Product Representation Research Articles

DeepEMPR: coffee leaf disease detection with deep learning and enhanced multivariance product representation

Plant diseases threaten agricultural sustainability by reducing crop yields. Rapid and accurate disease identification is crucial for effective management. Recent advancements in artificial intelligence (AI) have facilitated the development of automated systems for disease detection. This study focuses on enhancing the classification of diseases and estimating their severity in coffee leaf images. To do so, we propose a novel approach as the preprocessing step for the classification in which enhanced multivariance product representation (EMPR) is used to decompose the considered image into components, a new image is constructed using some of those components, and the contrast of the new image is enhanced by applying high-dimensional model representation (HDMR) to highlight the diseased parts of the leaves. Popular convolutional neural network (CNN) architectures, including AlexNet, VGG16, and ResNet50, are evaluated. Results show that VGG16 achieves the highest classification accuracy of approximately 96%, while all models perform well in predicting disease severity levels, with accuracies exceeding 85%. Notably, the ResNet50 model achieves accuracy levels surpassing 90%. This research contributes to the advancement of automated crop health management systems.

Enhancing Accessibility in Online Shopping: A Dataset and Summarization Method for Visually Impaired Individuals

A visually impaired individual (VI) encounters numerous challenges in their daily activities, particularly in tasks reliant on visual systems such as navigation, educational pursuits, and shopping. Online shopping poses a heightened difficulty due to its reliance on visual representations of products in digital formats. The impact of visual impairment on product selection based on reviews remains inadequately investigated. This study endeavors to address two primary objectives. Firstly, we propose the creation of a dataset comprising product review videos (referred to as PVS10) tailored for visually impaired individuals. Secondly, we present a foundational summarization methodology designed to facilitate access to pertinent and informative content within extensive video collections for visually impaired individuals. Our dataset, gathered from YouTube, encompasses 10 distinct products, each associated with the top 10 review videos, totaling 100 videos of varying lengths. Utilizing the search term "review videos of PRODUCT NAME", we assembled the dataset to facilitate automated summarization processes aimed at maximizing salient information, minimizing redundant content, and preserving the overarching sentiment conveyed in the reviews. This research focuses on the challenges faced by visually impaired people in online shopping, particularly when selecting products based on customer reviews. Our study demonstrates that people with visual impairments may actively explore product reviews and only acquire the information they require.

A Fourth-Order Tensorial Wiener Filter Using the Conjugate Gradient Method

The recently developed iterative Wiener filter using a fourth-order tensorial (FOT) decomposition owns appealing performance in the identification of long length impulse responses. It relies on the nearest Kronecker product representation (with particular intrinsic symmetry features), together with low-rank approximations. Nevertheless, this new iterative filter requires matrix inversion operations when solving the Wiener–Hopf equations associated with the component filters. In this communication, we propose a computationally efficient version that relies on the conjugate gradient (CG) method for solving these sets of equations. The proposed solution involves a specific initialization of the component filters and sequential connections between the CG cycles. Different FOT-based decomposition setups are also analyzed from the point of view of the resulting parameter space. Experimental results obtained in the context of echo cancellation confirm the good behavior of the proposed approach and its superiority in comparison to the conventional Wiener filter and other decomposition-based versions.

Dignified Branding: Upholding Authenticity and Integrity in Product Representation

Dignified branding is a concept that emphasizes authenticity, respect for a product's inherent qualities, and transparent communication with consumers. This research paper explores the principles of dignified branding, contrasting it with fake branding practices characterized by deceit and exaggeration. By analyzing the implications of each approach, this study highlights the importance of maintaining integrity in brand representation. It posits that dignified branding fosters trust, loyalty, and sustainable growth, while fake branding undermines consumer confidence and damages brand reputation. The paper concludes that organizations should prioritize dignified branding to create meaningful connections with consumers and uphold their brand's dignity.

