Search

Lotic fish species distributions are frequently predicted using remotely sensed habitat variables that characterize the adjacent landscape and serve as proxies for instream habitat. Recent advancements in statistical methodology, however, allow for leveraging fish assemblage data when predicting distributions. This is important because assemblage composition likely provides better information about instream habitat compared to landscape-derived metrics and therefore may improve predictions. To better understand the value of using multi-species fish data in species distribution modeling, we fit two conditional random fields (CRF) models to quantify the relative importance of fish assemblage co-occurrence, landscape-derived habitat variables, and interactions between these two predictor groups (i.e., effects of co-occurrence could be context-dependent) at over 1200 stream catchments in Pennsylvania, USA. We first compared predictive performance of CRF models against traditionally used single-species logistic regressions (generalized linear models; GLMs) and found that inclusion of fish assemblage data often improved predictive performance. The multi-species CRF models performed significantly better at predicting occurrence for 63% of species with an average percent increase in AUC of 25% compared to GLMs. Furthermore, the CRF identified species co-occurrences as more informative, and thus relatively more important, at predicting occurrence than the other effect types. The CRF also suggested that allowing these biotic effects to be context-dependent was important for predicting occurrence of many species. These findings illustrate the value of fish assemblage data for landscape-scale species distribution modeling and leveraging this information can improve predictions and inferences to help inform the management and conservation of freshwater fishes. [ABSTRACT FROM AUTHOR]

We examine named entity recognition (NER), an essential and commonly used first step in many natural language processing tasks, including chatbots and language translation. We focus on the application of NER to texts that have a lot of noise, such as tweets, which is difficult due to the casual and unstructured language often used in these mediums. In this study, we make use of the largest available labeled data sets for Turkish NER, specifically targeting three informal platforms, namely Twitter, Facebook and Donanimhaber. We choose Turkish as a representative agglutinative language, which has a significantly different structure than other well-known languages such as English, French, and German. We emphasize that the methodologies and insights gained from this study can be extended to other agglutinative languages, like Finnish, Hungarian, Japanese, and Korean. We apply NER to these datasets using 16 different named entity tags through a framework that employs bidirectional long short-term memory (BiLSTM) networks followed by conditional random fields (CRF), known together as the BiLSTM-CRF model. Our experiments show an F1 score of 84% on a combined dataset, which indicates that deep learning models can also be effectively used for business applications in informal settings in agglutinative languages such as Turkish. [ABSTRACT FROM AUTHOR]

Valued constraint satisfaction problems with ordered variables (VCSPO) are a special case of Valued CSPs in which variables are totally ordered and soft constraints are imposed on tuples of variables that do not violate the order. We study a restriction of VCSPO, in which soft constraints are imposed on a segment of adjacent variables and a constraint language Γ consists of { 0 , 1 } -valued characteristic functions of predicates. This kind of potentials generalizes the so-called pattern-based potentials, which were applied in many tasks of structured prediction. For a constraint language Γ we introduce a closure operator, Γ ∩ ¯ ⊇ Γ , and give examples of constraint languages for which | Γ ∩ ¯ | is small. If all predicates in Γ are cartesian products, we show that the minimization of a generalized pattern-based potential (or, the computation of its partition function) can be made in O (| V | · | D | 2 · | Γ ∩ ¯ | 2) time, where V is a set of variables, D is a domain set. If, additionally, only non-positive weights of constraints are allowed, the complexity of the minimization task drops to O (| V | · | Γ ∩ ¯ | · | D | · max ϱ ∈ Γ ‖ ϱ ‖ 2) where ‖ ϱ ‖ is the arity of ϱ ∈ Γ . For a general language Γ and non-positive weights, the minimization task can be carried out in O (| V | · | Γ ∩ ¯ | 2) time. We argue that in many natural cases Γ ∩ ¯ is of moderate size, though in the worst case | Γ ∩ ¯ | can blow up and depend exponentially on max ϱ ∈ Γ ‖ ϱ ‖ . [ABSTRACT FROM AUTHOR]

Understanding users' requirements are essential to developing an effective AI service system, in which requirement expressions of users can be resolved into intent detection and slot filling tasks. In a lot of literature, the two tasks are normally considered as independent tasks and obtain satisfactory performance. Recently, many researchers have found that intent detection and slot filling can benefit each other since they always appear together in a sentence and may include shared information. Most of the existing joint models employ the structures of encoder and decoder and capture the cross-impact between two tasks by concatenation of hidden state information from two encoders, which ignore the dependencies among slot tags in specific intent. In this paper, we propose a novel Double-Bi-LSTM-CRF Model (DBLC), which can fit the dependency among hidden slot tags while considering the cross-impact between intent detection and slot filling. We also design and implement an intention chatbot on the tourism area, which can assist users to complete a travel plan through human-computer interaction. Extensive experiments show that our DBLC achieves state-of-the-art results on the benchmark ATIS, SNIPS, and multi-domain datasets. [ABSTRACT FROM AUTHOR]

