Search results for: large language models

Authors: Guilherme Ribeiro, Alexandre Oliveira, Antonio Ferreira, Shyam Visweswaran, Gregory Cooper

Abstract:

Patient-specific models are instance-based learning algorithms that take advantage of the particular features of the patient case at hand to predict an outcome. We introduce two patient-specific algorithms based on decision tree paradigm that use AUC as a metric to select an attribute. We apply the patient specific algorithms to predict outcomes in several datasets, including medical datasets. Compared to the patient-specific decision path (PSDP) entropy-based and CART methods, the AUC-based patient-specific decision path models performed equivalently on area under the ROC curve (AUC). Our results provide support for patient-specific methods being a promising approach for making clinical predictions.

Keywords: approach instance-based, area under the ROC curve, patient-specific decision path, clinical predictions

Procedia PDF Downloads 466

Authors: Krzysztof Pomorski

Abstract:

In the given work we present universal mathematical framework for modeling of cattle farm system that can set and validate various hypothesis that can be tested against experimental data. The presented work is preliminary but it is expected to be valid tool for future deeper analysis that can result in new class of prediction methods allowing early detection of cow dieseaes as well as cow performance. Therefore the presented work shall have its meaning in agriculture models and in machine learning as well. It also opens the possibilities for incorporation of certain class of biological models necessary in modeling of cow behavior and farm performance that might include the impact of environment on the farm system. Particular attention is paid to the model of coupled oscillators that it the basic building hypothesis that can construct the model showing certain periodic or quasiperiodic behavior.

Keywords: coupled ordinary differential equations, cattle farm system, numerical methods, stochastic differential equations

Procedia PDF Downloads 132

Authors: Hsiaoping Wu

Abstract:

With the wide use of Internet, learners are exposed to the wider world. This exposure permits learners to discover new information and combine a variety of media in order to reach in-depth and broader understanding of their literacy and the world. Many paper-based teaching, learning and assessment modalities can be transferred to a digital platform. This study examines the use of e-portfolios for ESL (English as a second language) pre-service teacher. The data were collected by reviewing 100 E-portfolio from 2013 to 2015 in order to synthesize meaningful information about e-portfolios for ESL pre-service teachers. Participants were generalists, bilingual and ESL pre-service teachers. The studies were coded into two main categories: learning gains, including assessment, and technical skills. The findings showed that using e-portfolios enhanced and developed ESL pre-service teachers’ teaching and assessment skills. Also, the E-portfolio also developed the pre-service teachers’ technical stills to prepare a comprehensible portfolio to present who they are. Finally, the study and presentation suggested e-portfolios for ecological issues and educational purposes.

Keywords: assessment, e-portfolio, pre-service teacher, reflection

Procedia PDF Downloads 305

Authors: Xiuying Wang, Changliang Li, Bo Xu

Abstract:

A fixed-length feature vector is required for many machine learning algorithms in NLP field. Word embeddings have been very successful at learning lexical information. However, they cannot capture the compositional meaning of sentences, which prevents them from a deeper understanding of language. In this paper, we introduce a novel hierarchical tree long short-term memory (HTLSTM) model that learns vector representations for sentences of arbitrary syntactic type and length. We propose to split one sentence into three hierarchies: short phrase, long phrase and full sentence level. The HTLSTM model gives our algorithm the potential to fully consider the hierarchical information and long-term dependencies of language. We design the experiments on both English and Chinese corpus to evaluate our model on sentiment analysis task. And the results show that our model outperforms several existing state of the art approaches significantly.

Keywords: deep learning, hierarchical tree long short-term memory, sentence representation, sentiment analysis

Procedia PDF Downloads 338

Authors: Wiem Zaabi, Yemna Bensalem, Hafedh Trabelsi

Abstract:

The paper presents a finite element (FE) based efficient analysis procedure for induction machine (IM). The FE formulation approaches are proposed to achieve this goal: the magnetostatic and the non-linear transient time stepped formulations. The study based on finite element models offers much more information on the phenomena characterizing the operation of electrical machines than the classical analytical models. This explains the increase of the interest for the finite element investigations in electrical machines. Based on finite element models, this paper studies the influence of the stator and the rotor faults on the behavior of the IM. In this work, a simple dynamic model for an IM with inter-turn winding fault and a broken bar fault is presented. This fault model is used to study the IM under various fault conditions and severity. The simulation results are conducted to validate the fault model for different levels of fault severity. The comparison of the results obtained by simulation tests allowed verifying the precision of the proposed FEM model. This paper presents a technical method based on Fast Fourier Transform (FFT) analysis of stator current and electromagnetic torque to detect the faults of broken rotor bar. The technique used and the obtained results show clearly the possibility of extracting signatures to detect and locate faults.

