Search results for: emotion models
Commenced in January 2007
Frequency: Monthly
Edition: International
Paper Count: 7151

Search results for: emotion models

6821 A Super-Efficiency Model for Evaluating Efficiency in the Presence of Time Lag Effect

Authors: Yanshuang Zhang, Byungho Jeong

Abstract:

In many cases, there is a time lag between the consumption of inputs and the production of outputs. This time lag effect should be considered in evaluating the performance of organizations. Recently, a couple of DEA models were developed for considering time lag effect in efficiency evaluation of research activities. Multi-periods input(MpI) and Multi-periods output(MpO) models are integrated models to calculate simple efficiency considering time lag effect. However, these models can’t discriminate efficient DMUs because of the nature of basic DEA model in which efficiency scores are limited to ‘1’. That is, efficient DMUs can’t be discriminated because their efficiency scores are same. Thus, this paper suggests a super-efficiency model for efficiency evaluation under the consideration of time lag effect based on the MpO model. A case example using a long-term research project is given to compare the suggested model with the MpO model.

Keywords: DEA, super-efficiency, time lag, multi-periods input

Procedia PDF Downloads 474
6820 Exploring Tweet Geolocation: Leveraging Large Language Models for Post-Hoc Explanations

Authors: Sarra Hasni, Sami Faiz

Abstract:

In recent years, location prediction on social networks has gained significant attention, with short and unstructured texts like tweets posing additional challenges. Advanced geolocation models have been proposed, increasing the need to explain their predictions. In this paper, we provide explanations for a geolocation black-box model using LIME and SHAP, two state-of-the-art XAI (eXplainable Artificial Intelligence) methods. We extend our evaluations to Large Language Models (LLMs) as post hoc explainers for tweet geolocation. Our preliminary results show that LLMs outperform LIME and SHAP by generating more accurate explanations. Additionally, we demonstrate that prompts with examples and meta-prompts containing phonetic spelling rules improve the interpretability of these models, even with informal input data. This approach highlights the potential of advanced prompt engineering techniques to enhance the effectiveness of black-box models in geolocation tasks on social networks.

Keywords: large language model, post hoc explainer, prompt engineering, local explanation, tweet geolocation

Procedia PDF Downloads 28
6819 Classification of Business Models of Italian Bancassurance by Balance Sheet Indicators

Authors: Andrea Bellucci, Martina Tofi

Abstract:

The aim of paper is to analyze business models of bancassurance in Italy for life business. The life insurance business is very developed in the Italian market and banks branches have 80% of the market share. Given its maturity, the life insurance market needs to consolidate its organizational form to allow for the development of non-life business, which nowadays collects few premiums but represents a great opportunity to enlarge the market share of bancassurance using its strength in the distribution channel while the market share of independent agents is decreasing. Starting with the main business model of bancassurance for life business, this paper will analyze the performances of life companies in the Italian market by balance sheet indicators and by main discriminant variables of business models. The study will observe trends from 2013 to 2015 for the Italian market by exploiting a database managed by Associazione Nazionale delle Imprese di Assicurazione (ANIA). The applied approach is based on a bottom-up analysis starting with variables and indicators to define business models’ classification. The statistical classification algorithm proposed by Ward is employed to design business models’ profiles. Results from the analysis will be a representation of the main business models built by their profile related to indicators. In that way, an unsupervised analysis is developed that has the limit of its judgmental dimension based on research opinion, but it is possible to obtain a design of effective business models.

Keywords: bancassurance, business model, non life bancassurance, insurance business value drivers

Procedia PDF Downloads 299
6818 Reducing Uncertainty in Climate Projections over Uganda by Numerical Models Using Bias Correction

Authors: Isaac Mugume

Abstract:

Since the beginning of the 21st century, climate change has been an issue due to the reported rise in global temperature and changes in the frequency as well as severity of extreme weather and climatic events. The changing climate has been attributed to rising concentrations of greenhouse gases, including environmental changes such as ecosystems and land-uses. Climatic projections have been carried out under the auspices of the intergovernmental panel on climate change where a couple of models have been run to inform us about the likelihood of future climates. Since one of the major forcings informing the changing climate is emission of greenhouse gases, different scenarios have been proposed and future climates for different periods presented. The global climate models project different areas to experience different impacts. While regional modeling is being carried out for high impact studies, bias correction is less documented. Yet, the regional climate models suffer bias which introduces uncertainty. This is addressed in this study by bias correcting the regional models. This study uses the Weather Research and Forecasting model under different representative concentration pathways and correcting the products of these models using observed climatic data. This study notes that bias correction (e.g., the running-mean bias correction; the best easy systematic estimator method; the simple linear regression method, nearest neighborhood, weighted mean) improves the climatic projection skill and therefore reduce the uncertainty inherent in the climatic projections.