Named Entity Recognition (NER) is the task of automatically identifying named entities and classifying them into predefined categories such as person, place, organization, among other. This task is important and challenging, especially when the system must be able to recognize named entities in many textual genres, including genres that differ from those for which it was trained. This paper aims to report the initial efforts made to adapt a NER system for many textual genres in accordance with the proposed Portuguese Named Entity Recognition task in IberLEF 2019. To achieve this goal, the system was trained in an augmented training corpus. In addition, a Local Grammar (handmade rules to identify named entities within the text) was adapted to capture rules of different textual genres. We discuss the results of this study and some difficulties involved in this task. [ABSTRACT FROM AUTHOR]

Appellate Court Modifications Extraction consists of, given an Appellate Court decision, identifying the proposed modifications by the upper Court of the lower Court judge's decision. In this work, we propose a system to extract Appellate Court Modifications for Portuguese. Information extraction for legal texts has been previously addressed using different techniques and for several languages. Our proposal differs from previous work in two ways: (1) our corpus is composed of Brazilian Appellate Court decisions, in which we look for a set of modifications provided by the Court; and (2) to automatically extract the modifications, we use a traditional Machine Learning approach and a Deep Learning approach, both as alternative solutions and as a combined solution. We tackle the Appellate Court Modifications Extraction task, experimenting with a wide variety of methods. In order to train and evaluate the system, we have built the KauaneJunior corpus, using public data disclosed by the Appellate State Court of Rio de Janeiro jurisprudence database. Our best method, which is a Bidirectional Long Short-Term Memory network combined with Conditional Random Fields, obtained an F β = 1 score of 94.79%. [ABSTRACT FROM AUTHOR]

The conditional random fields (CRFs) model plays an important role in the machine learning field. Driven by the development of the artificial intelligence, the CRF models have enjoyed great advancement. To analyze the recent development of the CRFs, this paper presents a comprehensive review of different versions of the CRF models and their applications. On the basis of elaborating on the background and definition of the CRFs, it analyzes three basic problems faced by the CRF models and reviews their latest improvements. Based on that, it presents the applications of the CRFs in the natural language processing, computer vision, biomedicine, Internet intelligence and other relevant fields. At last, specific analysis and future directions of the CRFs are discussed. [ABSTRACT FROM AUTHOR]

In order to realize real-time 3D environment perception of UAVs and autopilot in low-altitude complex road scenes, a neural network model RobNet based on SqueezeNet and cyclic CRF for real-time 3D point cloud segmentation is proposed to segment the road objects in real time. Firstly, the unordered, scattered 3D point cloud data are preprocessed into a standard data format similar to an image by a spherical mapping method. Then, at the macro-level of the model design, the SqueezeNet network with the residual connection optimization is selected as the basic network of the model, and then, the conditional random field (CRF) algorithm which is processed into the cyclic network structure is used to refine the segmentation result. Finally, the construction of the basic network, the cyclic network and the network parameter settings in the model is elaborated at the micro-level. The experimental results show that the RobNet model proposed in this paper can segment the target in the road scene better. The segmentation callback rate of the three types of vehicles, pedestrians and cyclists is increased by 28, 2 and 17%, respectively, compared with the VoxelNet network. The higher callback rate is in line with the safe movement specifications for drones and autonomous driving. At the same time, the proposed model parameters are small, 98.5% smaller than the classic network AlexNet, and are easy to deploy on a platform with limited computing resources. The RobNet model in the Robot Operating System (ROS) framework engineering deployment and implementation experimental data shows that the model meets the real-time and stability requirements of the drone and automatic driving application, engineering code can run in real time at 12 Hz, the standard deviation of each frame's running time is around 4.5 ms. [ABSTRACT FROM AUTHOR]

This paper explores the usefulness of conditional random fields through the idea of semantic face segmentation in the challenging task of head pose estimation. A multi-class face segmentation algorithm based on conditional random fields is implemented to develop a model for each discrete pose. When a new test image is given as input to the face segmentation framework, the trained model predicts probabilities for each face part. These probabilities are then used for estimation of head pose. The proposed framework is evaluated on four standard databases, namely Pointing'04, AFLW, BU and ICT-3DHPE. Two standard metrics, mean absolute error and pose estimation accuracy are used for evaluation of the head pose estimation part. Pixel labeling accuracy is used to assess the segmentation results. The experimental results show that better results are obtained as compared to state-of-the-art. [ABSTRACT FROM AUTHOR]

