The Use of Ateji as Figurative Speech in Japanese Song Lyrics

This study aims to analyze the types of ateji used in Japanese song lyrics, and to analyze ateji as a form of figurative speech by the semantic-semiotic connection between the words used in ateji’s writing and reading. Analysis is done qualitatively based on Shirose’s theory of ateji classification and Japan’s hiyu hyougen (figurative speech) theory. This research resulted in the finding of 4 ateji types in Japanese song lyrics, which are ateji for foreign words pronunciation, ateji for pronouns, ateji for replacing words, and ateji for words used in specific titles/works. The use of metaphor (in’yu), metonymy (kan’yu), and synecdoche (teiyu) are also found between the uses of ateji, based on the relation of the words in said ateji. The words can be linked through contextual relation, conceptual relation, or semantical relation. The connection of the words can also result in similar uses as other figurative speeches not included in Japanese’s hiyu hyougen, which indicated that ateji can be handled and understood as a general form of figurative speech in written Japanese language.

Inductive Contextual Relation Learning for Personalization

Web personalization, e.g., recommendation or relevance search, tailoring a service/product to accommodate specific online users, is becoming increasingly important. Inductive personalization aims to infer the relations between existing entities and unseen new ones, e.g., searching relevant authors for new papers or recommending new items to users. This problem, however, is challenging since most of recent studies focus on transductive problem for existing entities. In addition, despite some inductive learning approaches have been introduced recently, their performance is sub-optimal due to relatively simple and inflexible architectures for aggregating entity’s content. To this end, we propose the inductive contextual personalization (ICP) framework through contextual relation learning. Specifically, we first formulate the pairwise relations between entities with a ranking optimization scheme that employs neural aggregator to fuse entity’s heterogeneous contents. Next, we introduce a node embedding term to capture entity’s contextual relations, as a smoothness constraint over the prior ranking objective. Finally, the gradient descent procedure with adaptive negative sampling is employed to learn the model parameters. The learned model is capable of inferring the relations between existing entities and inductive ones. Thorough experiments demonstrate that ICP outperforms numerous baseline methods for two different applications, i.e., relevant author search and new item recommendation.

"I am either all the way there or I am not. I don't know how to walk this middle line" (Skorczewski 109). The Impact of Manic and Depressive States on Anne Sexton's Poetic Depictions of Food.

This research explores poet Anne Sexton’s use of food imagery in To Bedlam and Part Way Back, focusing on the way food is presented in contextual relation to the manic-depressive cycles Sexton was experiencing. This is placed within the socio-political climate of the 1950s-1960s which influences Sexton’s relationship with food.

Walking-sensing as a decolonial art and pedagogical practice

How can walking, as a sensate experience and a recollective engagement with our memories, lead us to imagine new ways of knowing, being and sensing otherwise? This article conceptualizes walking-sensing as a decolonial art and pedagogical practice, which offers anti-colonial critiques and activates decolonial imaginations. By combining walking and sensing together, I first highlight how our experience of walking is intrinsically intertwined with our act of sensing that is already oriented and attuned in the contextual relation to things in the world. The notion of walking-sensing is used to describe not only our physical movement and the sensibilities of our bodies, but also as recollective and communal engagements, such as connecting memories with others, (re)collecting personal and local stories, and imagining the ways of living and being otherwise. I further elucidate how walking-sensing can be a form of anti-colonial critiques and decolonial imaginations that valorize multiple knowledges and sensibilities as well as to pave a path towards a new liberatory way of being in solidarity. With pedagogical scenarios, I demonstrate in what ways walking-sensing can be utilized as a critical intervention towards decoloniality. Lastly, introduce two artists’ art-making practice and how they are linked to the concept of walking-sensing. In this way, I elucidate the inextricable relationship between art and pedagogical practice and how walking-sensing can lead to decolonial resistance.

Landscaping an Ecuadorian neighborhood in Queens, NY

Diaspora studies on migration communities have shown how these transnational groups appropriate the new space in the receptive country through a process of deterritorialization and reterritorialization. These processes involve a reinterpretation and reconceptualization of the linguistic relationship between the language of the diaspora group and those spoken in the new home. One of the most visible places where this contextual relation must be negotiated is in the public sphere, where language, culture and identity are inevitably interwoven (Blackwood, Lanza & Woldemariam, 2016). Using a multimodal approach and using both qualitative and quantitative methodologies, this article analyzes how an Ecuadorian-American community in Queens transformed the linguistic landscape of their surroundings to make it similar to what this community had in their home country.

Relation-Aware Pedestrian Attribute Recognition with Graph Convolutional Networks

In this paper, we propose a new end-to-end network, named Joint Learning of Attribute and Contextual relations (JLAC), to solve the task of pedestrian attribute recognition. It includes two novel modules: Attribute Relation Module (ARM) and Contextual Relation Module (CRM). For ARM, we construct an attribute graph with attribute-specific features which are learned by the constrained losses, and further use Graph Convolutional Network (GCN) to explore the correlations among multiple attributes. For CRM, we first propose a graph projection scheme to project the 2-D feature map into a set of nodes from different image regions, and then employ GCN to explore the contextual relations among those regions. Since the relation information in the above two modules is correlated and complementary, we incorporate them into a unified framework to learn both together. Experiments on three benchmarks, including PA-100K, RAP, PETA attribute datasets, demonstrate the effectiveness of the proposed JLAC.