The Design of Fast Type-V Discrete Cosine Transform Algorithms for Short-Length Input Sequences

Fast algorithms for type-five discrete cosine transform (DCT-V) for sequences of input data of short length in the range of two to eight are elaborated in the paper. A matrix–vector product representation of the DCT-V is the starting point for designing the algorithms. In each specific case, the DCT-V matrices have remarkable structural properties that follow from the localization of identical entries within the matrices. Each matrix of the DCT-V has only a few distinct entries that are repeated at different positions in its structure. Using simple transformations such as permutations of the rows and/or columns of this matrix or its favorable decomposition into two or more matrix components, it is possible to obtain efficient matrix structures that lead to useful factorization schemes. Based on the suitable factorization schemes we obtained, we developed fast algorithms that reduce the number of arithmetic operations when calculating the DCT-V. The correctness of the obtained algorithmic solutions was justified theoretically using a strict mathematical background of each of them. The developed algorithms were then further tested using MATLAB R2023b software to finally confirm their correctness. Finally, an evaluation of the computational complexity for each obtained solution is presented. The evaluation results were compared with the computational complexity of the direct calculation of matrix–vector products. The resulting factorizations of the matrices of the DCT-V reduce the average number of multiplications by 57% but increase the number of additions by 29%.

Keywords-enhanced Contrastive Learning Model for travel recommendation

Travel recommendation aims to infer travel intentions of users by analyzing their historical behaviors on Online Travel Agencies (OTAs). However, crucial keywords in clicked travel product titles, such as destination and itinerary duration, indicating tourists’ intentions, are often overlooked. Additionally, most previous studies only consider stable long-term user interests or temporary short-term user preferences, making the recommendation performance unreliable. To mitigate these constraints, this paper proposes a novel Keywords-enhanced Contrastive Learning Model (KCLM). KCLM simultaneously implements personalized travel recommendation and keywords generation tasks, integrating long-term and short-term user preferences within both tasks. Furthermore, we design two kinds of contrastive learning tasks for better user and travel product representation learning. The preference contrastive learning aims to bridge the gap between long-term and short-term user preferences. The multi-view contrastive learning focuses on modeling the coarse-grained commonality between clicked products and their keywords. Extensive experiments are conducted on two tourism datasets and a large-scale e-commerce dataset. The experimental results demonstrate that KCLM achieves substantial gains in both metrics compared to the best-performing baseline methods. Specifically, HR@20 improved by 5.79%–14.13%, MRR@20 improved by 6.57%–18.50%. Furthermore, to have an intuitive understanding of the keyword generation by the KCLM model, we provide a case study for several randomized examples.

Sensitivity of Arctic Clouds to Ice Microphysical Processes in the NorESM2 Climate Model

Abstract Ice formation remains one of the most poorly represented microphysical processes in climate models. While primary ice production (PIP) parameterizations are known to have a large influence on the modeled cloud properties, the representation of secondary ice production (SIP) is incomplete and its corresponding impact is therefore largely unquantified. Furthermore, ice aggregation is another important process for the total cloud ice budget, which also remains largely unconstrained. In this study, we examine the impact of PIP, SIP, and ice aggregation on Arctic clouds, using the Norwegian Earth System Model, version 2 (NorESM2). Simulations with both prognostic and diagnostic PIP show that heterogeneous freezing alone cannot reproduce the observed cloud ice content. The implementation of missing SIP mechanisms (collisional breakup, drop shattering, and sublimation breakup) in NorESM2 improves the modeled ice properties, while improvements in liquid content occur only in simulations with prognostic PIP. However, results are sensitive to the description of collisional breakup. This mechanism, which dominates SIP in the examined conditions, is very sensitive to the treatment of the sublimation correction factor, a poorly constrained parameter that is included in the utilized parameterization. Finally, variations in ice aggregation treatment can also significantly impact cloud properties, mainly through their impact on collisional breakup efficiency. Overall, enhancement in ice production through the addition of SIP mechanisms and the reduction in ice aggregation (in line with radar observations of shallow Arctic clouds) result in enhanced cloud cover and decreased TOA radiation biases, compared to satellite measurements, especially during the cold months. Significance Statement Arctic clouds remain a large source of uncertainty in projections of the future climate due to the poor representation of the microphysical processes that govern their life cycle. Ice formation is among the least understood processes. While it is widely recognized that better constraints on primary ice production (PIP) are needed to improve existing parameterizations, we show that secondary ice production (SIP) and ice aggregation can have also a significant impact on ice number concentrations. Constraining ice formation through the addition of missing SIP mechanisms and reducing ice aggregation can improve the representation of the cloud macrophysical properties and enhance total cloud cover in the Arctic region, which in turn contributes to decreased TOA radiation biases in the cold months.