Accurate and fast sketch segmentation and labeling is a hard task, since sketches have much fewer features than natural images. This paper proposes a novel hybrid approach for fast automatic sketch labeling, which is based on convolutional neural network (CNN) and conditional random field (CRF). Firstly, we design a CNN for stroke classification. The CNN is equipped with larger first layer filters and larger pooling, which is suitable for extracting descriptive features from strokes. Secondly, we integrate each stroke with its host sketch to construct a more informative input for the CNN model. Finally, we leverage the spatio-temporal relations among strokes in the same sketch to create a connected graph, based on which we apply a CRF model to further refine the result of the CNN. We evaluate our method on two public benchmark datasets. Experimental results demonstrate that our method achieves the state-of-the-art level on both accuracy and runtime. [ABSTRACT FROM AUTHOR]

Many detailed geoscience reports lie unused, offering both challenges and opportunities for information extraction. In geoscience research, geological named entity recognition (GNER) is an important task in the field of geoscience information extraction. Regarding numerical geoscience data, research on information extraction remains limited. Most conventional NER approaches are heavily dependent on feature engineering, and such sentence-level-based methods suffer from the tagging inconsistency problem. Based on the above observations, this paper proposes a neural network approach, namely, attention-based bidirectional long short-term memory with a conditional random field layer (Att-BiLSTM-CRF), for name entity recognition to extract information entities describing geoscience information from geoscience reports. This approach leverages global information learned from an attention mechanism to enforce tagging consistency across multiple instances of the same token in a document. Experiments on the constructed dataset show that our method achieves comparable performance to that of other state-of-the-art systems. Additionally, our method achieved an average F1 score of 91.47% in the NER extraction task. [ABSTRACT FROM AUTHOR]

Short texts have become the prevalent format of information on the Internet. Inferring the topics of this type of messages becomes a critical and challenging task for many applications. Due to the length of short texts, conventional topic models (e.g., latent Dirichlet allocation and its variants) suffer from the severe data sparsity problem which makes topic modeling of short texts difficult and unreliable. Recently, word embeddings have been proved effective to capture semantic and syntactic information about words, which can be used to induce similarity measures and semantic correlations among words. Enlightened by this, in this paper, we design a novel model for short text topic modeling, referred as Conditional Random Field regularized Topic Model (CRFTM). CRFTM not only develops a generalized solution to alleviate the sparsity problem by aggregating short texts into pseudo-documents, but also leverages a Conditional Random Field regularized model that encourages semantically related words to share the same topic assignment. Experimental results on two real-world datasets show that our method can extract more coherent topics, and significantly outperform state-of-the-art baselines on several evaluation metrics. [ABSTRACT FROM AUTHOR]

Part of speech (POS) tagging, the assignment of syntactic categories for words in running text, is significant to natural language processing as a preliminary task in applications such as speech processing, information extraction, and others. Urdu language processing presents a challenge due to the dual behaviour of various Urdu POS tags in differing situations (morphosyntactic ambiguity). This paper addresses this challenge by developing a novel tagging approach using linear-chain conditional random fields (CRF). Our work is the first instance of a CRF approach for Urdu POS tagging. The proposed model employs a strong, stable and balanced language-independent as well as language dependent feature set. The language-dependent feature considered includes part-of-speech tag of the previous word and suffix of the current word while the language-independent features includes the 'context words window'. Our approach was evaluated against support vector machine techniques for Urdu POS—considered as state of the art—on two benchmark datasets. The results show our CRF approach to improve upon the F-measure of prior attempts by 8.3–8.5%. [ABSTRACT FROM AUTHOR]

Radiological measurements are reported in free text reports, and it is challenging to extract such measures for treatment planning such as lesion summarization and cancer response assessment. The purpose of this work is to develop and evaluate a natural language processing (NLP) pipeline that can extract measurements and their core descriptors, such as temporality, anatomical entity, imaging observation, RadLex descriptors, series number, image number, and segment from a wide variety of radiology reports (MR, CT, and mammogram). We created a hybrid NLP pipeline that integrates rule-based feature extraction modules and conditional random field (CRF) model for extraction of the measurements from the radiology reports and links them with clinically relevant features such as anatomical entities or imaging observations. The pipeline was trained on 1117 CT/MR reports, and performance of the system was evaluated on an independent set of 100 expert-annotated CT/MR reports and also tested on 25 mammography reports. The system detected 813 out of 806 measurements in the CT/MR reports; 784 were true positives, 29 were false positives, and 0 were false negatives. Similarly, from the mammography reports, 96% of the measurements with their modifiers were extracted correctly. Our approach could enable the development of computerized applications that can utilize summarized lesion measurements from radiology report of varying modalities and improve practice by tracking the same lesions along multiple radiologic encounters. [ABSTRACT FROM AUTHOR]

Human behavior recognition has become one of the most active topics in computer vision and pattern recognition, which has a wide range of promising applications. In order to overcome the deficiency of single representation feature, a new recognition algorithm of human behavior based on multi-feature fusion of image and conditional random fields (CRF) is presented in this paper. The proposed algorithm consists of three essential cascade modules. First, AE features and RNN features were obtained by extracting the behaviors of the action by the recurrent neural network (RNN) and the AutoEncoder (AE), Then, feature similarity was introduced, the AE features and RNN features were fused to form a more comprehensive and accurate AE-RNN feature by using feature similarity. Finally, the multiple features were using for recognizing the human behavior of image by conditional random fields. The experimental results show that the proposed algorithm is effective and promising and has higher accurate recognition rate which can adapt to complex background and behavioral changes. [ABSTRACT FROM AUTHOR]

