Sparse Text Research Articles

Sex Discrimination and the Anti-Stereotyping Principle: Anuj Garg vs. Hotel Association of India

In this paper, I will examine sex discrimination under the Indian Constitution, through the lens of the judgment of the Supreme Court in Anuj Garg vs Hotel Association (2008). I will begin by excavating the sparse text of Article 15(1) of the Constitution, and highlighting the many potential meanings attributable to words such as “discriminate”, “grounds”, “only”, and “sex” (I). I will then examine the history of Indian sex discrimination jurisprudence in the six decades leading up to Anuj Garg. This history, I will argue, can broadly be divided into two intellectual traditions. I will call the first the “formal reading.” The formal reading justifies differential treatment of men and women on the basis of presumed “natural differences” between the sexes, and also holds that discrimination results only from conscious, hostile intentions of lawmakers. Set against this approach is what I call the “transformative reading”. The transformative reading rejects the notion of “natural difference”, and understands discrimination as a product of social, economic, and cultural structures and institutions that create “patterns of... exclusion”, rather than of individual hostile intentions. The judgment in Anuj Garg – which applied what I will define as “the anti-stereotyping principle” to sex discrimination claims – is the culmination of this second tradition (II). In the third section, I will trace the roots of this anti-stereotyping principle to the American suffrage movement, and the rejection of a social and political consensus that divided the world into two “separate spheres” – a public sphere to be occupied by men, and a private sphere that was the domain of women (III). I will then argue that this idea of separate spheres – with a few important differences – was prevalent in the colonial Indian polity as well (IV), and was itself challenged by the political movement for women’s right to vote, a movement whose vision triumphed with the framing of the Constitution (V). Consequently, Anuj Garg was correct in making anti-stereotyping, and a rejection of the conception of “natural differences”, the basis of its judgment. I will conclude by discussing the transformative potential of the reasoning in Anuj Garg for sex discrimination jurisprudence in the years to come (VI).

Short Text Classification Based on Latent Topic Modeling and Word Embedding

With the rapid development of the social network and e-commerce, we are exposed to enormous short text every day, ranging from twitters, movie comments, search snippets to news summaries. To classify the short and sparse text accurately is always the basic need for us to deal with information efficiently. However, previous methods fail to achieve high performance due to the sparseness and meaningless of the representation of text. The key breakout lies on the appropriate representation of the words, on which we excogitate a new framework. By discovering the latent topics in the related data crawled from the web, topic distribution can describe the text content in general. Combining with the word embedding generated from the online universal data, the proposed method is a more dense representation, containing semantic information from two different aspects. With this semantic representation of the texts, this framework greatly outperform the previous methods even using the most common SVM classifier, improving the accuracy by 11.58% on standard data set.