Breaking the Boundary between Permanent Capital Farmland and Arable Land in China: Understanding State and Drivers of Permanent Capital Farmland Non-Grain Production in a Rapid Urbanizing County

Amid unprecedented challenges to protect arable land, China’s permanent capital farmland (PCF) has played a crucial role in grain production. However, a clear legal and physical boundary between PCF and arable land seems to be unable to stop the spread of non-grain production. To address it, an analysis framework for PCF non-grain production was developed to examine the state and drivers of village-scale PCF non-grain production based on the logical relationship between PCF and arable land in the rapid urbanization of Liyang. The results suggested that PCF comprised approximately 70% arable land and 30% adjustable land. Meanwhile, forest land and aquaculture ponds occupied over 25% of PCF, while nearly 20% of PCF is unsuitable for the resumption of crop cultivation. The transition state (scenario SR) offered a realistic representation of PCF non-grain production, with an average non-grain production of 48.88%. This is 14.00% lower than the current state (scenario SD) and 9.65% higher than the future state (scenario ST). Furthermore, PCF area and agricultural income per capita significantly encouraged PCF non-grain production, with explanatory powers of 51.60% and 42.40%, respectively. In contrast, urbanization rate (with an explanatory power of 35.30%) significantly discouraged it. Therefore, this paper proposed PCF redefinition, flexible PCF, and diversified economic incentives to mitigate PCF non-grain production at the village scale.

Fostering Students’ Development of Productive Representation Systems for Infinite Series

Fostering Students’ Development of Productive Representation Systems for Infinite Series

Association change and its relationship with designs in context and form: how product evaluation is affected by negative framings

This study explores the interplay between internal factors (e.g. personal values) and external factors (e.g. product framing and representation form) in shaping user perceptions, attitudes and behaviours towards products. A between-participants experiment (N = 52) was designed using deliberately an iconic product (Haribo gummy bears) to be presented either as image or a tangible item, as well as framed either positively or negatively. The aim is to examine the impact of these manipulated internal and external factors on users’ evaluations and their subsequent behaviour. The effects of framing, and markedly negative framing, were considerably larger than those of forms of representations, whose role was limited for the studied product. However, the effect of framing on evaluations is amplified after interacting with the real product. Despite its significance, negative product framing cannot change behaviour as much as user personal values. These insights suggest designers pay attention to how products are portrayed to potential users. Particular attention should be paid to aligning products with the personal values of the target users since this turned as the main determinant for product appreciation. Hence, this research urges designers to pay attention to the fact that products have to resonate with users so to enhance their acceptance.

Skip-Gram and Transformer Model for Session-Based Recommendation

Session-based recommendation uses past clicks and interaction sequences from anonymous users to predict the next item most likely to be clicked. Predicting the user’s subsequent behavior in online transactions becomes a problem mainly due to the lack of user information and limited behavioral information. Existing methods, such as recurrent neural network (RNN)-based models that model user’s past behavior sequences and graph neural network (GNN)-based models that capture potential relationships between items, miss different time intervals in the past behavior sequence and can only capture certain types of user interest patterns due to the characteristics of neural networks. Graphic models created to improve the current session reduce the model’s success due to the addition of irrelevant items. Moreover, attention mechanisms in recent approaches have been insufficient due to weak representations of users and products. In this study, we propose a model based on the combination of skip-gram and transformer (SkipGT) to solve the above-mentioned drawbacks in session-based recommendation systems. In the proposed method, skip-gram both captures chained user interest in the session thread through item-specific subreddits and learns complex interaction information between items. The proposed method captures short-term and long-term preference representations to predict the next click with the help of a transformer. The transformer in our proposed model overcomes many limitations in turn-based models and models longer contextual connections between items more effectively. In our proposed model, by giving the transformer trained item embeddings from the skip-gram model as input, the transformer has better performance because it does not learn item representations from scratch. By conducting extensive experiments with three real-world datasets, we confirm that SkipGT significantly outperforms state-of-the-art solutions with an average MRR score of 5.58%.

Panorama de pesquisas sobre ecossistemas de inovação:

Innovation ecosystems are composed of interconnected and interdependent network actors that shape scientific, technological and political-economic dynamics with high transformative capacity related to information and innovation. The study analyzes the panorama of research on “innovation ecosystems”, as a theme produced and communicated in scientific literature. As an exploratory and descriptive research, it uses the Dimensions database and VOSviewer network analysis software, for the collection, representation and inferential analysis of production and connection indicators on this topic. It demonstrates the exponential growth of publications on innovation ecosystems, its thematic coverage and its scope of emerging research highlighting themes such as technological transfer, digital transformation, internet, regional ecosystem and open innovation. Innovation ecosystems are configured as a transversal theme that allows for multilateral and complex theoretical approaches due to the broad and profound possibility of connecting other themes, concepts, contents and contexts of social production.