Named entity recognition (NER) is one of the fundamental problems in many natural language processing applications and the study on NER has great significance. Combining words segmentation and parts of speech analysis, the paper proposes a new NER method based on conditional random fields considering the graininess of candidate entities. The recognition granularity can be divided into two levels: word-based and character-based. We use segmented text to extract characteristics according to the characteristic templates which had been trained in the training phase, and then calculate P (y | x) to get the best result from the input sequence. The paper valuates the algorithm for different graininess on large-scale corpus experimentally, and the results show that this method has high research value and feasibility. [ABSTRACT FROM AUTHOR]

Automated image annotation (AIA) is an important issue in computer vision and pattern recognition, and plays an extremely important role in retrieving large-scale images. In many image annotation approaches, different regions of the image are processed equally, which is inconsistent with the mechanism by which humans understand images. In order to improve the annotation performance of existing AIA approaches, a hybrid AIA approach based on visual attention mechanism (VAM) and the conditional random field (CRF) is proposed. First, since people pay more attention to the salient region of an image during the image recognition process, VAM is implemented for acquiring the salient and non salient regions of the image. Second, support vector machine (SVM) is used to annotate the salient region, and k nearest neighbor (kNN) voting algorithm is used to annotate the non salient regions. Finally, due to the existence of a certain relationship between any two annotation words (also called labels), CRF is calculated to obtain the final label set of each given image. The experimental results confirm that the proposed hybrid AIA approach has ideal annotation performance. [ABSTRACT FROM AUTHOR]

This work proposes a general multi-layer framework for image labeling, which targets the challenging problem of classifying the occluded parts of the 3D scene depicted in a 2D image. Our framework is based on the mixed graphical models, which explicitly encode causal relationship between the visible and occluded regions. Unlike other image labeling techniques where a single label is determined for each pixel, layered model assigns multiple labels to pixels. We propose a novel "Multi-Layer-CRF" framework that allows for the integration of sophisticated occlusion potentials into the model and enables the automatic inference of the layer decomposition. We use a special message-passing algorithm to perform maximum a posterior inference on mixed graphs and demonstrate the ability to infer the correct labels of occluded regions in both the aerial near-vertical dataset and urban street-view dataset. It is shown to increase the classification accuracy in occluded areas significantly. [ABSTRACT FROM AUTHOR]

Personal webpages of researchers or faculty members make up a percentage of the academic web. These webpages contain semi-structured or plain text information, and research has shown the importance of combining information extracted from multiple academic websites to create a unified database that can help in expert finding, and thus improve information retrieval for end users. This research identifies the kind of named entities that could be present in academic biographies by manually examining the biographies extracted from ORCID public profiles, and describes a method that uses natural language processing techniques and supervised machine learning to automatically extract these named entities from the plain text biographies. Up to 86% accuracy was achieved with support vector machines, demonstrating that the method used in this research can be suitable for creating a reusable trained model that extracts useful academic information from researchers’ personal profiles in webpages or other data sources. [ABSTRACT FROM AUTHOR]

The paper analyzes implementation of intelligence-led policing into the National Police of Ukraine. The architecture of analytical system is proposed based on the analysis of state-of-the-art in the National Police of Ukraine. The mathematical apparatus is considered for the future components of the analytical system. Successful implementation and application of new methods of criminal analysis will allow expanding it to the entire system of the National Police of Ukraine in the future and to actively use analytical methods and technologies that provide efficient investigation of criminal offenses. [ABSTRACT FROM AUTHOR]

Many of today’s online news websites and aggregator apps have enabled users to publish their opinions without respect to time and place. Existing works on topic-based sentiment analysis of product reviews cannot be applied to online news directly because of the following two reasons: (1) The dynamic nature of news streams require the topic and sentiment analysis model also to be dynamically updated. (2) The user interactions among news comments can easily lead to inaccurate topic extraction and sentiment classification. In this paper, we propose a novel probabilistic generative model (DTSA) to extract topics and the specified sentiments from news streams and analyze their evolution over time simultaneously. In DTSA, three different timescale models are studied to account for the historical dependencies of sentiment-topic word distributions at current epoch, continuous, skip and multiple timescale models. Additionally, we further consider the links among news comments to avoid the error caused by user interactions. In order to mine more interpretable topics, a Conditional Random Fields (CRF) model is adopted to label a set of meaningful phrases for augmenting the bag-of-word features. Finally, we derive distributed online inference procedures to update the model with newly arrived data and show the effectiveness of our proposed model on real-world data sets. [ABSTRACT FROM AUTHOR]