Keywords: bias correction, climatic projections, numerical models, representative concentration pathways

Procedia PDF Downloads 121
6817 A Nonlinear Dynamical System with Application

Authors: Abdullah Eqal Al Mazrooei

Abstract:

In this paper, a nonlinear dynamical system is presented. This system is a bilinear class. The bilinear systems are very important kind of nonlinear systems because they have many applications in real life. They are used in biology, chemistry, manufacturing, engineering, and economics where linear models are ineffective or inadequate. They have also been recently used to analyze and forecast weather conditions. Bilinear systems have three advantages: First, they define many problems which have a great applied importance. Second, they give us approximations to nonlinear systems. Thirdly, they have a rich geometric and algebraic structures, which promises to be a fruitful field of research for scientists and applications. The type of nonlinearity that is treated and analyzed consists of bilinear interaction between the states vectors and the system input. By using some properties of the tensor product, these systems can be transformed to linear systems. But, here we discuss the nonlinearity when the state vector is multiplied by itself. So, this model will be able to handle evolutions according to the Lotka-Volterra models or the Lorenz weather models, thus enabling a wider and more flexible application of such models. Here we apply by using an estimator to estimate temperatures. The results prove the efficiency of the proposed system.

Keywords: Lorenz models, nonlinear systems, nonlinear estimator, state-space model

Procedia PDF Downloads 254
6816 Models of State Organization and Influence over Collective Identity and Nationalism in Spain

Authors: Muñoz-Sanchez, Victor Manuel, Perez-Flores, Antonio Manuel

Abstract:

The main objective of this paper is to establish the relationship between models of state organization and the various types of collective identity expressed by the Spanish. The question of nationalism and identity ascription in Spain has always been a topic of special importance due to the presence in that country of territories where the population emits very different opinions of nationalist sentiment than the rest of Spain. The current situation of sovereignty challenge of Catalonia to the central government exemplifies the importance of the subject matter. In order to analyze this process of interrelation, we use a secondary data mining by applying the multiple correspondence analysis technique (MCA). As a main result a typology of four types of expression of collective identity based on models of State organization are shown, which are connected with the party position on this issue.

Keywords: models of organization of the state, nationalism, collective identity, Spain, political parties

Procedia PDF Downloads 444
6815 Mosaic Augmentation: Insights and Limitations

Authors: Olivia A. Kjorlien, Maryam Asghari, Farshid Alizadeh-Shabdiz

Abstract:

The goal of this paper is to investigate the impact of mosaic augmentation on the performance of object detection solutions. To carry out the study, YOLOv4 and YOLOv4-Tiny models have been selected, which are popular, advanced object detection models. These models are also representatives of two classes of complex and simple models. The study also has been carried out on two categories of objects, simple and complex. For this study, YOLOv4 and YOLOv4 Tiny are trained with and without mosaic augmentation for two sets of objects. While mosaic augmentation improves the performance of simple object detection, it deteriorates the performance of complex object detection, specifically having the largest negative impact on the false positive rate in a complex object detection case.

Keywords: accuracy, false positives, mosaic augmentation, object detection, YOLOV4, YOLOV4-Tiny

Procedia PDF Downloads 127
6814 Investigation of Different Control Stratgies for UPFC Decoupled Model and the Impact of Location on Control Parameters

Authors: S. A. Al-Qallaf, S. A. Al-Mawsawi, A. Haider

Abstract:

In order to evaluate the performance of a unified power flow controller (UPFC), mathematical models for steady state and dynamic analysis are to be developed. The steady state model is mainly concerned with the incorporation of the UPFC in load flow studies. Several load flow models for UPFC have been introduced in literature, and one of the most reliable models is the decoupled UPFC model. In spite of UPFC decoupled load flow model simplicity, it is more robust compared to other UPFC load flow models and it contains unique capabilities. Some shortcoming such as additional set of nonlinear equations are to be solved separately after the load flow solution is obtained. The aim of this study is to investigate the different control strategies that can be realized in the decoupled load flow model (individual control and combined control), and the impact of the location of the UPFC in the network on its control parameters.

Keywords: UPFC, decoupled model, load flow, control parameters

Procedia PDF Downloads 555
6813 A Study on Characteristics of Hedonic Price Models in Korea Based on Meta-Regression Analysis

Authors: Minseo Jo

Abstract:

The purpose of this paper is to examine the factors in the hedonic price models, that has significance impact in determining the price of apartments. There are many variables employed in the hedonic price models and their effectiveness vary differently according to the researchers and the regions they are analysing. In order to consider various conditions, the meta-regression analysis has been selected for the study. In this paper, four meta-independent variables, from the 65 hedonic price models to analysis. The factors that influence the prices of apartments, as well as including factors that influence the prices of apartments, regions, which are divided into two of the research performed, years of research performed, the coefficients of the functions employed. The covariance between the four meta-variables and p-value of the coefficients and the four meta-variables and number of data used in the 65 hedonic price models have been analyzed in this study. The six factors that are most important in deciding the prices of apartments are positioning of apartments, the noise of the apartments, points of the compass and views from the apartments, proximity to the public transportations, companies that have constructed the apartments, social environments (such as schools etc.).

