Capturing high-level structure of naturally occurring, extended explanations using bottom-up strategies

This paper notes some structural features of naturally occurring, extended explanations and argues that current generation methodologies are inadequate for determining high-level structure. We then present a computational model for generating high-level structure, which is based on the hypothesis that high-level structure is distinct from text structure at lower levels and is determined by the material to be conveyed rather than by independent, recursive text structuring devices. We propose that high-level structure can be determined by bottom-up processes that attempt to satisfy speaker, listener, and compositional goals, after which top-down strategies can be used to organize the material about the selected framework.
Cet article relève certaines caractéristiques structurales ďexplications étendues à occurrence naturelle et soutient que les méthodes de génération courantes ne permettent pas de déterminer adéquatement une stratégie de haut niveau. Les auteurs présentent ensuite un modèle computationnel de géneration de structures de haut niveau qui est basé sur ľhypothèse qu'une structure de haut niveau est différente ďune structure de texte aux nivaux inféiieurs, et qu'elle est déterminée par les données qui doivent ětre transmises plutt que par des mécanismes indépendants et récursifs de structure de texte. lis soutiennent qu'une structure de haut niveau peut ětre déterminée àľaide ď'un processus ascendant qui essaie de rencontrer les objectifs du locuteur, de ľauditeur et de composition, et que des stratégies descendantes peuvent ensuite ětre utilisées pour organiser les données a propos du cadre choisi.     

The role of the user's domain knowledge in generation

A question-answering program that provides access to a large amount of data will be most useful if it can tailor its answers to each individual user. In particular, a user's level of knowledge about the domain of discourse is an important factor in this tailoring if the answer provided is to be both informative and understandable to the user. In this research, we address the issue of how the user's domain knowledge, or the level of expertise, might affect an answer. We present TAILOR, a flexible computer system that takes into account this knowledge to provide an answer that is appropriate for users with varying levels of expertise (including novices and experts), without requiring an a priori set of user types.
Un programme qui donne accès à un système de connaissances à de nombreux et divers utilisateurs serait plus utile s'il pouvait adapter ses réponses pour chaque utilisateur. En particulier, le savoir de l'utilisateur est un facteur important pour que la réponse soit compréhensible et informative. Dans cet article, nous étudions le rle de la connaissance de l'utilisateur pour former un texte qui soit approprié pour la personne qui le lira. Nous présentons TAILOR, un système de génération flexible qui est sensitif au niveau de connaissance de l'utilisateur. TAILOR peut générer des textes appropries pour des utilisateurs aux niveaux de connaissances très variés, sans avoir besoin ďun groupe de stéréotypes définis a priori.     

Bridging the generation gap between text planning and linguistic realization

Text planning is traditionally done as a separate process, independent of linguistic realization. However, I show in this paper that it is essential for a text planner to know the linguistic consequences of its actions. The choice of how to realize an element affects how much information is conveyed (e.g., “Floyd arrived safely in Boston” vs. “Floyd landed safely at Logan”) and what other information can be added (e.g., “Peter was deciding for an hour” vs. *“Peter made a decision for an hour”). The choice of realization also affects how the relative salience of the elements being expressed will be perceived (e.g., “The green car is in the garage” vs. “The car in the garage is green”). I have defined an intermediate level of representation for text planning, called the Text Structure. It is an abstract linguistic level that reflects germane linguistic constraints while abstracting away from syntactic detail. This representation allows the text planner to have greater control over the decisions, so that it can take advantage of the expressiveness of language to convey subtleties of meaning. More importantly, the Text Structure allows the generation process overall to be incremental, since it ensures that the text plan being composed will always be expressible in the language. La planification de textes est habituellement réalisée séparément, sans tenir compte de la réalisation linguistique. Cependant, ľauteur demontre dans cet article qu'il est essentiel à un planificateur de textes de connai̊tre les conséquences linguistiques de ses actions. Le choix de la méhode de réalisation ?un élément influe sur la quantité?informations qui est transmise (par ex.: 〈〈 Floyd est arrivé sain et sauf à Boston 〉〉 et 〈〈 Floyd a atterri sain et sauf àľaéroport Logan 〉〉) et sur quelle autre information peut ětre ajoutée. Le choix de la méthode de réalisation influe également sur la façon dont les caractères saillants des ééments qui sont exprimés seront perçus (par ex.: 〈〈 la voiture verte est dans le garage 〉〉 et 〈〈 la voiture dans le garage est verte 〉〉). L'auteur a défini un niveau intérmediaire de représentation pour la planification de textes qu'il a appelé structure de texte. Il; s'agit ?'un niveau linguistique abstrait qui refléte les contraintes linguistiques appropriées tout en s'éloignant du détail syntaxique. Cette représentation permet au planificateur de textes ?avoir un meilleur contrǒle des décisions, et done de tirer profit de la force ?expression du langage afin de tenir compte des subtiliés de sens. Plus important encore, la structure de texte permet au processus de génération ?ětre incrémentiel, car elle s'assure que le plan de texte en voie de composition soit toujours exprimable dans le langage.     