“I Hate You Everything”: Reading Adolescent Bad Feelings in Tamaki’s Skim

“I Hate You Everything”: Reading Adolescent Bad Feelings in Tamaki’s Skim Michelle Miller (bio) At the end of part 2 in Mariko and Jillian Tamaki’s graphic coming-of-age narrative Skim, a full-page panel shows tenth grade Kim Cameron (generally referred to as Skim) alone in a very large, snowy area.1 At the top of the page, the narration, which is drawn onto the scene as it is throughout the text, reads “Dear Diary.” Much further down, the diary entry continues, “It’s snowing.” In the bottom corner, there’s an inset panel with a photograph of a boy taken from a newspaper. On his forehead, someone has written “fag” with a marker. Readers have seen this photo before. It is of John Reddear, a local boy who has recently committed suicide. Reddear haunts the text—his image resurfaces at several points, as his ghostly presence is called up imaginatively and in both official and subversive curricula. While she never really knew him, having met him only once (she thinks), his life and death weigh heavily on Skim’s mind. [End Page 83] The photo is mounted on a large memorial bulletin board that hangs in the hallway of the private girls’ school Skim attends. The headline from the newspaper article accompanying the photo is “Teen Taken: too soon” (Tamaki and Tamaki 89). Throughout the book, this bulletin board, created and maintained by the “Girls Celebrate Life” club, sits at the centre of the conflicting atmospheres surrounding John’s death. The school is officially concerned with mourning John by “celebrating” life and investing in staying alive themselves. This official program of mourning, which involves participating in balloon-releasing ceremonies, watching Dead Poet’s Society, and taking part in public and semi-private grief exercises, is required of all students, whose compliance is monitored by teachers and counselors as well as popular students. Despite the intensity of the demand for compliance—or perhaps due to it—an oppositional discourse squeezes over and over into the school arena. This other discourse resists the requirement to have appropriately sad feelings about John and positive feelings about the future, joking instead that “death is cool” (77). The bulletin board, adamantly proclaimed to be “school property,” and therefore, ideally at least, immune to being undermined, offers students a site to express aggression, either passively, by sticking gum to it, or actively, by vandalizing or even destroying it. The homophobic inscription of “fag” is one such expression of aggression. In the larger scene, Skim walks a trail of letters—as long as she is tall—into the snow. This primary scene would be unseen by eyes on the ground: it’s only visible from our vantage above the action. We can thus read this as a deeply personal moment, in which Skim simultaneously expresses her feelings and keeps them obscured. Shuffling one foot in front of the other she inscribes the field with “I hate you,” then crosses out the “you” three times, replacing it with “everything.” We catch her before she finishes (she has written only “everythin”) and find her gazing back at her progress, even as we know she would be unable to make out her own words from her perspective. The co-presence of these separate but clearly related panels sharing the page signal that her school, social, and romantic/sexual situations are weighing on her. We can read the force of her hatred as directed at John, at the bulletin board containing this image, at the school hallways where the board hangs and where the unseen hand has inscribed him with “fag,” or at what readers know about Skim’s queer and complicated sexual life, which feels further complicated by the likelihood that this popular and supposedly well-adjusted boy’s suicide is attributed to the rumours that he is gay. [End Page 84] This simple panel, which temporally stretches through a quiet afternoon, arrests me every time I arrive at it. Like other quiet, expansive moments in this sparse text, the panel invites me to dwell inside of it: not only to read its surface but to feel and read its textures, to experience the time...

GMove: Group-Level Mobility Modeling Using Geo-Tagged Social Media.

Understanding human mobility is of great importance to various applications, such as urban planning, traffic scheduling, and location prediction. While there has been fruitful research on modeling human mobility using tracking data (e.g., GPS traces), the recent growth of geo-tagged social media (GeoSM) brings new opportunities to this task because of its sheer size and multi-dimensional nature. Nevertheless, how to obtain quality mobility models from the highly sparse and complex GeoSM data remains a challenge that cannot be readily addressed by existing techniques. We propose GMove, a group-level mobility modeling method using GeoSM data. Our insight is that the GeoSM data usually contains multiple user groups, where the users within the same group share significant movement regularity. Meanwhile, user grouping and mobility modeling are two intertwined tasks: (1) better user grouping offers better within-group data consistency and thus leads to more reliable mobility models; and (2) better mobility models serve as useful guidance that helps infer the group a user belongs to. GMove thus alternates between user grouping and mobility modeling, and generates an ensemble of Hidden Markov Models (HMMs) to characterize group-level movement regularity. Furthermore, to reduce text sparsity of GeoSM data, GMove also features a text augmenter. The augmenter computes keyword correlations by examining their spatiotemporal distributions. With such correlations as auxiliary knowledge, it performs sampling-based augmentation to alleviate text sparsity and produce high-quality HMMs. Our extensive experiments on two real-life data sets demonstrate that GMove can effectively generate meaningful group-level mobility models. Moreover, with context-aware location prediction as an example application, we find that GMove significantly outperforms baseline mobility models in terms of prediction accuracy.

Ensemble subspace clustering of text data using two-level features

This paper proposes a new integrated method for ensemble subspace clustering of high dimensional sparse text data. Our method employs two-level feature representation of text data (words and topics) to generate clusters from subspaces. We also use ensemble clustering to increase the robustness of the clusters. This method depends on topic modeling to get the two-level feature representation of text data and to generate different ensemble components. By using both topics and words to cluster text data, we can get more interpretable clusters as we can measure the weight of words and topics in each cluster. In order to evaluate the proposed method, we have conducted several experiments on seven real-life data sets. While some of these data sets are easy to cluster, others are hard, and some others contain unbalanced data. Experimental results on this diversity of data sets show that our method outperforms other methods for ensemble clustering.