Keywords: hedonic price model, housing price, meta-regression analysis, characteristics

Procedia PDF Downloads 403
6812 Convolutional Neural Networks Architecture Analysis for Image Captioning

Authors: Jun Seung Woo, Shin Dong Ho

Abstract:

The Image Captioning models with Attention technology have developed significantly compared to previous models, but it is still unsatisfactory in recognizing images. We perform an extensive search over seven interesting Convolutional Neural Networks(CNN) architectures to analyze the behavior of different models for image captioning. We compared seven different CNN Architectures, according to batch size, using on public benchmarks: MS-COCO datasets. In our experimental results, DenseNet and InceptionV3 got about 14% loss and about 160sec training time per epoch. It was the most satisfactory result among the seven CNN architectures after training 50 epochs on GPU.

Keywords: deep learning, image captioning, CNN architectures, densenet, inceptionV3

Procedia PDF Downloads 134
6811 Models and Metamodels for Computer-Assisted Natural Language Grammar Learning

Authors: Evgeny Pyshkin, Maxim Mozgovoy, Vladislav Volkov

Abstract:

The paper follows a discourse on computer-assisted language learning. We examine problems of foreign language teaching and learning and introduce a metamodel that can be used to define learning models of language grammar structures in order to support teacher/student interaction. Special attention is paid to the concept of a virtual language lab. Our approach to language education assumes to encourage learners to experiment with a language and to learn by discovering patterns of grammatically correct structures created and managed by a language expert.

Keywords: computer-assisted instruction, language learning, natural language grammar models, HCI

Procedia PDF Downloads 521
6810 Automatic Calibration of Agent-Based Models Using Deep Neural Networks

Authors: Sima Najafzadehkhoei, George Vega Yon

Abstract:

This paper presents an approach for calibrating Agent-Based Models (ABMs) efficiently, utilizing Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks. These machine learning techniques are applied to Susceptible-Infected-Recovered (SIR) models, which are a core framework in the study of epidemiology. Our method replicates parameter values from observed trajectory curves, enhancing the accuracy of predictions when compared to traditional calibration techniques. Through the use of simulated data, we train the models to predict epidemiological parameters more accurately. Two primary approaches were explored: one where the number of susceptible, infected, and recovered individuals is fully known, and another using only the number of infected individuals. Our method shows promise for application in other ABMs where calibration is computationally intensive and expensive.

Keywords: ABM, calibration, CNN, LSTM, epidemiology

Procedia PDF Downloads 27
6809 Brain Tumor Detection and Classification Using Pre-Trained Deep Learning Models

Authors: Aditya Karade, Sharada Falane, Dhananjay Deshmukh, Vijaykumar Mantri

Abstract:

Brain tumors pose a significant challenge in healthcare due to their complex nature and impact on patient outcomes. The application of deep learning (DL) algorithms in medical imaging have shown promise in accurate and efficient brain tumour detection. This paper explores the performance of various pre-trained DL models ResNet50, Xception, InceptionV3, EfficientNetB0, DenseNet121, NASNetMobile, VGG19, VGG16, and MobileNet on a brain tumour dataset sourced from Figshare. The dataset consists of MRI scans categorizing different types of brain tumours, including meningioma, pituitary, glioma, and no tumour. The study involves a comprehensive evaluation of these models’ accuracy and effectiveness in classifying brain tumour images. Data preprocessing, augmentation, and finetuning techniques are employed to optimize model performance. Among the evaluated deep learning models for brain tumour detection, ResNet50 emerges as the top performer with an accuracy of 98.86%. Following closely is Xception, exhibiting a strong accuracy of 97.33%. These models showcase robust capabilities in accurately classifying brain tumour images. On the other end of the spectrum, VGG16 trails with the lowest accuracy at 89.02%.

Keywords: brain tumour, MRI image, detecting and classifying tumour, pre-trained models, transfer learning, image segmentation, data augmentation

Procedia PDF Downloads 74
6808 Continuum-Based Modelling Approaches for Cell Mechanics

Authors: Yogesh D. Bansod, Jiri Bursa

Abstract:

The quantitative study of cell mechanics is of paramount interest since it regulates the behavior of the living cells in response to the myriad of extracellular and intracellular mechanical stimuli. The novel experimental techniques together with robust computational approaches have given rise to new theories and models, which describe cell mechanics as a combination of biomechanical and biochemical processes. This review paper encapsulates the existing continuum-based computational approaches that have been developed for interpreting the mechanical responses of living cells under different loading and boundary conditions. The salient features and drawbacks of each model are discussed from both structural and biological points of view. This discussion can contribute to the development of even more precise and realistic computational models of cell mechanics based on continuum approaches or on their combination with microstructural approaches, which in turn may provide a better understanding of mechanotransduction in living cells.