Augmenting a conceptual model with geospatiotemporal annotations

While many real-world applications need to organize data based on space (e.g., geology, geomarketing, environmental modeling) and/or time (e.g., accounting, inventory management, personnel management), existing conventional conceptual models do not provide a straightforward mechanism to explicitly capture the associated spatial and temporal semantics. As a result, it is left to database designers to discover, design, and implement - on an ad hoc basis - the temporal and spatial concepts that they need. We propose an annotation-based approach that allows a database designer to focus first on nontemporal and nongeospatial aspects (i.e., "what") of the application and, subsequently, augment the conceptual schema with geospatiotemporal annotations (i.e., "when" and "where"). Via annotations, we enable a supplementary level of abstraction that succinctly encapsulates the geospatiotemporal data semantics and naturally extends the semantics of a conventional conceptual model. An overarching assumption in conceptual modeling has always been that expressiveness and formality need to be balanced with simplicity. We posit that our formally defined annotation-based approach is net only expressive, but also straightforward to understand and implement.     

Symbolic user-defined periodicity in temporal relational databases

Calendars and periodicity play a fundamental role in many applications. Recently, some commercial databases started to support user-defined periodicity in the queries in order to provide "a human-friendly way of handling time" (see, e.g., TimeSeries in Oracle 8). On the other hand, only few relational data models support user-defined periodicity in the data, mostly using "mathematical" expressions to represent periodicity. In this paper, we propose a high-level "symbolic" language for representing user-defined periodicity which seems to us more human-oriented than mathematical ones, and we use the domain of Gadia's temporal elements in order to define its properties and its extensional semantics. We then propose a temporal relational model which supports user-defined "symbolic" periodicity (e.g., to express "on the second Monday of each month") in the validity time of tuples and also copes with frame times (e.g., "from 1/1/98 to 28/2/98"). We define the temporal counterpart of the standard operators of the relational algebra, and we introduce new temporal operators and functions. We also prove that our temporal algebra is a consistent extension of the classical (atemporal) one. Moreover, we define both a fully symbolic evaluation method for the operators on the periodicities in the validity times of tuples, which is correct but not complete, and semisymbolic one, which is correct and complete, and study their computational complexity.     

Spatial ordering and encoding for geographic data mining and visualization

Geographic information (e.g., locations, networks, and nearest neighbors) are unique and different from other aspatial attributes (e.g., population, sales, or income). It is a challenging problem in spatial data mining and visualization to take into account both the geographic information and multiple aspatial variables in the detection of patterns. To tackle this problem, we present and evaluate a variety of spatial ordering methods that can transform spatial relations into a one-dimensional ordering and encoding which preserves spatial locality as much possible. The ordering can then be used to spatially sort temporal or multivariate data series and thus help reveal patterns across different spaces. The encoding, as a materialization of spatial clusters and neighboring relations, is also amenable for processing together with aspatial variables by any existing (non-spatial) data mining methods. We design a set of measures to evaluate nine different ordering/encoding methods, including two space-filling curves, six hierarchical clustering based methods, and a one-dimensional Sammon mapping (a multidimensional scaling approach). Evaluation results with various data distributions show that the optimal ordering/encoding with the complete-linkage clustering consistently gives the best overall performance, surpassing well-known space-filling curves in preserving spatial locality. Moreover, clustering-based methods can encode not only simple geographic locations, e.g., x and y coordinates, but also a wide range of other spatial relations, e.g., network distances or arbitrarily weighted graphs.     

Space-time super-resolution

We propose a method for constructing a video sequence of high space-time resolution by combining information from multiple low-resolution video sequences of the same dynamic scene. Super-resolution is performed simultaneously in time and in space. By "temporal super-resolution," we mean recovering rapid dynamic events that occur faster than regular frame-rate. Such dynamic events are not visible (or else are observed incorrectly) in any of the input sequences, even if these are played in "slow-motion." The spatial and temporal dimensions are very different in nature, yet are interrelated. This leads to interesting visual trade-offs in time and space and to new video applications. These include: 1) treatment of spatial artifacts (e.g., motion-blur) by increasing the temporal resolution and 2) combination of input sequences of different space-time resolutions (e.g., NTSC, PAL, and even high quality still images) to generate a high quality video sequence. We further analyze and compare characteristics of temporal super-resolution to those of spatial super-resolution. These include: the video cameras needed to obtain increased resolution; the upper bound on resolution improvement via super-resolution; and, the temporal analogue to the spatial "ringing" effect.     