Instance-Wise Denoising Autoencoder for High Dimensional Data

Denoising Autoencoder (DAE) is one of the most popular fashions that has reported significant success in recent neural network research. To be specific, DAE randomly corrupts some features of the data to zero as to utilize the cooccurrence information while avoiding overfitting. However, existing DAE approaches do not fare well on sparse and high dimensional data. In this paper, we present a Denoising Autoencoder labeled here as Instance-Wise Denoising Autoencoder (IDA), which is designed to work with high dimensional and sparse data by utilizing the instance-wise cooccurrence relation instead of the feature-wise one. IDA works ahead based on the following corruption rule: if an instance vector of nonzero feature is selected, it is forced to become a zero vector. To avoid serious information loss in the event that too many instances are discarded, an ensemble of multiple independent autoencoders built on different corrupted versions of the data is considered. Extensive experimental results on high dimensional and sparse text data show the superiority of IDA in efficiency and effectiveness. IDA is also experimented on the heterogenous transfer learning setting and cross-modal retrieval to study its generality on heterogeneous feature representation.

Naave Textualism in Patent Law

This symposium asks, “How much work does language do?” The answer these days is “too much.” Courts are letting statutory language do the work that used to be done by judges’ paying sensitive attention to context, history, policy, and background understandings. Or at least, they are apparently doing so — the even less appealing possibility is that courts are using statutory language as a cover for decisions reached on other grounds. I have long argued that part of the judicial function in statutory interpretation is to apply “background principles” of law, or “field-specific canons of construction.” Courts, in construing statutes, should — and do — discern the background principles of the area of law of which a statute is a part and interpret statutory text in light of them. Background principles of law frequently influence statutory interpretation, and in appropriate cases, the force of field-specific canons of construction may be so great as to cause courts to depart from apparently clear statutory text. Textualist interpreters, however, are pushing more and more in the direction of insistently following statutory text. Textualists are becoming increasingly radical, as they gradually realize that the accommodations they previously allowed in order to reach sensible results are inconsistent with fundamental textualist premises. This trend has resulted in the creation of a “naive textualism.” This mode of interpretation is not sharply differentiated from textualism per se, but is distinguished by its naive attitude that statutes can be best understood by simply looking up their words in a dictionary, applying a few canons of statutory construction, and eschewing other considerations. The Supreme Court recently provided an excellent example of its radical shift in the direction of naive textualism in the field of patent law. For decades — indeed, for centuries — patent law was a paradigm of richly contextualized judicial interpretation. Courts understood the sparse text of patent statutes in light of history, policy, and background understandings of the field of patent law. In the recent case of Bilski v. Kappos, however, the Supreme Court looked to little more than the dictionary in deciding fundamentally important questions under the patent statute. Bilski shows the dangers of language doing too much work.

Effective and Efficient Classification of Topically-Enriched Domain-Specific Text Snippets

Due to the explosive growth in the amount of text snippets over the past few years and their sparsity of text, organizations are unable to effectively and efficiently classify them, missing out on business opportunities. This paper presents TETSC: the Topically-Enriched Text Snippet Classification method. TETSC aims to solve the classification problem for text snippets in any domain. TETSC recognizes that there are different types of text snippets and, therefore, allows for stop word removal, named-entity recognition, and topical enrichment for the different types of text snippets. TETSC has been implemented in the production systems of a personal finance organization, which resulted in a classification error reduction of over 21%. Highlights: The authors create the TETSC method for classifying topically-enriched text snippets; the authors differentiate between different types of text snippets; the authors show a successful application of Named-Entity Recognition to text snippets; using multiple enrichment strategies appears to reduce effectivity.

TASC:Topic-Adaptive Sentiment Classification on Dynamic Tweets

Sentiment classification is a topic-sensitive task, i.e., a classifier trained from one topic will perform worse on another. This is especially a problem for the tweets sentiment analysis. Since the topics in Twitter are very diverse, it is impossible to train a universal classifier for all topics. Moreover, compared to product review, Twitter lacks data labeling and a rating mechanism to acquire sentiment labels. The extremely sparse text of tweets also brings down the performance of a sentiment classifier. In this paper, we propose a semi-supervised topic-adaptive sentiment classification (TASC) model, which starts with a classifier built on common features and mixed labeled data from various topics. It minimizes the hinge loss to adapt to unlabeled data and features including topic-related sentiment words, authors’ sentiments and sentiment connections derived from “@” mentions of tweets, named as topic-adaptive features. Text and non-text features are extracted and naturally split into two views for co-training. The TASC learning algorithm updates topic-adaptive features based on the collaborative selection of unlabeled data, which in turn helps to select more reliable tweets to boost the performance. We also design the adapting model along a timeline (TASC-t) for dynamic tweets. An experiment on 6 topics from published tweet corpuses demonstrates that TASC outperforms other well-known supervised and ensemble classifiers. It also beats those semi-supervised learning methods without feature adaption. Meanwhile, TASC-t can also achieve impressive accuracy and F-score. Finally, with timeline visualization of “river” graph, people can intuitively grasp the ups and downs of sentiments’ evolvement, and the intensity by color gradation.

