution enhancement, monitoring global land usage, and medical imaging. Numerous algorithms for image mosaicing have been proposed over the last two decades. In this paper the authors present a review on different approaches for image mosaicing and the literature over the past few years in the field of image masaicing methodologies. The authors take an overview on the various methods for image mosaicing. This review paper also provides an in depth survey of the existing image mosaicing algorithms by classifying them into several groups. For each group, the fundamental concepts are first clearly explained. Finally this paper also reviews and discusses the strength and weaknesses of all the mosaicing groups.

The colorimetric method is usually used to test the concentration of substances. However, this method has a big error since different people have different sensitivities to colors. In this paper, in order to solve the identification problem of the color and the concentration of the test paper, firstly, we found out that the concentration of substance is correlated with the color reading by using the Pearson’s Chi-squared test method. And by the concentration coefficient of Pearson correlation analysis, the concentration of substance and color reading is highly correlated. Secondly, according to the RGB value of the paper image, the color moments of the image are calculated as the characteristics of the image, and the Levenberg–Marquardt (LM) neural network is established to classify the concentration of the substance. The accuracy of the training set model is 94.5%, and the accuracy of the test set model is 87.5%. The model precision is high, and the model has stronger generalization ability. Therefore, according to the RGB value of the test paper image, it is effective to establish the LM neural network model to identify the substance concentration.

We present RusNLP, a web service implementing semantic search engine and recommendation system over proceedings of three major Russian NLP conferences (Dialogue, AIST and AINL). The collected corpus spans across 12 years and contains about 400 academic papers in English. The presented web service allows searching for publications semantically similar to arbitrary user queries or to any given paper. Search results can be filtered by authors and their affiliations, conferences or years. They are also interlinked with the NLPub.ru service, making it easier to quickly capture the general focus of each paper. The search engine source code and the publications metadata are freely available for all interested researchers.

Thousands of academic papers are submitted at top venues each year. Manual audits are time-consuming and laborious. And the result may be influenced by human factors. This paper investigates a modularized and attention-based recurrent convolutional network model to represent academic paper and predict aspect scores. This model treats input text as module-document hierarchies, uses attention pooling CNN and LSTM to represent text, and outputs prediction with a linear layer. Empirical results on PeerRead data show that this model give the best performance among the baseline models.

Due to the recent boom in artificial intelligence (AI) research, including computer vision (CV), it has become impossible for researchers in these fields to keep up with the exponentially increasing number of manuscripts. In response to this situation, this paper proposes the paper summary generation (PSG) task using a simple but effective method to automatically generate an academic paper summary from raw PDF data. We realized PSG by combination of vision-based supervised components detector and language-based unsupervised important sentence extractor, which is applicable for a trained format of manuscripts. We show the quantitative evaluation of ability of simple vision-based components extraction, and the qualitative evaluation that our system can extract both visual item and sentence that are helpful for understanding. After processing via our PSG, the 979 manuscripts accepted by the Conference on Computer Vision and Pattern Recognition (CVPR) 2018 are available. It is believed that the proposed method will provide a better way for researchers to stay caught with important academic papers., Comment: International Conference on Machine Vision 2018, Munich, Germany