Video parsing via spatiotemporally analysis with images

Effective parsing of video through the spatial and temporal domains is vital to many computer vision problems because it is helpful to automatically label objects in video instead of manual fashion, which is tedious. Some literatures propose to parse the semantic information on individual 2D images or individual video frames, however, these approaches only take use of the spatial information, ignore the temporal continuity information and fail to consider the relevance of frames. On the other hand, some approaches which only consider the spatial information attempt to propagate labels in the temporal domain for parsing the semantic information of the whole video, yet the non-injective and non-surjective natures can cause the black hole effect. In this paper, inspirited by some annotated image datasets (e.g., Stanford Background Dataset, LabelMe, and SIFT-FLOW), we propose to transfer or propagate such labels from images to videos. The proposed approach consists of three main stages: I) the posterior category probability density function (PDF) is learned by an algorithm which combines frame relevance and label propagation from images. II) the prior contextual constraint PDF on the map of pixel categories through whole video is learned by the Markov Random Fields (MRF). III) finally, based on both learned PDFs, the final parsing results are yielded up to the maximum a posterior (MAP) process which is computed via a very efficient graph-cut based integer optimization algorithm. The experiments show that the black hole effect can be effectively handled by the proposed approach.     

基于LSTM的POI个性化推荐框架

近年来,随着基于位置的社会网络（Location-Based Social Network,LBSN）热度的不断增加,为用户推荐下一个POI （Point-Of-Interests）也显得越来越重要.而对应的各种应用搜集到的用户的行为时间、地理、好友以及标签等信息的增多,使得POI推荐变得更加容易.目前针对POI推荐,已经有部分算法提出,但是他们受限于自身的局限性,还都不能很好的解决这个问题,例如,个性化马尔科夫链（Factorizing Personalized Markov Chain,FPMC）、张量分解（Tensor Factorization,TF）、RNN （Recurrent Neural Networks）等.但是,这些模型由于其本身缺陷,都不能完美的糅合POI场景中的所有信息.在这篇文章中,我们扩展了长短时记忆循环神经网络（Long-ShorT Memory recurrent neural networks,LSTM）,提出一种全新的推荐框架POI-LSTM来解决POI推荐问题.POI-LSTM借鉴Embedding的思想,对用户信息、好友关系、POI信息和评论信息进行向量化后,输入到神经网络中,同时利用LSTM捕捉用户的兴趣特征和兴趣的变化趋势,最终能够在不同的输入层拟合社交网络信息和语义信息,同时利用用户历史行为的时间和地理位置信息来为用户推荐下一个兴趣点.     

Time limitations and geometrical parameters in the design of microfluidic comparators

The ability to control the flow of particles (e.g., droplets and cells) in microfluidic environments can enable new methods for synthesis of biomaterials (Mann and Ozin in Nature 382:313–318, 1996), biocharacterization, and medical diagnosis (Pipper et al. in Nat Med 13:1259–1263, 2007). Understanding the factors that affect the particle passage can improve the control over the particles’ flow through microchannels (Vanapalli et al. in Lab Chip 9:982, 2009). The first step to understand the particle passage is to measure the resulting flow rate, induced pressure drop across the channel, and other parameters. Flow rates and pressure drops during passage of a particle through microchannels are typically measured using microfluidic comparators. Since the first microfluidic comparators were reported, a few design factors have been explored experimentally and theoretically, e.g., sensitivity (Vanapalli et al. in Appl Phys Lett 90:114109, 2007). Nevertheless, there is still a gap in the understanding of the temporal and spatial resolution limits of microfluidic comparators. Here we explore, theoretically and experimentally, the factors that affect the spatial and temporal resolution. We determined that the comparator sensitivity is defined by the device geometry adjacent and upstream the measuring point in the comparator. Further, we determined that, in order of importance, the temporal resolution is limited by the convective timescale, capacitive timescale due to channel expansion, and unsteady timescale due to the flow inertia. Finally, we explored the flow velocity limits by characterizing the transition between low to moderate Reynolds numbers (Re <<1 to Re ~ 50). The present work can guide the design of microfluidic comparators and clarify the limits of this technique.     