Electronic Medical Records (EMR) with unstructured sentences and various conceptual expressions provide rich information for medical information extraction. However, common Named Entity Recognition (NER) in Natural Language Processing (NLP) are not well suitable for clinical NER in EMR. This study aims at applying neural networks to clinical concept extractions. We integrate Bidirectional Long Short-Term Memory Networks (Bi-LSTM) with a Conditional Random Fields (CRF) layer to detect three types of clinical named entities. Word representations fed into the neural networks are concatenated by character-based word embeddings and Continuous Bag of Words (CBOW) embeddings trained both on domain and non-domain corpus. We test our NER system on i2b2/VA open datasets and compare the performance with six related works, achieving the best result of NER with F1 value 0.853 7. We also point out a few specific problems in clinical concept extractions which will give some hints to deeper studies. [ABSTRACT FROM AUTHOR]

This paper proposes several recurrent neural network-based models for recognizing requisite and effectuation (RE) parts in Legal Texts. Firstly, we propose a modification of BiLSTM-CRF model that allows the use of external features to improve the performance of deep learning models in case large annotated corpora are not available. However, this model can only recognize RE parts which are not overlapped. Secondly, we propose two approaches for recognizing overlapping RE parts including the cascading approach which uses the sequence of BiLSTM-CRF models and the unified model approach with the multilayer BiLSTM-CRF model and the multilayer BiLSTM-MLP-CRF model. Experimental results on two Japan law RRE datasets demonstrated advantages of our proposed models. For the Japanese National Pension Law dataset, our approaches obtained an F1 score of 93.27% and exhibited a significant improvement compared to previous approaches. For the Japan Civil Code RRE dataset which is written in English, our approaches produced an F1 score of 78.24% in recognizing RE parts that exhibited a significant improvement over strong baselines. In addition, using external features and in-domain pre-trained word embeddings also improved the performance of RRE systems. [ABSTRACT FROM AUTHOR]

A large number of cross-references to various bodies of text are used in legal texts, each serving a different purpose. It is often necessary for authorities and companies to look into certain types of these citations. Yet, there is a lack of automatic tools to aid in this process. Recently, citation graphs have been used to improve the intelligibility of complex rule frameworks. We propose an algorithm that builds the citation graph from a document and automatically labels each edge according to its purpose. Our method uses the citing text only and thus works only on citations who’s purpose can be uniquely identified by their surrounding text. This framework is then applied to the US code. This paper includes defining and evaluating a standard gold set of labels that cover a vast majority of citation types which appear in the “US Code” but are still short enough for practical use. We also proposed a novel linear-chain conditional random field model that extracts the features required for labeling the citations from the surrounding text. We then analyzed the effectiveness of different clustering methods such as K-means and support vector machine to automatically label each citation with the corresponding label. Besides this, we talk about the practical difficulties of this task and give a comparison of human accuracy compared to our end-to-end algorithm. [ABSTRACT FROM AUTHOR]

Predicting object location using a top-down saliency model has grown increasingly popular in recent years. In this work, we combine locality-constrained linear coding (LLC) with a conditional random field (CRF), and construct a top-down saliency model to generate a specific object-based saliency map. During the training phase, we use the LLC codes as the latent variables of the CRF model, and meanwhile learn a class-specific codebook by CRF modulation. In the testing phase, we use this top-down model to distinguish specific objects from a cluttered background. Finally, we evaluate the experimental results on the MSRA-B, Garz-02, Weizmann Horse, and Plane datasets by applying the developed object-based saliency model. The performance shows that our approach can not only improve the precision but also dramatically reduce the computational complexity. [ABSTRACT FROM AUTHOR]

The automatic analysis of the surgical process, from videos recorded during surgeries, could be very useful to surgeons, both for training and for acquiring new techniques. The training process could be optimized by automatically providing some targeted recommendations or warnings, similar to the expert surgeon's guidance. In this paper, we propose to reuse videos recorded and stored during cataract surgeries to perform the analysis. The proposed system allows to automatically recognize, in real time, what the surgeon is doing: what surgical phase or, more precisely, what surgical step he or she is performing. This recognition relies on the inference of a multilevel statistical model which uses 1) the conditional relations between levels of description (steps and phases) and 2) the temporal relations among steps and among phases. The model accepts two types of inputs: 1) the presence of surgical tools, manually provided by the surgeons, or 2) motion in videos, automatically analyzed through the Content Based Video retrieval (CBVR) paradigm. Different data-driven statistical models are evaluated in this paper. For this project, a dataset of 30 cataract surgery videos was collected at Brest University hospital. The system was evaluated in terms of area under the ROC curve. Promising results were obtained using either the presence of surgical tools ( A = 0.983) or motion analysis ( A = 0.759). The generality of the method allows to adapt it to other kinds of surgeries. The proposed solution could be used in a computer assisted surgery tool to support surgeons during the surgery. [ABSTRACT FROM AUTHOR]

