152 research outputs found
RS_GV at SemEval-2021 Task 1: Sense Relative Lexical Complexity Prediction
No full textStodden R, Venugopal G. RS_GV at SemEval-2021 Task 1: Sense Relative Lexical Complexity Prediction. In: Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021). Stroudsburg, PA, USA: Association for Computational Linguistics; 2021: 640-649.We present the technical report of the system called RS_GV at SemEval-2021 Task 1 on lexical complexity prediction of English words. RS_GV is a neural network using hand-crafted linguistic features in combination with character and word embeddings to predict target words’ complexity. For the generation of the hand-crafted features, we set the target words in relation to their senses. RS_GV predicts the complexity well of biomedical terms but it has problems with the complexity prediction of very complex and very simple target words
Entrepreneurial marketing in subsistence marketplaces
No full textThere are more than a billion poverty-stricken entrepreneurs in the world who run micro-enterprises to meet basic consumption needs. This pervasive phenomenon presents an interesting theoretical conundrum - that of consumer-entrepreneur duality. This duality blurs the boundaries between consumption and entrepreneurship, which have traditionally been distinct domains of scholarly inquiry. The research reported in this dissertation aims to a) provide a theoretical foundation for the notion of consumer-entrepreneur duality and b) test the implications of the aforementioned duality empirically. A key insight flowing from the investigations is that factors in the consumption domain impact important outcomes in the entrepreneurial domain and vice versa.Submission published under a 24 month embargo labeled 'Closed Access', the embargo will last until 2018-05-01The student, Srinivas Venugopal, accepted the attached license on 2016-04-18 at 08:47.The student, Srinivas Venugopal, submitted this Dissertation for approval on 2016-04-18 at 09:04.This Dissertation was approved for publication on 2016-04-19 at 08:14.DSpace SAF Submission Ingestion Package generated from Vireo submission #9286 on 2016-07-07 at 14:17:05Made available in DSpace on 2016-07-07T21:17:37Z (GMT). No. of bitstreams: 4
VENUGOPAL-DISSERTATION-2016.pdf: 1854109 bytes, checksum: f8e3d9c290a0109c220b8b0fc51c60c1 (MD5)
SrinivasVenugopal-DissertationApr17-Final.docx: 5793780 bytes, checksum: b6083d1e61eed44327c6ca98d0843dc4 (MD5)
LICENSE.txt: 4215 bytes, checksum: 3a0d71a95b961c52e415358c38df4270 (MD5)
PROQUEST_LICENSE.txt: 4561 bytes, checksum: 191925090206f5324017b16a1d5401bd (MD5)
Previous issue date: 2016-04-19Embargo set by: Seth Robbins for item 93274
Lift date: 2018-07-07T21:18:16Z
Reason: Author requested closed access (OA after 2yrs) in Vireo ETD systemLimited Restriction Lifted for Item 93274 on 2018-07-08T09:15:20Z
Aggregate breakdown of nanoparticulate titania
No full textSix nanosized titanium dioxide powders synthesized from a sulfate process were investigated. The targeted end-use of this powder was for a de-NOx catalyst honeycomb monolith. Alteration of synthesis parameters had resulted principally in differences in soluble ion level and specific surface area of the powders. The goal of this investigation was to understand the role of synthesis parameters in the aggregation behavior of these powders. Investigation via scanning electron microscopy of the powders revealed three different aggregation iterations at specific length scales.
Secondary and higher order aggregate strength was investigated via oscillatory stress rheometry as a means of simulating shear conditions encountered during extrusion. G' and G'' were measured as a function of the applied oscillatory stress. Oscillatory rheometry indicated a strong variation as a function of the sulfate level of the particles in the viscoelastic yield strengths. Powder yield stresses ranged from 3.0 Pa to 24.0 Pa of oscillatory stress. Compaction curves to 750 MPa found strong similarities in extrapolated yield point of stage I and II compaction for each of the powders (at approximately 500 MPa) suggesting that the variation in sulfate was greatest above the primary aggregate level. Scanning electron microscopy of samples at different states of shear in oscillatory rheometry confirmed the variation in the linear elastic region and the viscous flow regime.