Fotografia e cidadania: prática comunicacional em Curralinho – Marajó

This study departs from a multimedia workshop of the ÀWÚRE Ribeirinhos Project in Curralinho, in the region of Marajó, Pará, Brazil. After it, academic restlessness arose to reflect on what would be the perceptions of the young people of Curralinho about the municipality and their daily lives based on the photographic essay they produced to promote active participation and awareness of citizenship in the construction of their own narratives. This study proposed to understand how young people portray their local reality by photography, exploring the themes, perspectives, and image approaches in the produced material. This research also seeks to investigate the influence of the geographical, social, and cultural context in these visual production representations and identify the difficulties this population faces and their expression by photographs.

Direct Retrieval of Orthographic Representations in Chinese Handwritten Production: Evidence from a Dynamic Causal Modeling Study.

This present study identified an optimal model representing the relationship between orthography and phonology in Chinese handwritten production using dynamic causal modeling, and further explored how this model was modulated by word frequency and syllable frequency. Each model contained five volumes of interest in the left hemisphere (angular gyrus [AG], inferior frontal gyrus [IFG], middle frontal gyrus [MFG], superior frontal gyrus [SFG], and supramarginal gyrus [SMG]), with the IFG as the driven input area. Results showed the superiority of a model in which both the MFG and the AG connected with the IFG, supporting the orthography autonomy hypothesis. Word frequency modulated the AG → SFG connection (information flow from the orthographic lexicon to the orthographic buffer), and syllable frequency affected the IFG → MFG connection (information transmission from the semantic system to the phonological lexicon). This study thus provides new insights into the connectivity architecture of neural substrates involved in writing.

Star exponentials from propagators and path integrals

In this paper we address the relation between the star exponentials emerging within the Deformation Quantization formalism and Feynman’s path integrals associated with propagators in quantum dynamics. In order to obtain such a relation, we start by visualizing the quantum propagator as an integral transform of the star exponential by means of the symbol corresponding to the time evolution operator and, thus, we introduce Feynman’s path integral representation of the propagator as a sum over all the classical histories. The star exponential thus constructed has the advantage that it does not depend on the convergence of formal series, as commonly understood within the context of Deformation Quantization. We include some basic examples to illustrate our findings, recovering standard results reported in the literature. Further, for an arbitrary finite dimensional system, we use the star exponential introduced here in order to find a particular representation of the star product which may be recognized as the one encountered in the context of the quantum field theory for a Poisson sigma model.

Transnational flows of living cultural heritage: African screen media in the world

Transnational flows of living cultural heritage: African screen media in the world Dr. Sheila Petty, FRSC, Professor of Media Studies and SaskPower Research Chair in Cultural Heritage, at the University of Regina, explores transnational movements of living cultural heritage in African screen media. My research program investigates transnational methodologies in the expression and safeguarding of intangible and tangible cultural heritage (screen media). It proposes a reflection on current debates around identity politics in globalizing cultures, political and cultural sovereignty, control of representation in image production, and modes of resistance to colonialism in African screen media.

Packaging of disposable vaping products and e-liquids in England, Canada and the United States: A content analysis.

Vaping product packaging is varied and often features bright colours and novel designs, particularly among recently marketed disposable vapes. This study provides an overview of attributes found on the packaging of popular disposable vapes and e-liquid bottles in England, Canada and the United States (US) and assesses compliance with local packaging regulations. Content analysis. Brick-and-mortar and online shops in England (London), Canada (Ontario) and the US (New Hampshire and South Carolina). 108 vaping products (including packaging) from 76 brands in a range of flavours and nicotine levels. Specifically, 48 disposable vapes (15 from England, 16 from Canada, 17 from the US) and 60 e-liquid bottles (20 per country). Textual and graphic branding and marketing elements, independently coded by two researchers and checked by a third. Compliance with local packaging regulations varied across countries. Health warnings were present on the packaging of all but one nicotine-containing product, although 33% of disposables and 17% of e-liquids featuring the warning did not adhere to formatting requirements. Leaflets were seldom included with e-liquid bottles, even in England (45%) where mandatory, and omitted elsewhere. Labelling of nicotine type and batch numbers was inconsistent. Vaping product packaging featured claims relating to sensory perceptions (41%), most often flavours, and some (32%) featured youth-appealing content. Common graphic elements included stylised brand fonts (80%), brand logos (54%), product representations on the external packaging (47%) and abstract graphic elements (64%). Colours on packaging, disposable vapes and e-liquid bottles were associated with product flavour. In England, Canada and the United States, popular disposable vapes and e-liquid bottles appear to have varying compliance with local packaging regulations and inconsistent labelling of nicotine and product characteristics. The use of colourful designs, evocative descriptors and appealing graphics to promote flavours underscores the need for comprehensive packaging regulations and enforcement.