For many real-world applications, structured regression is commonly used for predicting output variables that have some internal structure. Gaussian conditional random fields (GCRF) are a widely used type of structured regression model that incorporates the outputs of unstructured predictors and the correlation between objects in order to achieve higher accuracy. However, applications of this model are limited to objects that are symmetrically correlated, while interaction between objects is asymmetric in many cases. In this work we propose a new model, called Directed Gaussian conditional random fields (DirGCRF), which extends GCRF to allow modeling asymmetric relationships (e.g. friendship, influence, love, solidarity, etc.). The DirGCRF models the response variable as a function of both the outputs of unstructured predictors and the asymmetric structure. The effectiveness of the proposed model is characterized on six types of synthetic datasets and four real-world applications where DirGCRF was consistently more accurate than the standard GCRF model and baseline unstructured models. [ABSTRACT FROM AUTHOR]

As one of the famous probabilistic graph models in machine learning, the conditional random fields (CRFs) can merge different types of features, and encode known relationships between observations and construct consistent interpretations, which have been widely applied in many areas of the Natural Language Processing (NLP). With the high-speed development of the internet and information systems, some performance issues are certain to arise when the traditional CRFs deals with such massive data. This paper proposes SCRFs, which is a parallel optimization of CRFs based on the Resilient Distributed Datasets (RDD) in the Spark computing framework. SCRFs optimizes the traditional CRFs from these stages: First, with all features are generated in parallel, the intermediate data which will be used frequently are all cached into the memory to speed up the iteration efficiency. By removing the low-frequency features of the model, SCRFs can also prevent the overfitting of the model to improve the prediction effect. Second, some specific features are dynamically added in parallel to correct the model in the training process. And for implementing the efficient prediction, a max-sum algorithm is proposed to infer the most likely state sequence by extending the belief propagation algorithm. Finally, we implement SCRFs base on the version of Spark 1.6.0, and evaluate its performance using two widely used benchmarks: Named Entity Recognition and Chinese Word Segmentation. Compared with the traditional CRFs models running on the Hadoop and Spark platforms respectively, the experimental results illustrate that SCRFs has obvious advantages in terms of the model accuracy and the iteration performance. [ABSTRACT FROM AUTHOR]

Tagging biomedical entities such as gene, protein, cell, and cell-line is the first step and an important pre-requisite in biomedical literature mining. In this paper, we describe our hybrid named entity tagging approach namely BCC-NER (bidirectional, contextual clues named entity tagger for gene/protein mention recognition). BCC-NER is deployed with three modules. The first module is for text processing which includes basic NLP pre-processing, feature extraction, and feature selection. The second module is for training and model building with bidirectional conditional random fields (CRF) to parse the text in both directions (forward and backward) and integrate the backward and forward trained models using margin-infused relaxed algorithm (MIRA). The third and final module is for post-processing to achieve a better performance, which includes surrounding text features, parenthesis mismatching, and two-tier abbreviation algorithm. The evaluation results on BioCreative II GM test corpus of BCC-NER achieve a precision of 89.95, recall of 84.15 and overall F-score of 86.95, which is higher than the other currently available open source taggers. [ABSTRACT FROM AUTHOR]

Deep convolutional neural networks (DCNNs) based methods recently keep setting new records on the tasks of predicting depth maps from monocular images. When dealing with video-based applications such as 2D (2-dimensional) to 3D (3-dimensional) video conversion, however, these approaches tend to produce temporally inconsistent depth maps, since their CNN models are optimized over single frames. In this paper, we address this problem by introducing a novel spatial-temporal conditional random fields (CRF) model into the DCNN architecture, which is able to enforce temporal consistency between depth map estimations over consecutive video frames. In our approach, temporally consistent superpixel (TSP) is first applied to an image sequence to establish the correspondence of targets in consecutive frames. A DCNN is then used to regress the depth value of each temporal superpixel, followed by a spatial-temporal CRF layer to model the relationship of the estimated depths in both spatial and temporal domains. The parameters in both DCNN and CRF models are jointly optimized with back propagation. Experimental results show that our approach not only is able to significantly enhance the temporal consistency of estimated depth maps over existing single-frame-based approaches, but also improves the depth estimation accuracy in terms of various evaluation metrics. [ABSTRACT FROM AUTHOR]

State-of-the-art methods for disparity estimation achieve good results for single stereo frames, but temporal coherence in stereo videos is often neglected. In this paper, we present a method to compute temporally coherent disparity maps. We define an energy over whole stereo sequences and optimize their conditional random field (CRF) distributions using the mean-field approximation. In addition, we introduce novel terms for smoothness and consistency between the left and right views. We perform CRF optimization by fast, iterative spatio-temporal filtering with linear complexity in the total number of pixels. We propose two CRF optimization techniques, using parallel and sequential updates, and compare them in detail. While parallel updates are not guaranteed to converge, we show that, in practice with appropriate initialization, they provide the same quality as sequential updates and they also lead to faster implementations. Finally, we demonstrate that the results of our approach rank among the state of the art while having significantly less flickering artifacts in stereo sequences. [ABSTRACT FROM AUTHOR]