Keywords: Finite element Method (FEM), Induction motor (IM), short-circuit fault, broken rotor bar, Fast Fourier Transform (FFT) analysis

Procedia PDF Downloads 283

Authors: Raf E. Ul Shougat, Ezazul Haque Sabuz, G. M. Najmul Quader, Monon Mahboob

Abstract:

Since there are lots of ways for building and post processing of parts or models in 3D printing technology, the main objective of this research is to provide an understanding how mechanical characteristics of 3D printed parts get changed for different building orientations and infiltrates. Tensile, compressive, flexure, and hardness test were performed for the analysis of mechanical properties of those models. Specimens were designed in CAD software, printed on Z-printer 450 with five different build orientations and post processed with four different infiltrates. Results show that with the change of infiltrates or orientations each of the above mechanical property changes and for each infiltrate the highest tensile strength, flexural strength, and hardness are found for such orientation where there is the lowest number of layers while printing.

Keywords: 3D printing, building orientations, infiltrates, mechanical characteristics, number of layers

Procedia PDF Downloads 267

Authors: Mahmoud Ahmed Farrag, Ali Zain Elabdeen Heikal, Mohamed Shawky Ahmed, Ahmed Osama Amer

Abstract:

Crowdsourced data refers to data that is collected and shared by a large number of individuals or organizations, often through the use of digital technologies such as mobile devices and social media. The shortage in crash data collection in developing countries makes it difficult to fully understand and address road safety issues in these regions. In developing countries, crowdsourced data can be a valuable tool for improving road safety, particularly in urban areas where the majority of road crashes occur. This study is the first to develop safety performance functions using crowdsourced data by adopting a negative binomial structure model and Full Bayes model to investigate traffic safety for urban road networks and provide insights into the impact of roadway characteristics. Furthermore, as a part of the safety management process, network screening has been undergone through applying two different methods to rank the most hazardous road segments: PCR method (adopted in the Highway Capacity Manual HCM) as well as a graphical method using GIS tools to compare and validate. Lastly, recommendations were suggested for policymakers to ensure safer roads.

Keywords: crowdsourced data, road crashes, safety performance functions, Full Bayes models, network screening

Procedia PDF Downloads 13

Authors: Kohei Shimasaki, Tomoaki Okamura, Idaku Ishii

Abstract:

We propose a vibration imaging method for high frame rate (HFR)-video-based localization of vibrating objects with large translations. When the ratio of the translation speed of a target to its vibration frequency is large, obtaining its frequency response in image intensities becomes difficult because one or no waves are observable at the same pixel. Our method can precisely localize moving objects with vibration by virtually translating multiple image sequences for pixel-level short-time Fourier transform to observe multiple waves at the same pixel. The effectiveness of the proposed method is demonstrated by analyzing several HFR videos of flying insects in real scenarios.

Keywords: HFR video analysis, pixel-level vibration source localization, short-time Fourier transform, virtual translation

Procedia PDF Downloads 91

Authors: C. F. Kumru, C. Kocatepe, O. Arikan

Abstract:

In this study, electric field distribution analyses for three pylon models are carried out by a Finite Element Method (FEM) based software. Analyses are performed in both stationary and time domains to observe instantaneous values along with the effective ones. Considering the results of the study, different line geometries is considerably affecting the magnitude and distribution of electric field although the line voltages are the same. Furthermore, it is observed that maximum values of instantaneous electric field obtained in time domain analysis are quite higher than the effective ones in stationary mode. In consequence, electric field distribution analyses should be individually made for each different line model and the limit exposure values or distances to residential buildings should be defined according to the results obtained.

Keywords: electric field, energy transmission line, finite element method, pylon

Procedia PDF Downloads 712

Authors: Farah Diab, Mohamad Khalil, Francesca Storace, Francesca Baldini, Piero Portincasaa, Giulio Lupidi, Laura Vergani

Abstract:

Thymbra spicata is a thyme-like plant belonging to the Lamiaceae family that shows a global distribution, especially in the eastern Mediterranean region. Leaves of T. spicata contain large amounts of phenols such as phenolic acids (rosmarinic acid), phenolic monoterpenes (carvacrol), and flavonoids. In Lebanon, T. spicata is currently used as a culinary herb in salad and infusion, as well as for traditional medicinal purposes. Carvacrol (5-isopropyl-2-methyl phenol), the most abundant polyphenol in the organic extract and essential oils, has a great array of pharmacological properties. In fact, carvacrol is largely employed as a food additive and neutraceutical agent. Our aim is to investigate the beneficial effects of T. spicata ethanolic extract (TE) and its main component, carvacrol, using in vitro models of hepatic steatosis and endothelial dysfunction. As a further point, we focused on investigating if and how the binding of carvacrol to albumin, the physiological transporter for drugs in the blood, might be altered by the presence of high levels of fatty acids (FAs), thus impairing the carvacrol bio-distribution in vivo. For that reason, hepatic FaO cells treated with exogenous FAs such as oleate and palmitate mimic hepatosteatosis; endothelial HECV cells exposed to hydrogen peroxide are a model of endothelial dysfunction. In these models, we measured lipid accumulation, free radical production, lipoperoxidation, and nitric oxide release before and after treatment with carvacrol. The carvacrol binding to albumin with/without high levels of long-chain FAs was assessed by absorption and emission spectroscopies. Our findings show that both TE and carvacrol (i) counteracted lipid accumulation in hepatocytes by decreasing the intracellular and extracellular lipid contents in steatotic FaO cells; (ii) decreased oxidative stress in endothelial cells by significantly reducing lipoperoxidation and free radical production, as well as, attenuating the nitric oxide release; (ii) high levels of circulating FAs reduced the binding of carvacrol to albumin. The beneficial effects of TE and carvacrol on both hepatic and endothelial cells point to a nutraceutical potential. However, high levels of circulating FAs, such as those occurring in metabolic disorders, might hinder the carvacrol transport, bio-distribution, and pharmacodynamics.

Keywords: carvacrol, endothelial dysfunction, fatty acids, non-alcoholic fatty liver diseases, serum albumin

Procedia PDF Downloads 171

Authors: Monday Agbonica Bello Eje

Abstract:

The language of the print media is characterized by the use of apposition. This linguistic element function strategically in journalistic discourse where it is communicatively necessary to name individuals and provide information about them. Linguistic studies on the language of the print media with bias for apposition have largely dwelt on other areas but the examination of the typology of appositive reference in newspaper discourse. Yet, it is capable of revealing ways writers communicate and provide information necessary for readers to follow and understand the message. The study, therefore, analyses the patterns of appositional occurrences and the typology of reference in newspaper articles. The data were obtained from The Punch and Daily Trust Newspapers. A total of six editions of these newspapers were collected randomly spread over three months. News and feature articles were used in the analysis. Guided by the referential theory of meaning in discourse, the appositions identified were subjected to analysis. The findings show that the semantic relation of coreference and speaker coreference have the highest percentage and frequency of occurrence in the data. This is because the subject matter of news reports and feature articles focuses on humans and the events around them; as a result, readers need to be provided with some form of detail and background information in order to identify as well as follow the discourse. Also, the non-referential relation of absolute synonymy and speaker synonymy no doubt have fewer occurrences and percentages in the analysis. This is tied to a major feature of the language of the media: simplicity. The paper concludes that appositions is mainly used for the purpose of providing the reader with much detail. In this way, the writer transmits information which helps him not only to give detailed yet concise descriptions but also in some way help the reader to follow the discourse.

Keywords: apposition, discourse, newspaper, Nigeria, reference

Procedia PDF Downloads 142

Authors: Ramoshweu Melvin Sethobya, Emmanuel Chirenje, Mihlali Hobo, Simon Sebothoma

Abstract:

The recent surge in residential infrastructure development around the metropolitan areas of South Africa has necessitated conditions for thorough geotechnical assessments to be conducted prior to site developments to ensure human and infrastructure safety. This paper appraises the success in the application of multi-method geophysical techniques for the delineation of sinkhole vulnerability in a residential landscape. Geophysical techniques ERT, MASW, VES, Magnetics and gravity surveys were conducted to assist in mapping sinkhole vulnerability, using an existing sinkhole as a constraint at Venterspost town, West of Johannesburg city. A combination of different geophysical techniques and results integration from those proved to be useful in the delineation of the lithologic succession around sinkhole locality, and determining the geotechnical characteristics of each layer for its contribution to the development of sinkholes, subsidence and cavities at the vicinity of the site. Study results have also assisted in the determination of the possible depth extension of the currently existing sinkhole and the location of sites where other similar karstic features and sinkholes could form. Results of the ERT, VES and MASW surveys have uncovered dolomitic bedrock at varying depths around the sites, which exhibits high resistivity values in the range 2500-8000ohm.m and corresponding high velocities in the range 1000-2400 m/s. The dolomite layer was found to be overlain by a weathered chert-poor dolomite layer, which has resistivities between the range 250-2400ohm.m, and velocities ranging from 500-600m/s, from which the large sinkhole has been found to collapse/ cave in. A compiled 2.5D high resolution Shear Wave Velocity (Vs) map of the study area was created using 2D profiles of MASW data, offering insights into the prevailing lithological setup conducive for formation various types of karstic features around the site. 3D magnetic models of the site highlighted the regions of possible subsurface interconnections between the currently existing large sinkhole and the other subsidence feature at the site. A number of depth slices were used to detail the conditions near the sinkhole as depth increases. Gravity surveys results mapped the possible formational pathways for development of new karstic features around the site. Combination and correlation of different geophysical techniques proved useful in delineation of the site geotechnical characteristics and mapping the possible depth extend of the currently existing sinkhole.