Ontological model of structure and parameters of components in high-tech product modernization projects

The subject of the article is models representing the structure and parameters of components in high-tech products based on the analysis of technical documentation in modernization projects. The purpose of the proposed research is to enhance the quality of modernization processes of high-tech products by forming an ontological model of the product with innovative components, considering the diversity of information support. The article addresses the following tasks: studying the main methods and technologies of system representation of complex product structure; forming a model of structural-functional decomposition of a high-tech product, creating an ontological model of the structure and parameters of a high-tech product based on technical documentation. The following methods are applied: systemic approach, methods of functional-structural decomposition, set theory, ontology construction methods, semantic models. The following results were obtained: The main methods of system representation of complex product structure were investigated, based on the following principles: decomposition of complex product architecture, layering of complex product representation, multivariate synthesis of component architecture. The main directions and advantages of using 3D technologies for solving design tasks in modernization projects of high-tech product components are considered. The decomposition of a high-tech product into component parts and partial parameters is proposed, taking into account functional, structural, and parametric characteristics. An ontological model of the structure and parameters of a high-tech product is formed based on a set of technical documentation and considering additional heterogeneous sources of information. Innovative elements are identified, the description of which may be fuzzy. Conclusions. The proposed ontological model can serve as a basis for finding similar solutions regarding the design of innovative components in the precedent database, which is properly structured. In the absence of similar design solutions, innovative component design can be carried out using 3D technologies based on supplementing fuzzy information in the semantic model.

Embedding-based retrieval: measures of threshold recall and precision to evaluate product search

Modern product retrieval systems are becoming increasingly complex due to the use of extra product representations, such as user behavior, language semantics and product images. However, adding new information and complicating machine learning models does not necessarily lead to an improvement in online and business search performance, since after retrieval the product list is ranked, which introduces its own bias. Nevertheless, the business performance of a product search will be worse from ranking an incomplete list of products than a complete one, and the relevance of search results will not improve from perfect sorting of products that do not match the search query. Therefore, the main quality indicators for the products retrieval phase remain Recall and Precision at the k threshold. This paper compares several architectures of product retrieval systems in product search for e-commerce. To do this, the concepts of threshold Recall and Precision for information retrieval are investigated and the dependence of these measures on the order of issuance is revealed. An automatic procedure has been developed for calculating R@k and P@k, which allows us to compare the effectiveness of information retrieval systems. The proposed automatic procedure has been tested on the WANDS public dataset for several key architectures. The obtained values R@1000 = 84% ± 9% and P@10 = 67% ± 17% are at the level of SOTA models.

Study on the Factors Influencing the Purchase Intention of Celebrity Endorsements

In tandem with market evolution, advertising strategies within enterprises are progressively gaining ubiquity, and the phenomenon of celebrity product representation is becoming increasingly pervasive. Instances of celebrities endorsing products are commonplace in daily life, with around 50% of the top 100 global brands adopting endorsement strategies. In the contemporary era, under the influence of the “fan economy”, the distinct characteristics of renowned endorsers and their fan base exert a direct influence on consumers, particularly within the general consumer demographic’s proclivity towards products endorsed by celebrities.This research employed the SPSS tool to empirically validate hypotheses, probing into the multifaceted impact of various factors associated with renowned endorsers on the purchase intentions of consumer groups within the backdrop of the “fan economy”. Commencing with three dimensions of celebrity endorser allure, expertise, and reliability, and three dimensions involving consumer congruence with renowned products, familiarity, and peer opinions, this study meticulously constructed a comprehensive relationship model delineating the nexus between celebrity endorsements and consumer purchase intentions. The research endeavors to deepen insights into the market strategies entwined with celebrity-endorsed products. From the vantage point of the seller, it furnishes nuanced recommendations tailored to brands, aspiring to equip enterprises with effective means to leverage celebrity advertising for heightened market share and sales.