Linguistic Theory Based Contextual Evidence Mining for Statistical Chinese Co-Reference Resolution

Under statistical learning framework, the paper focuses on how to use traditional linguistic findings on anaphora resolution as a guide for mining and organizing contextual features for Chinese co-reference resolution. The main achievements are as follows. （1） In order to simulate ＂syntactic and semantic parallelism factor＂, we extract ＂bags of word form and POS＂ feature and ＂bag of seines＂ feature from the contexts of the entity mentions and incorporate them into the baseline feature set. （2） Because it is too coarse to use the feature of bags of word form, POS tag and seme to determine the syntactic and semantic parallelism between two entity mentions, we propose a method for contextual feature reconstruction based on semantic similarity computation, in order that the reconstructed contextual features could better approximate the anaphora resolution factor of ＂Syntactic and Semantic Parallelism Preferences＂. （3） We use an entity-mention-based contextual feature representation instead of isolated word-based contextual feature representation, and expand the size of the contextual windows in addition, in order to approximately simulate ＂the selectional restriction factor＂ for anaphora resolution. The experiments show that the multi-level contextual features are useful for co-reference resolution, and the statistical system incorporated with these features performs well on the standard ACE datasets.     

Continuous implementation of information technology: The development of an interview guide and a cross-national comparison of Austrian and American organizations

An important contextual factor in the implementation of information technology is the management of the implementation or the “implementation style,” for example, number of project tools and user involvement. The implementation style is assumed to mediate the relationship between implementation characteristics (e.g., the size and type of the implementation) and implementation effects (e.g., negative or positive effects on employees and the organization). An interview guide was developed in German and English language versions for evaluating the characteristics, management, and effects of information technology implementation. Data were collected on 60 technology implementations in Austria and 18 implementations in the United States from the project managers using the interview guide. The implementations were “continuous” (no first-time introduction of information technology), were considered as major implementations (e.g., update of a software package was excluded) and affected at least 10 employees. Compared with the Austrian implementations, the American implementations were characterized by a higher degree of professionalism (that is, more use of project tools), more participation measures, but at the same time by more negative effects for the employees, such as personnel reduction. © 1999 John Wiley & Sons, Inc.     

Linguistic modelling of imperfect spatial information as a basis for simplifying spatial analysis

Spatial analysis is at the basis of several decision processes in fields such as urban planning and environmental management. In order to carry it out effectively more flexible Geographic Information Systems are needed to be able to represent and manage the imperfection that invariably affects geographic information. Starting from the consideration that experts usually deal with imperfect spatial information through linguistic terms (e.g., they identify the approximate position of a phenomenon on a map and classify spatial properties through linguistic labels), this contribution proposes the use of linguistic granule of information to represent and manage imperfect spatial information. A fuzzy object-based data model is proposed as a tool for supporting spatial analysis based on the management of linguistic granule. In particular, the problem of defining methods to manage imperfect information depending on the type of imperfection is discussed. Finally, an example of spatial analysis applied to support a decision problem in environmental impact assessment is described.     

过程工业大数据建模研究展望

人们对大数据的认识已从"3Vs" (Volume-大容量; Variety-多样性; Velocity-处理实时性)、"4Vs" ("3Vs"与Value-价值)、到现今的"5Vs" ("4Vs"与Veracity-真实性).在此背景下, 首先分析过程工业大数据的"5Vs"特性; 接下来, 综述现有数据建模方法, 并结合过程工业大数据特有性质 (包括:多层面不规则采样性、多时空时间序列性、不真实数据混杂性) 论述现有数据建模方法应用于工业大数据建模时的局限; 最后, 探讨过程工业大数据建模有待研究的问题, 包括:1) 多层面不规则采样数据的潜结构建模; 2) 用于事件发现、决策和因果分析的多时空时间序列数据建模; 3) 含有不真实数据的鲁棒建模; 4) 支持实时建模的大容量数据计算架构与方法.     

The user model-based summarize and refine approach improves information presentation in spoken dialog systems

A common task for spoken dialog systems (SDS) is to help users select a suitable option (e.g., flight, hotel, and restaurant) from the set of options available. As the number of options increases, the system must have strategies for generating summaries that enable the user to browse the option space efficiently and successfully. In the user-model based summarize and refine approach (UMSR, Demberg and Moore, 2006), options are clustered to maximize utility with respect to a user model, and linguistic devices such as discourse cues and adverbials are used to highlight the trade-offs among the presented items. In a Wizard-of-Oz experiment, we show that the UMSR approach leads to improvements in task success, efficiency, and user satisfaction compared to an approach that clusters the available options to maximize coverage of the domain (Polifroni et al., 2003). In both a laboratory experiment and a web-based experimental paradigm employing the Amazon Mechanical Turk platform, we show that the discourse cues in UMSR summaries help users compare different options and choose between options, even though they do not improve verbatim recall. This effect was observed for both written and spoken stimuli.     