Keywords: resistivity, magnetics, sinkhole, gravity, karst, delineation, VES

Procedia PDF Downloads 55

Authors: Seyed-Ali Sadegh-Zadeh, Kaveh Kavianpour, Hamed Atashbar, Elham Heidari, Saeed Shiry Ghidary, Amir M. Hajiyavand

Abstract:

Evaluation metrics play a critical role in assessing the performance of machine learning models. In this review paper, we provide a comprehensive overview of performance measurement approaches for machine learning models. For each category, we discuss the most widely used metrics, including their mathematical formulations and interpretation. Additionally, we provide a comparative analysis of performance measurement approaches for metric combinations. Our review paper aims to provide researchers and practitioners with a better understanding of performance measurement approaches and to aid in the selection of appropriate evaluation metrics for their specific applications.

Keywords: evaluation metrics, performance measurement, supervised learning, unsupervised learning, reinforcement learning, model robustness and stability, comparative analysis

Procedia PDF Downloads 45

Authors: Gohar Rahman, Rabia Saad Ullah

Abstract:

This research aims to examine the use of discourse markers within the dominion of political speeches, differentiating between genders. The analysis centers on twelve speakers, comprising six males and six females. Speeches selected include commencement, victory, state union addresses, campaigns, and presidential speeches. Halliday and Hasan's cohesion framework, specifically discourse markers, is utilized as a theoretical framework. Data is quantitatively analyzed using AntConc to identify marker frequency. The findings are presented through Excel's tables and graphs, suggesting differences in discourse marker preferences between genders. The findings suggest a divergence in the preferences for discourse markers between males and females. However, asserting that females utilize discourse markers more frequently due to the increased use of filler words, face threat mitigation, and polite speech would be an exaggeration. The disparity in frequency is not substantial, suggesting that males and females exhibit varying language inclinations to some degree.

Keywords: discourse markers, political discourse, gender, speeches, language

Procedia PDF Downloads 46

Authors: Sara ElElimy, Samir Moustafa

Abstract:

Mobile network operators start to face many challenges in the digital era, especially with high demands from customers. Since mobile network operators are considered a source of big data, traditional techniques are not effective with new era of big data, Internet of things (IoT) and 5G; as a result, handling effectively different big datasets becomes a vital task for operators with the continuous growth of data and moving from long term evolution (LTE) to 5G. So, there is an urgent need for effective Big data analytics to predict future demands, traffic, and network performance to full fill the requirements of the fifth generation of mobile network technology. In this paper, we introduce data science techniques using machine learning and deep learning algorithms: the autoregressive integrated moving average (ARIMA), Bayesian-based curve fitting, and recurrent neural network (RNN) are employed for a data-driven application to mobile network operators. The main framework included in models are identification parameters of each model, estimation, prediction, and final data-driven application of this prediction from business and network performance applications. These models are applied to Telecom Italia Big Data challenge call detail records (CDRs) datasets. The performance of these models is found out using a specific well-known evaluation criteria shows that ARIMA (machine learning-based model) is more accurate as a predictive model in such a dataset than the RNN (deep learning model).

Keywords: big data analytics, machine learning, CDRs, 5G

Procedia PDF Downloads 124

Authors: Yun-Tao Zhang, Jong-Yeop Bae, Whoi-Yul Kim

Abstract:

Background modeling and subtraction in video analysis has been widely proved to be an effective method for moving objects detection in many computer vision applications. Over the past years, a large number of approaches have been developed to tackle different types of challenges in this field. However, the dynamic background and illumination variations are two of the most frequently occurring issues in the practical situation. This paper presents a new two-layer model based on codebook algorithm incorporated with local binary pattern (LBP) texture measure, targeted for handling dynamic background and illumination variation problems. More specifically, the first layer is designed by block-based codebook combining with LBP histogram and mean values of RGB color channels. Because of the invariance of the LBP features with respect to monotonic gray-scale changes, this layer can produce block-wise detection results with considerable tolerance of illumination variations. The pixel-based codebook is employed to reinforce the precision from the outputs of the first layer which is to eliminate false positives further. As a result, the proposed approach can greatly promote the accuracy under the circumstances of dynamic background and illumination changes. Experimental results on several popular background subtraction datasets demonstrate a very competitive performance compared to previous models.