Tracking the spatio-temporal activity is highly relevant for domains like security, health, and quality management. Since animal welfare became a topic in politics and legislation locomotion patterns of livestock have received increasing interest. In contrast to the monitoring of pedestrians cattle activity tracking poses special challenges to both sensors and data analysis. Interesting states are not directly observable by a single sensor. In addition, sensors must be accepted by cattle and need to be robust enough to cope with a rough environment. In this article, we introduce the novel combination of heart rate and positioning sensors. Attached to neck and chest they are less interfering than accelerometers at the ankles. Exploiting the potential of such combined sensor system that records locomotion and non-spatial information from the heart rate sensor however is challenging. We introduce a novel two level method for the activity tracking focused on the duration and sequence of activity states. We combine Support Vector Machine (SVM) with Conditional Random Field (CRF) and extend Conditional Random fields by an explicit representation of duration. The SVM characterizes local activity states, whereas the CRF addresses sequences of local states to sequences incorporating spatial and non-spatial contextual knowledge. This combination provides a reliable and comprehensive identification of defined activity patterns, as well as their chronology and durations, suitable for the integration in an activity data base. This data base is used to extract physiological parameters and promises insights into internal states such as fitness, well-being and stress. Interestingly we were able to demonstrate a significant correlation between resting pulse rate and the day of pregnancy. [ABSTRACT FROM AUTHOR]

We consider Conditional random fields ( CRFs) with pattern-based potentials defined on a chain. In this model the energy of a string (labeling) $$x_1\ldots x_n$$ is the sum of terms over intervals [ i, j] where each term is non-zero only if the substring $$x_i\ldots x_j$$ equals a prespecified pattern w. Such CRFs can be naturally applied to many sequence tagging problems. We present efficient algorithms for the three standard inference tasks in a CRF, namely computing (i) the partition function, (ii) marginals, and (iii) computing the MAP. Their complexities are respectively $$O(\textit{nL})$$ , $$O(\textit{nL} \ell _{\max })$$ and $$O(\textit{nL} \min \{|D|,\log (\ell _{\max }\!+\!1)\})$$ where L is the combined length of input patterns, $$\ell _{\max }$$ is the maximum length of a pattern, and D is the input alphabet. This improves on the previous algorithms of Ye et al. (NIPS, 2009) whose complexities are respectively $$O(\textit{nL} |D|)$$ , $$O\left( n |\varGamma | L^2 \ell _{\max }^2\right) $$ and $$O(\textit{nL} |D|)$$ , where $$|\varGamma |$$ is the number of input patterns. In addition, we give an efficient algorithm for sampling, and revisit the case of MAP with non-positive weights. [ABSTRACT FROM AUTHOR]

Background: In this paper, we investigate the effectiveness of a CRF-based learning method for identifying necessary Web of Things (WoT) application components that would satisfy the users' requests issued in natural language. For instance, a user request such as 'archive all sports breaking news' can be satisfied by composing a WoT application that consists of ESPN breaking news service and Dropbox as a storage service. Findings: We built an engine that can identify the necessary application components by recognizing a main act (MA) or named entities (NEs) from a given request. We trained this engine with the descriptions of WoT applications (called recipes) that were collected from IFTTT WoT platform. IFTTT hosts over 300 WoT entities that offer thousands of functions referred to as triggers and actions. There are more than 270,000 publicly-available recipes composed with those functions by real users. Therefore, the set of these recipes is well-qualified for the training of our MA and NE recognition engine. Conlusions: We share our unique experience of generating the training and test set from these recipe descriptions and assess the performance of the CRF-based language method. Based on the performance evaluation, we introduce further research directions. [ABSTRACT FROM AUTHOR]

Automatically recognizing the e-learning activities is an important task for improving the online learning process. Probabilistic graphical models such as hidden Markov models and conditional random fields have been successfully used in order to identify a Web users' activity. For such models, the sequences of observation are crucial for training and inference processes. Despite the efficiency of these probabilistic graphical models in segmenting and labeling stochastic sequences, their performance is adversely affected by the imperfect quality of data used for the construction of sequences of observation. In this paper, a formalism of the possibilistic theory will be used in order to propose a new approach for observation sequences preparation. The eminent contribution of our approach is to evaluate the effect of possibilistic reasoning-during the generation of observation sequences-on the effectiveness of hidden Markov models and conditional random fields models. Using a dataset containing 51 real manipulations related to three types of learners' tasks, the preliminary experiments demonstrate that the sequences of observation obtained based on possibilistic reasoning significantly improve the performance of hidden Marvov models and conditional random fields models in the automatic recognition of the e-learning activities. [ABSTRACT FROM AUTHOR]