二维逻辑PPTLSL的可满足性检查

由于指针的灵活性以及别名现象的存在,程序的运行可能会出现悬空指针引用、内存泄漏等诸多问题.PPTLSL是一种二维(时间和空间)时序逻辑,它结合了分离逻辑(Separation Logic)与命题投影时序逻辑PPTL(Propositional Projection Temporal Logic),能够描述和验证操作链表的指针程序的时序性质.本文简要回顾了PPTLSL的相关理论,并详细介绍工具SAT-PPTLSL的工作原理.该工具主要利用PPTLSL与PPTL之间构建起来的“同构”关系进行PPTLSL公式的可满足性检查.此外,本文结合一些实例展示了SAT-PPTLSL的执行过程,并通过实验分析了关键参数对SAT-PPTLSL执行效率的影响.     

Linear Sequence-to-Sequence Alignment

In this paper, we consider the problem of estimating the spatiotemporal alignment between N unsynchronized video sequences of the same dynamic 3D scene, captured from distinct viewpoints. Unlike most existing methods, which work for N=2 and rely on a computationally intensive search in the space of temporal alignments, we present a novel approach that reduces the problem for general N to the robust estimation of a single line in {{hbox{rlap{I}kern 2.0pt{hbox{R}}}}}^{N}. This line captures all temporal relations between the sequences and can be computed without any prior knowledge of these relations. Considering that the spatial alignment is captured by the parameters of fundamental matrices, an iterative algorithm is used to refine simultaneously the parameters representing the temporal and spatial relations between the sequences. Experimental results with real-world and synthetic sequences show that our method can accurately align the videos even when they have large misalignments (e.g., hundreds of frames), when the problem is seemingly ambiguous (e.g., scenes with roughly periodic motion), and when accurate manual alignment is difficult (e.g., due to slow-moving objects).     

Interactive dialogue model: a design technique for multichannel applications

Multichannel applications deliver the same content and a "similar interactive experience" using different devices and different technologies(e.g., web sites, palm held devices, car navigators, or interactive TVs). Various channels imply a number of differences, including screen (size), keyboard(size), pointing devices, output devices, performances, and the context of use (standing, sitting, walking, etc.). In most cases, today, applications for different channels are designed and implemented almost "independently", with ineffectiveness for the developers (high costs) and ineffectiveness for the users (loss of consistency across the different channels and the perception that they are "different applications"). This paper presents an interactive dialogue model (IDM), a novel design model specifically tailored for multichannel applications. The background research, moving from linguistic theories and practices, has led us to the development of a "channel-independent" design model (based on dialogue primitives). Design can start in a "conceptual", channel-independent fashion, and then proceed into a further "logical" design oriented toward specific channels of communication. Designing an interactive application in two steps (channel-independent first, and channel-dependent later) allows a number of advantages without making more cumbersome the overall design process. Beside the emphasis on multichannel, IDM has additional distinctive features: it is lightweight, providing a few set of primitives (and a simple graphic notation) which are easy to learn and teach. Moreover, it is suitable for brainstorming and generating ideas at early stage during design (or during the shift from requirements to design); finally, it is cost-effective (it requires little effort from designers) and modular (designers can take the part they wish, not being forced to "all or nothing"). IDM has been validated both in the academic and industry environments,providing excellent results so far.     

M-FFN: multi-scale feature fusion network for image captioning

In this work, we present a novel multi-scale feature fusion network (M-FFN) for image captioning task to incorporate discriminative features and scene contextual information of an image. We construct multi-scale feature fusion network by leveraging spatial transformation and multi-scale feature pyramid networks via feature fusion block to enrich spatial and global semantic information. In particular, we take advantage of multi-scale feature pyramid network to incorporate global contextual information by employing atrous convolutions on top layers of convolutional neural network (CNN). And, the spatial transformation network is exploited on early layers of CNN to remove intra-class variability caused by spatial transformations. Further, the feature fusion block integrates both global contextual information and spatial features to encode the visual information of an input image. Moreover, spatial-semantic attention module is incorporated to learn attentive contextual features to guide the captioning module. The efficacy of the proposed model is evaluated on the COCO dataset.