Keywords: background subtraction, codebook model, local binary pattern, dynamic background, illumination change

Procedia PDF Downloads 201

Authors: Joy Aworo-Okoroh

Abstract:

Politics is about trust and trust is challenged by the speaker’s ability to manipulate language before the electorate. Critical discourse analysis investigates the role of language in constructing social relationships between a political speaker and his audience. This paper explores the linguistic choices made by President Muhammad Buhari that enshrines his ideologies as well as the socio-political relations of power between him and Nigerians in his speeches. Two speeches of President Buhari –inaugural and Independence Day speeches are analyzed using Norman Fairclough’s perspective on Halliday’s Systemic functional grammar. The analysis is at two levels. The first level of analysis is the identification of transitivity and modality choices in the speeches and how they reveal the covert ideologies. The second analysis is premised on Normal Fairclough’s model, the clauses are analyzed to identify elements of power, hesistation, persuasion, threat and religious statement. It was discovered that Buhari is a dominant character who manipulates the material processes a lot.

Keywords: politics, critical discourse analysis, Norman Fairclough, systemic functional grammar

Procedia PDF Downloads 527

Authors: Lipaz Shamoa-Nir, Irene Razpurker-Apfeld

Abstract:

This research examined the impact of subliminal exposure to religious content on the mutual attitudes of majority group members (Jews) and minority group members (Muslims). Participants were subliminally exposed to religious concepts (e.g., Mezuzah, yarmulke or veil) and then they filled questionnaires assessing their stereotypes towards the out-group members. Each participant was primed with either in-group religious concepts, out-group concepts or neutral ones. The findings show that the Muslim participants were not influenced by the religious content to which they were exposed while the Jewish participants perceived the Muslims as less 'hostile' when subliminally exposed to religious concepts, regardless of concept type (out-group/in-group). This research highlights the influence of evoked religious content on out-group attitudes even when the perceiver is unaware of prime content. The power that exposure to content in a non-native language has in activating attitudes towards the out-group is also discussed.

Keywords: intergroup attitudes, stereotypes, majority-minority, religious out-group, implicit content, native language

Procedia PDF Downloads 231

Authors: Hiromi Kawai

Abstract:

This study investigates the nature of EFL older children’s L2 perceptive and productive abilities using classroom data, in order to find a pedagogical solution to the teaching of L2 sounds at an early stage of learning in a formal school setting. It is still inconclusive whether older children with only EFL formal school instruction at the initial stage of L2 learning are able to attain native-like perception and production in English within the very limited amount of exposure to the target language available. Based on the notion of the lack of study of EFL Japanese children’s acquisition of English segments, the researcher uses a model of L1 speech processing which was developed for investigating L1 English children’s speech and literacy difficulties using a psycholinguistic framework. The model is composed of input channel, output channel, and lexical representation, and examines how a child receives information from spoken or written language, remembers and stores it within the lexical representations and how the child selects and produces spoken or written words. Concerning language universality and language specificity in the language acquisitional process, the aim of finding any sound errors in L1 English children seemed to conform to the author’s intention to find abilities of English sounds in older Japanese children at the novice level of English in an EFL setting. 104 students in Grade 5 (between the ages of 10 and 11 years old) of an elementary school in Tokyo participated in this study. Four tests to measure their perceptive ability and three oral repetition tests to measure their productive ability were conducted with/without reference to lexical representation. All the test items were analyzed to calculate item facility (IF) indices, and correlational analyses and Structural Equation Modeling (SEM) were conducted to examine the relationship between the receptive ability and the productive ability. IF analysis showed that (1) the participants were better at perceiving a segment than producing a segment, (2) they had difficulty in auditory discrimination of paired consonants when one of them does not exist in the Japanese inventory, (3) they had difficulty in both perceiving and producing English vowels, and (4) their L1 loan word knowledge had an influence on their ability to perceive and produce L2 sounds. The result of the Multiple Regression Modeling showed that the two production tests could predict the participants’ auditory ability of real words in English. The result of SEM showed that the hypothesis that perceptive ability affects productive ability was supported. Based on these findings, the author discusses the possible explicit method of teaching English segments to EFL older children in a formal school setting.

Keywords: EFL older children, english segments, perception, production, speech processing system

Procedia PDF Downloads 230

Authors: Atef Ibrahim, Hamed Elsimary, Abdullah Aljumah, Fayez Gebali

Abstract:

This paper presents a novel semi-systolic array architecture for an optimized parallel sequence alignment algorithm. This architecture has the advantage that it can be modified to be reused for multiple pass processing in order to increase the number of processing elements that can be packed into a single FPGA and to increase the number of sequences that can be aligned in parallel in a single FPGA. This resolves the potential problem of many FPGA resources left unused for designs that have large values of short read length. When using the previously published conventional hardware design. FPGA implementation results show that, for large values of short read lengths (M>128), the proposed design has a slightly higher speed up and FPGA utilization over the the conventional one.