Although the hierarchical conditional random field (HCRF) model has been successfully applied to multi-class object segmentation, there is still room for improvement. Firstly, the pairwise potential in the HCRF model has the tendency to over-smooth boundaries of regions that are similar to their neighbors. Secondly, the higher-order potential associated with multiple unsupervised segments is prone to producing incorrect guidance to inference in the under-segmentation situation. Finally, the co-occurrence potential as a measure of inter-object relationships cannot completely suppress some uncommon combinations of object classes due to joint optimization of multi-potential cost function. To alleviate these problems, we propose an improved HCRF model that efficiently combines information from global, middle and local scales for object segmentation in this paper. At the global scale, scene categorization technique is adopted to recognize the scene category of an image. The scene consistency then enforces object segmentation to align with feasible labels in specific scenes at the local and middle scales. Furthermore, an improved pairwise potential and a segment-reliable consistency potential are developed at the local and middle scales, respectively. These potentials rectify the over-smoothness issues by propagating the believed labeling from the unary potential and perform coherent inference by ensuring reliable segment consistency. Experimental results on the MSRC-21 dataset demonstrate that the improved HCRF model achieves better subjective results, as well as state-of-the-art objective results in terms of both global accuracy of 87.98 % and average accuracy of 81.43 %. [ABSTRACT FROM AUTHOR]

A high-order conditional random field (CRF) for depth estimation from a single image is proposed in this paper. Instead of formulating the problem with the Guassian or Laplacian CRF modeling techniques, which cannot exploit the full potential offered by the probabilistic modeling, this paper proposes a depth estimation CRF model with field of experts (FoE) as the prior. The minimum mean square error (MMSE) criteria is used to infer depth. Moreover, it is assumed that the variance of depth estimation error varies spatially in depth estimation model. This allows the proposed method to enjoy the benefits offered by the flexible prior and have the advantages of making use of the non-stationary variance probability model. Experimental results indicate that the proposed method outperforms state-of-the-art approaches in terms of RMSE-error and log10-error. [ABSTRACT FROM AUTHOR]

Image automatic annotation is an important issue of semantic-based image retrieval, and it is still a challenging problem for the reason of semantic gap. In this paper, a novel model with three parts is proposed. The first one is multi-layer image segmentation, in which saliency analysis and normalized cut are combined to segment images into semantic regions in the first layer. While in the second layer, the semantic regions are segmented into grids further . The second one is image content representation by region-based bag-of-words (RBoW) model, which is the variant of BoW model. Considering the correlations of labels, we adopt second-order CRFs as the third part of our model to ensure the accuracy of automatic image annotation. Experimental results show that our multi-layer segmentation-based image annotation model can achieve promising performance for multi-labeling and outperform the model based on single-layer segmentation and previous algorithm on Corel 5K and Pascal VOC 2007 datasets . [ABSTRACT FROM AUTHOR]

As the rapid growth of the biomedical literature, the model training time in biomedical named entity recognition increases sharply when dealing with large-scale training samples. How to increase the efficiency of named entity recognition in biomedical big data becomes one of the key problems in biomedical text mining. For the purposes of improving the recognition performance and reducing the training time, this paper proposes an optimization method for two-phase recognition using conditional random fields. In the first stage, each named entity boundary is detected to distinguish all real entities. In the second stage, we label the semantic class of the entity detected. To expedite the training speed, in these two phases, we implement the model training process on a parallel optimization program framework based on MapReduce. Through dividing the training set into several parts, the iterations in the training algorithm are designed as map tasks which can be executed simultaneously in a cluster, where each map function is designed to complete the calculation of a gradient vector component for each part in the training set. Our experiments show that the proposed method in this paper can achieve high performance with short training time, which has important implications for the current biological big data processing. [ABSTRACT FROM AUTHOR]

We present a new descriptor for activity recognition from skeleton data acquired by Kinect. Previous approaches tend to employ complex descriptors which require extensively long computation time. In this study, we present an efficient and effective descriptor which we name as Histogram-of-Oriented-Velocity-Vectors (HOVV). It is a scale-invariant, speed-invariant and length-invariant descriptor for human actions represented by 3D skeletons acquired by Kinect. We describe the skeleton sequence using 2D spatial histogram capturing the distribution of the orientations of velocity vectors of the joint in a spherical coordinate system. We make use of three methods to classify actions represented by HOVV descriptor. These are k-nearest neighbour classifier, Support Vector Machines classifier and Extreme Learning Machines. For the cases when HOVV descriptor is not sufficient, such as to differentiate actions which involve tiny movement of joints such as 'sit-still', we also incorporate a simple skeleton descriptor as a prior to the action descriptor. Through extensive experiments, we test our system with different configurations. We also demonstrate that our HOVV descriptor outperforms the state-of-the-art methods. The results demonstrate that our descriptor has much shorter computational time due to the simpler computations needed for feature extraction. Moreover our descriptor shows a higher recognition accuracy compared with the state-of-the-art methods. [ABSTRACT FROM AUTHOR]