A technique of this investigation was to approach aggregation via a novel perspective: aggregates are distinguished as being loose open structures that are highly disordered and stochastic in nature. The methodology used was to investigate the shear stresses required to rupture the various aggregation stages encountered and investigate the attempt to realign the now free-flowing constituents comprising the aggregate into a denser configuration. Mercury porosimetry was utilized to measure the pore size of the compact resulting from compaction via dry pressing and tape casting secondary scale aggregates. Mercury porosimetry of tapes cast at 0.85 and 9.09 cm/sec exhibited pore sizes ranging from 200-500 nm suggesting packing of intact micron-sized primary aggregates. Porosimetry further showed that this peak was absent in pressed pellets corroborating arguments of ruptured primary aggregates during compaction to 750 MPa.Ph.D.Includes bibliographical references (p. 166-170)
Study of a Pedagogy Adopted to Generate Interest in Students Taking a Programming Course
No full textHindi Complex Word Feature Dataset
No full textThis dataset was created by conducting a human intelligence test, wherein native and non-native Hindi speakers annotated words they could not understand in Hindi text. They were then asked to rank the complexity of these words along with their synonyms. A word that received an average rank of 3 is labeled 0. 1 indicates complex and 0 indicates simple
From clicks to quick bites—social media engagement and logistics efficiency as drivers of fast-food consumption in the United Arab Emirates
No full textIntroductionThe fast-food industry in the United Arab Emirates has expanded rapidly due to urbanization, a digitally active young population, and extensive use of social media for consumer engagement. Fast-food brands increasingly rely on digital touchpoints to influence purchase decisions, brand loyalty, and post-purchase satisfaction. This study examines how social media marketing attributes, consumer engagement, logistics efficiency, and sales platform usability affect fast-food purchase behavior among young adults in the UAE, with specific reference to major fried chicken and burger chains such as McDonald’s, KFC, and Hardee’s.MethodsThe study follows a positivist research philosophy with a deductive approach. A quantitative research design was adopted using a structured questionnaire administered through Google Forms. The target population comprised young adults aged 18 to 25 in the UAE. A total of 100 valid responses were collected. Data analysis was conducted using IBM SPSS, applying regression analysis, correlation analysis, ANOVA, and chi-square tests to test the proposed hypotheses.ResultsThe findings indicate that social media brand engagement and paid advertisements have a significant influence on fast-food purchase behavior. In contrast, influencer-driven content and promotional offers showed a comparatively lower impact. Frequent interaction with brands on social media platforms was positively associated with brand loyalty and repeat purchase intention. Logistics efficiency, particularly timely delivery and order accuracy, emerged as a key determinant of customer satisfaction and retention. The usability and convenience of digital ordering platforms significantly influenced consumer ordering behavior.DiscussionThe study highlights the interconnected role of marketing communication, logistics performance, and digital platform usability in shaping fast-food consumption among digitally native consumers in the UAE. The results suggest that fast-food brands should align social media engagement strategies with strong operational execution to enhance customer experience and loyalty. Emphasis on convenience, reliable delivery, and consistent brand interaction is critical for sustaining competitive advantage. Future research may extend this work by examining other geographical contexts, increasing the sample size, and adopting a longitudinal approach to capture evolving digital engagement patterns
Parallelizing Unstructured Sparse Matrix Computations on Large-Scale Multiprocessors
No full textProblems in the class of unstructured sparse matrix computations are characterized by highly irregular dependencies and communication patterns that are not known at compile-time, but can be completely determined at run-time before the computations are actually performed. For this class of problems, current parallelizing compilers are unable to produce efficient code on large-scale distributed memory MIMD multiprocessors, and manual techniques are inflexible and too ad hoc to be generally effective.
In this thesis, we propose a run-time automatic partitioning and scheduling methodology for unstructured sparse matrix computations on large-scale multiprocessors. Our methodology is based on extracting information from the problem instance by preprocessing its symbolic structure, and using this information to achieve high performance in repeated iterations of the computations during which the symbolic structure is unchanged. We present efficient software tools to help users build their parallelization system by following this methodology.
We demonstrate the efficacy of our methodology on sparse Cholesky factorization, which has historically proven to be hard to parallelize. The highlight of our approach is a new two-dimensional block partitioning scheme. We build a run-time parallel system for block sparse Cholesky factorization called Sparse Hybrid Automatic Parallelization Environment (SHAPE), consisting of a parallel partitioner, a parallel scheduler and a parallel communication optimization algorithm. These are modular tools tied together by an explicit representation for block-based unstructured computations. We employ SHAPE to carry out an extensive experimental study of sparse Cholesky factorization on the iPSC/860. The experimental results show that with a judicious choice of partitioning parameters, our block-based partitioning and scheduling method outperforms a well-known column-based method in delivering high performance on a variety of structured and unstructured matrices. The preprocessing itself is shown to be very efficient, its cost being recovered in a small number of iterations of the factorization.
Our methodology and tools may be used to parallelize other unstructured sparse matrix computations for which the same symbolic structure is used in several iterations of the computations. Such computations include sparse triangular solution and sparse
matrix-vector multiplication.Technical report DCS-TR-30
SHAPE: A Parallelization Tool for Sparse Matrix Computations
No full textWe describe the design, implementation and performance of a Sparse Hybrid Automatic Parallelization Environment (SHAPE). SHAPE partitions and schedules sparse matrix computations for Cholesky factorization with the goal of achieving good performance at low cost, while providing flexibility for use as an experimental tool. It employs efficient parallelization algorithms which reduce the communication cost without adversely affecting the load balance by using a hybrid mixture of column and block partitions. Through several parameters, SHAPE aims for portability across a diverse range of sparse matrix structures and message-passing multiprocessors with different communication cost parameters. We present preliminary timing results on the iPSC/860 and compare the performance of SHAPE with that of a commonly used column-based method. The results show that SHAPE significantly reduces computation time, number of messages, and overall communication time for a variety of test matrices.Technical report dcs-tr-29
- …