Keywords: bioinformatics, genome sequence alignment, re-sequencing applications, systolic array

Procedia PDF Downloads 512

Authors: Wala A. Bagunaid, Maram Meccawy, Arwa Allinjawi, Zilal Meccawy

Abstract:

Continuous self-assessment becomes crucial in self-paced online learning environments. Students often depend on themselves to assess their progress; which is considered an essential requirement for any successful learning process. Today’s education institutions face major problems around student motivation and engagement. Thus, personalized e-learning systems aim to help and guide the students. Gamification provides an opportunity to help students for self-assessment and social comparison with other students through attempting to harness the motivational power of games and apply it to the learning environment. Furthermore, Open Social Student Modeling (OSSM) as considered as the latest user modeling technologies is believed to improve students’ self-assessment and to allow them to social comparison with other students. This research integrates OSSM approach and gamification concepts in order to provide self-assessment for English language learners at King Abdulaziz University (KAU). This is achieved through an interactive visual representation of their learning progress.

Keywords: e-learning system, gamification, motivation, social comparison, visualization

Procedia PDF Downloads 134

Authors: Minghui Zou

Abstract:

This study investigates whether and how cognate status influences switching costs in bilingual language production. Two picture naming tasks will be conducted in this proposed study by manipulating the conditions of how cognates and non-cognates are presented, i.e., separately in two testing blocks vs intermixed in one single testing block. Participants of each experiment will be 24 L1-European Portuguese L2-English unbalanced speakers. Stimuli will include 12 pictures of cognate nouns and 12 of non-cognate nouns. It is hypothesized that there will be cognate switch facilitation effects among unbalanced bilinguals in both of their languages when stimuli are presented either in two single testing blocks or one mixed testing block. Shorter reaction times and higher naming accuracy are expected to be found in cognate switch trials than in non-cognate switch trials.

Keywords: cognates, language switching costs, picture naming, European Portuguese, cognate facilitation effect

Procedia PDF Downloads 11

Authors: Ankit Sinha, Soham Banerjee, Pratik Chattopadhyay

Abstract:

Automated product recognition in retail stores is an important real-world application in the domain of Computer Vision and Pattern Recognition. In this paper, we consider the problem of automatically identifying the classes of the products placed on racks in retail stores from an image of the rack and information about the query/product images. We improve upon the existing approaches in terms of effectiveness and memory requirement by developing a two-stage object detection and recognition pipeline comprising of a Faster-RCNN-based object localizer that detects the object regions in the rack image and a ResNet-18-based image encoder that classifies the detected regions into the appropriate classes. Each of the models is fine-tuned using appropriate data sets for better prediction and data augmentation is performed on each query image to prepare an extensive gallery set for fine-tuning the ResNet-18-based product recognition model. This encoder is trained using a triplet loss function following the strategy of online-hard-negative-mining for improved prediction. The proposed models are lightweight and can be connected in an end-to-end manner during deployment to automatically identify each product object placed in a rack image. Extensive experiments using Grozi-32k and GP-180 data sets verify the effectiveness of the proposed model.

Keywords: retail stores, faster-RCNN, object localization, ResNet-18, triplet loss, data augmentation, product recognition

Procedia PDF Downloads 134

Authors: William Chung

Abstract:

This paper is to study the use of multiple subproblems in Dantzig-Wolfe decomposition of linear programming (DW-LP). Traditionally, the decomposed LP consists of one LP master problem and one LP subproblem. The master problem and the subproblem is solved alternatively by exchanging the dual prices of the master problem and the proposals of the subproblem until the LP is solved. It is well known that convergence is slow with a long tail of near-optimal solutions (asymptotic convergence). Hence, the performance of DW-LP highly depends upon the number of decomposition steps. If the decomposition steps can be greatly reduced, the performance of DW-LP can be improved significantly. To reduce the number of decomposition steps, one of the methods is to increase the number of proposals from the subproblem to the master problem. To do so, we propose to add a quadratic approximation function to the LP subproblem in order to develop a set of approximate-LP subproblems (multiple subproblems). Consequently, in each decomposition step, multiple subproblems are solved for providing multiple proposals to the master problem. The number of decomposition steps can be reduced greatly. Note that each approximate-LP subproblem is nonlinear programming, and solving the LP subproblem must faster than solving the nonlinear multiple subproblems. Hence, using multiple subproblems in DW-LP is the tradeoff between the number of approximate-LP subproblems being formed and the decomposition steps. In this paper, we derive the corresponding algorithms and provide some simple computational results. Some properties of the resulting algorithms are also given.