Fuzzy partition based soft subspace clustering and its applications in high dimensional data

As one of the most popular clustering techniques for high dimensional data, soft subspace clustering (SSC) algorithms have been receiving a great deal of attention in recent years. Unfortunately, most existing works do not cluster high dimensional sparse data and noisy data in an effective manner. In this study, a novel soft subspace clustering algorithm called PI-SSC is proposed. By introducing a partition index (PI) into the objective function, a novel soft subspace clustering algorithm that combines the concepts of hard and fuzzy clustering is proposed. Furthermore, the robust property of PI-SSC is analyzed from the viewpoint of ε-insensitive distance. A convergence theorem for PI-SSC is also established by applying Zangwill’s convergence theorem. The results of the experiment demonstrate the effectiveness of the proposed algorithm in high dimensional sparse text data and noisy texture data.

Skew Estimation by Parts

This paper proposes a new part-based approach for skew estimation of document images. The proposed method first estimates skew angles on rather small areas, which are the local parts of characters, and subsequently determines the global skew angle by aggregating those local estimations. A local skew estimation on a part of a skewed character is performed by finding an identical part from prepared upright character images and calculating the angular difference. Specifically, a keypoint detector (e.g. SURF) is used to determine the local parts of characters, and once the parts are described as feature vectors, a nearest neighbor search is conducted in the instance database to identify the parts. Finally, a local skew estimation is acquired by calculating the difference of the dominant angles of brightness gradient of the parts. After the local skew estimation, the global skew angle is estimated by the majority voting of those local estimations, disregarding some noisy estimations. Our experiments have shown that the proposed method is more robust to short and sparse text lines and non-text backgrounds in document images compared to conventional methods.

Analyzing and Visualizing Web Opinion Development and Social Interactions With Density-Based Clustering

Due to the advancement of Web 2.0 technologies, a large volume of Web opinions is available on social media sites such as Web forums and Weblogs. These technologies provide a platform for Internet users around the world to communicate with each other and express their opinions. Analysis of developing Web opinions is potentially valuable for discovering ongoing topics of interests of the public like terrorist and crime detection, understanding how topics evolve together with the underlying social interaction between participants, and identifying important participants who have great influence in various topics of discussions. Nonetheless, the work of analyzing and clustering Web opinions is extremely challenging. Unlike regular documents, Web opinions are short and sparse text messages with noisy content. Typical document clustering techniques with the goal of clustering all documents applied to Web opinions produce unsatisfactory performance. In this paper, we investigated the density-based clustering algorithm and proposed the scalable distance-based clustering technique for Web opinion clustering. We conducted experiments and benchmarked with the density-based algorithm to show that the new algorithm obtains higher microaccuracy and macroaccuracy. This Web opinion clustering technique enables the identification of themes within discussions in Web social networks and their development, as well as the interactions of active participants. We also developed interactive visualization tools, which make use of the identified topic clusters to display social network development, the network topology similarity between topics, and the similarity values between participants.

A subspace decision cluster classifier for text classification

In this paper, a new classification method (SDCC) for high dimensional text data with multiple classes is proposed. In this method, a subspace decision cluster classification (SDCC) model consists of a set of disjoint subspace decision clusters, each labeled with a dominant class to determine the class of new objects falling in the cluster. A cluster tree is first generated from a training data set by recursively calling a subspace clustering algorithm Entropy Weighting k-Means algorithm. Then, the SDCC model is extracted from the subspace decision cluster tree. Various tests including Anderson–Darling test are used to determine the stopping condition of the tree growing. A series of experiments on real text data sets have been conducted. Their results show that the new classification method (SDCC) outperforms the existing methods like decision tree and SVM. SDCC is particularly suitable for large, high dimensional sparse text data with many classes.