Keywords: approximate subproblem, Dantzig-Wolfe decomposition, large-scale models, multiple subproblems

Procedia PDF Downloads 147

Authors: Mohammadreza Khanmohammadi, Noureddin Nakhostin Ansari, Soofia Naghdi

Abstract:

Purpose: Our study aimed to validate single item presenteeism question (SIPQ) into the Persian language for patients with low back pain. Background information: low back pain is a common health problem, and it is one of the most prevalent disorder in working people. There are the different subjective way to evaluate the effect of back pain on work productivity that one of them is by implementing single item presenteeism question. This question has not been validated into the Persian language. Method: Patients were asked to answer SIPQ and pain from 0 to 10 according to numerical rating scale (NRS). The functional rating index was administrated to evaluate construct validity. For test-retest reliability, almost 50 patients re-completed the Persian SIPQ. The construct validity of SIPQ was assessed by analyzing Spearman rank correlation between this question and the Persian version of Functional rating index questionnaire. To analyze test-retest reliability, we assessed intraclass correlation coefficient (agreement) (ICC agreement) (two-way random effects model, single measure). Results: The SIPQ score of two groups of patients (84 males, 16 females, mean age ±SD: 33.85±11.16 years, range: 19-67 years) and healthy subjects (48 male, 2 female ones, mean age ±SD: 24.24 ±8.07 years) was statistically significant. (Mann-Whitney U =198.00, P<.001). The Spearman correlation of data showed that there is a significant correlation between Persian SIPQ score and Persian FRI band (r= .559, P<.001). The ICC was .62. So, the analysis indicated good, test-retest reliability. Conclusion: This study showed that Persian version of SIPQ is reliable and valid when applied to back pain patients.

Keywords: cross cultural adaptation, economic burden, low back pain, Persian language, translation

Procedia PDF Downloads 400

Authors: Sujit Kumar Pradhan, Anil Kumar, Vijay Kumar

Abstract:

The testing process of the software during the software development time is a crucial step as it makes the software more efficient and dependable. To estimate software’s reliability through the mean value function, many software reliability growth models (SRGMs) were developed under the assumption that operating and testing environments are the same. Practically, it is not true because when the software works in a natural field environment, the reliability of the software differs. This article discussed an SRGM comprising change-point and imperfect debugging in a simulated testing environment. Later on, we extended it in a multi-release direction. Initially, the software was released to the market with few features. According to the market’s demand, the software company upgraded the current version by adding new features as time passed. Therefore, we have proposed a generalized multi-release SRGM where change-point and imperfect debugging concepts have been addressed in a simulated testing environment. The failure-increasing rate concept has been adopted to determine the change point for each software release. Based on nine goodness-of-fit criteria, the proposed model is validated on two real datasets. The results demonstrate that the proposed model fits the datasets better. We have also discussed the optimal release time of the software through a cost model by assuming that the testing and debugging costs are time-dependent.

Keywords: software reliability growth models, non-homogeneous Poisson process, multi-release software, mean value function, change-point, environmental factors

Procedia PDF Downloads 61

Authors: Milica Karadzic, Lidija Jevric, Sanja Podunavac-Kuzmanovic, Strahinja Kovacevic, Anamarija Mandic, Katarina Penov Gasi, Marija Sakac, Aleksandar Okljesa, Andrea Nikolic

Abstract:

The present paper deals with chromatographic lipophilicity prediction of newly synthesized steroid derivatives. The prediction was achieved using in silico generated molecular descriptors and quantitative structure-retention relationship (QSRR) methodology with the artificial neural networks (ANN) approach. Chromatographic lipophilicity of the investigated compounds was expressed as retention factor value logk. For QSRR modeling, a feedforward back-propagation ANN with gradient descent learning algorithm was applied. Using the novel sum of ranking differences (SRD) method generated ANN models were ranked. The aim was to distinguish the most consistent QSRR model that can be found, and similarity or dissimilarity between the models that could be noticed. In this study, SRD was performed with average values of retention factor value logk as reference values. An excellent correlation between experimentally observed retention factor value logk and values predicted by the ANN was obtained with a correlation coefficient higher than 0.9890. Statistical results show that the established ANN models can be applied for required purpose. This article is based upon work from COST Action (TD1305), supported by COST (European Cooperation in Science and Technology).

Keywords: artificial neural networks, liquid chromatography, molecular descriptors, steroids, sum of ranking differences

Procedia PDF Downloads 301

Authors: Farid Khosravikia, Patricia Clayton

Abstract:

The main objective of this work is to evaluate the advantages and disadvantages of various machine learning techniques in two key steps of seismic hazard and risk assessment of different types of structures. The first step is the development of ground-motion models, which are used for forecasting ground-motion intensity measures (IM) given source characteristics, source-to-site distance, and local site condition for future events. IMs such as peak ground acceleration and velocity (PGA and PGV, respectively) as well as 5% damped elastic pseudospectral accelerations at different periods (PSA), are indicators of the strength of shaking at the ground surface. Typically, linear regression-based models, with pre-defined equations and coefficients, are used in ground motion prediction. However, due to the restrictions of the linear regression methods, such models may not capture more complex nonlinear behaviors that exist in the data. Thus, this study comparatively investigates potential benefits from employing other machine learning techniques as statistical method in ground motion prediction such as Artificial Neural Network, Random Forest, and Support Vector Machine. The results indicate the algorithms satisfy some physically sound characteristics such as magnitude scaling distance dependency without requiring pre-defined equations or coefficients. Moreover, it is shown that, when sufficient data is available, all the alternative algorithms tend to provide more accurate estimates compared to the conventional linear regression-based method, and particularly, Random Forest outperforms the other algorithms. However, the conventional method is a better tool when limited data is available. Second, it is investigated how machine learning techniques could be beneficial for developing probabilistic seismic demand models (PSDMs), which provide the relationship between the structural demand responses (e.g., component deformations, accelerations, internal forces, etc.) and the ground motion IMs. In the risk framework, such models are used to develop fragility curves estimating exceeding probability of damage for pre-defined limit states, and therefore, control the reliability of the predictions in the risk assessment. In this study, machine learning algorithms like artificial neural network, random forest, and support vector machine are adopted and trained on the demand parameters to derive PSDMs for them. It is observed that such models can provide more accurate estimates of prediction in relatively shorter about of time compared to conventional methods. Moreover, they can be used for sensitivity analysis of fragility curves with respect to many modeling parameters without necessarily requiring more intense numerical response-history analysis.

Keywords: artificial neural network, machine learning, random forest, seismic risk analysis, seismic hazard analysis, support vector machine

Procedia PDF Downloads 89

Authors: Widi Astuti, Rizki Agus Hermawan, Hariono Mukti, Nurul Retno Sugiyono

Abstract:

The removal of lead ion (Pb²⁺) from aqueous solution by activated carbon with phosphoric acid activation employing mangrove propagule as precursor was investigated in a batch adsorption system. Batch studies were carried out to address various experimental parameters including pH and contact time. The Langmuir and Freundlich models were able to describe the adsorption equilibrium, while the pseudo first order and pseudo second order models were used to describe kinetic process of Pb²⁺ adsorption. The results show that the adsorption data are seen in accordance with Langmuir isotherm model and pseudo-second order kinetic model.

Keywords: activated carbon, adsorption, equilibrium, kinetic, lead, mangrove propagule

Procedia PDF Downloads 149

Authors: Andrea Fiorista

Abstract:

The learning of prepositions is a quite problematic aspect in foreign language instruction, and Italian is certainly not an exception. In their prototypical function, prepositions express schematic relations of two entities in a highly abstract, typically image-schematic way. In other terms, prepositions assume concepts such as directionality, collocation of objects in space and time and, in Cognitive Linguistics’ terms, the position of a trajector with respect to a landmark. Learners of different native languages may conceptualize them differently, implying that they are supposed to operate a recategorization (or create new categories) fitting with the target language. However, most current Italian Foreign/Second Language handbooks and didactic grammars do not facilitate learners in carrying out the task, as they tend to provide partial and idiosyncratic descriptions, with the consequent learner’s effort to memorize them, most of the time without success. In their prototypical meaning, prepositions are used to specify precise topographical positions in the physical environment which become less and less accurate as they radiate out from what might be termed a concrete prototype. According to that, the present study aims to elaborate a cognitive and conceptually well-grounded analysis of some extensive uses of the Italian preposition a, in order to propose effective pedagogical solutions in the Teaching/Learning process. Image schemas, cognitive metaphors and embodiment represent efficient cognitive tools in a task like this. Actually, while learning the merely spatial use of the preposition a (e.g. Sono a Roma = I am in Rome; vado a Roma = I am going to Rome,…) is quite straightforward, it is more complex when a appears in constructions such as verbs of motion +a + infinitive (e.g. Vado a studiare = I am going to study), inchoative periphrasis (e.g. Tra poco mi metto a leggere = In a moment I will read), causative construction (e.g. Lui mi ha mandato a lavorare = He sent me to work). The study reports data from a teaching intervention of Focus on Form, in which a basic cognitive schema is used to facilitate both teachers and students to respectively explain/understand the extensive uses of a. The educational material employed translates Cognitive Linguistics’ theoretical assumptions, such as image schemas and cognitive metaphors, into simple images or proto-scenes easily comprehensible for learners. Illustrative material, indeed, is supposed to make metalinguistic contents more accessible. Moreover, the concept of embodiment is pedagogically applied through activities including motion and learners’ bodily involvement. It is expected that replacing rote learning with a methodology that gives grammatical elements a proper meaning, makes learning process more effective both in the short and long term.

Keywords: cognitive approaches to language teaching, image schemas, embodiment, Italian as FL/SL

Procedia PDF Downloads 76