Constrained Coclustering for Textual Documents

In this paper, we present a constrained co-clustering approach for clustering textual documents. Our approach combines the benefits of information-theoretic co-clustering and constrained clustering. We use a two-sided hidden Markov random field (HMRF) to model both the document and word constraints. We also develop an alternating expectation maximization (EM) algorithm to optimize the constrained co-clustering model. We have conducted two sets of experiments on a benchmark data set: (1) using human-provided category labels to derive document and word constraints for semi-supervised document clustering, and (2) using automatically extracted named entities to derive document constraints for unsupervised document clustering. Compared to several representative constrained clustering and co-clustering approaches, our approach is shown to be more effective for high-dimensional, sparse text data.

Skew Estimation of Document Images Using Bagging

This paper proposes a general-purpose method for estimating the skew angles of document images. Rather than to derive a skew angle merely from text lines, the proposed method exploits various types of visual cues of image skew available in local image regions. The visual cues are extracted by Radon transform and then outliers of them are iteratively rejected through a floating cascade. A bagging (bootstrap aggregating) estimator is finally employed to combine the estimations on the local image blocks. Our experimental results show significant improvements against the state-of-the-art methods, in terms of execution speed and estimation accuracy, as well as the robustness to short and sparse text lines, multiple different skews and the presence of nontextual objects of various types and quantities.

Skew detection for complex document images using robust borderlines in both text and non-text regions

A new skew detection method for complex document images based on robust borderlines extracted from both text and non-text regions is proposed in this paper. First, borderlines are extracted from the borders of large connected components in a document image by using a run length based method. Second, after filtering out non-linear borderlines, a fast iteration algorithm is applied to optimize each linear borderline’s directional angle. Finally, the weighted median value of all the directional angles is calculated as the skew angle of the whole document. Experiments on 2000 various skew document images are implemented. Total correct rate is 95.2%, and the detecting time on average is less than 0.2 s for each document. The proposed skew detection method is efficient for complex documents with horizontal and vertical text layout, three kinds of linguistic characters in English, Japanese and Chinese, especially for documents with predominant non-text regions or sparse text regions.

A Text Clustering System based on k-means Type Subspace Clustering and Ontology

This paper presents a text clustering system developed based on a k-means type subspace clustering algorithm to cluster large, high dimensional and sparse text data. In this algorithm, a new step is added in the k-means clustering process to automatically calculate the weights of keywords in each cluster so that the important words of a cluster can be identified by the weight values. For understanding and interpretation of clustering results, a few keywords that can best represent the semantic topic are extracted from each cluster. Two methods are used to extract the representative words. The candidate words are first selected according to their weights calculated by our new algorithm. Then, the candidates are fed to the WordNet to identify the set of noun words and consolidate the synonymy and hyponymy words. Experimental results have shown that the clustering algorithm is superior to the other subspace clustering algorithms, such as PROCLUS and HARP and kmeans type algorithm, e.g., Bisecting-KMeans. Furthermore, the word extraction method is effective in selection of the words to represent the topics of the clusters. Keywords—Subspace Clustering, Text Mining, Feature Weighting, Cluster Interpretation, Ontology

“I talk them through it”: Teacher mediation of picturebooks with sparse verbal text during whole‐class readalouds

Picturebook readalouds have been promoted as a rich resource for strengthening literacy development in young children and discussions of picturebooks have become a common ‐ if threatened — practice in the primary grades. One subset of the diverse body of children's literature teachers can draw from is picturebooks with sparse verbal text. Such picturebooks may place additional demands on both teacher and student during whole class readalouds, who in their responses must attend even more closely to the print that is there as well as to the illustrations and other visual features of the books. The present study explores the particular “affordances” (Gibson, 1950) of such texts and how one teacher maximizes these potentials as she orchestrates three readaloud experiences in an urban kindergarten classroom.

A Colour Atlas of Posterior Chamber Implants

This atlas richly illustrates one clinical approach to extracapsular cataract extraction and posterior chamber lens implantation. Numerous high-quality color photographs, clearly reproduced and well organized, take the reader through each step in the operation. The most useful photographs are those depicting potential complications as they begin and methods to avoid them. Seven well-known extracapsular surgeons contributed photographs and several chapters to this text. outstanding is a concise chapter by David McIntyre illustrating his manual technique, which compares and contrasts with Lim's Particularly outstanding is a concise chapter by David McIntyre illustrating his manual technique. The sparse text accompanying the photographs is elementary and repetitious. In addition, the few references listed at the end of the book are not cited within the text. Complications are mentioned, but management of them is incompletely discussed and illustrated. Only a few alternative surgical techniques are presented; an established extracapsular surgeon would be