Keywords: cell mechanics, computational models, continuum approach, mechanical models

Procedia PDF Downloads 363
6807 Exploring Artistic Creation and Autoethnography in the Spatial Context of Geography

Authors: Sinem Tas

Abstract:

This research paper attempts to study the perspective of personal experience in relation to spatial dynamics and artistic outcomes within the realm of cultural identity. This article serves as a partial analysis within a broader PhD investigation that focuses on the cultural dynamics and political structures behind cultural identity through an autoethnography of narrative while presenting its correlation with artistic creation in the context of space and people. Focusing on the artistic/creative practice project AUTRUI, the primary goal is to analyse and understand the influence of personal experiences and culturally constructed identity as an artist in resulting in the compositional modality of the last image considering self-reflective experience. Referencing the works of Joyce Davidson and Christine Milligan - the scholars who emphasise the importance of emotion and spatial experience in geographical studies contribute to this work as they highlight the significance of emotion across various spatial scales in their work Embodying Emotion Sensing Space: Introducing Emotional Geographies (2004). Their perspective suggests that understanding emotions within different spatial contexts is crucial for comprehending human experiences and interactions with space. Incorporating the insights of scholars like Yi-Fu Tuan, particularly his seminal work Space and Place: The Perspective of Experience (1979), is important for creating an in-depth frame of geographical experience. Tuan's humanistic perspective on space and place provides a valuable theoretical framework for understanding the interplay between personal experiences and spatial contexts. A substantial contextualisation of the geopolitics of Turkey - the implications for national identity and cohesion - will be addressed by drawing an outline of the political and geographical frame as a methodological strategy to understand the dynamics behind this research. Besides the bibliographical reading, the methods used to study this relation are participatory observation, memory work along with memoir analysis, personal interviews, and discussion of photographs and news. The utilisation of the self as data requires the analysis of the written sources with personal engagement. By delving into written sources such as written communications or diaries as well as memoirs, the research gains a firsthand perspective, enriching the analytical depth of the study. Furthermore, the examination of photography and news articles serves as a valuable means of contextualising experiences from a journalist's background within specific geographical settings. The inclusion of interviews with close family members access provides firsthand perspectives and intimate insights rooted in shared experiences within similar geographical contexts, offering complementary insights and diversified viewpoints, enhancing the comprehensiveness of the investigation.

Keywords: art, autoethnography, place and space, Turkey

Procedia PDF Downloads 50
6806 Evaluation and Compression of Different Language Transformer Models for Semantic Textual Similarity Binary Task Using Minority Language Resources

Authors: Ma. Gracia Corazon Cayanan, Kai Yuen Cheong, Li Sha

Abstract:

Training a language model for a minority language has been a challenging task. The lack of available corpora to train and fine-tune state-of-the-art language models is still a challenge in the area of Natural Language Processing (NLP). Moreover, the need for high computational resources and bulk data limit the attainment of this task. In this paper, we presented the following contributions: (1) we introduce and used a translation pair set of Tagalog and English (TL-EN) in pre-training a language model to a minority language resource; (2) we fine-tuned and evaluated top-ranking and pre-trained semantic textual similarity binary task (STSB) models, to both TL-EN and STS dataset pairs. (3) then, we reduced the size of the model to offset the need for high computational resources. Based on our results, the models that were pre-trained to translation pairs and STS pairs can perform well for STSB task. Also, having it reduced to a smaller dimension has no negative effect on the performance but rather has a notable increase on the similarity scores. Moreover, models that were pre-trained to a similar dataset have a tremendous effect on the model’s performance scores.

Keywords: semantic matching, semantic textual similarity binary task, low resource minority language, fine-tuning, dimension reduction, transformer models

Procedia PDF Downloads 211
6805 A Comparative Analysis of ARIMA and Threshold Autoregressive Models on Exchange Rate

Authors: Diteboho Xaba, Kolentino Mpeta, Tlotliso Qejoe

Abstract:

This paper assesses the in-sample forecasting of the South African exchange rates comparing a linear ARIMA model and a SETAR model. The study uses a monthly adjusted data of South African exchange rates with 420 observations. Akaike information criterion (AIC) and the Schwarz information criteria (SIC) are used for model selection. Mean absolute error (MAE), root mean squared error (RMSE) and mean absolute percentage error (MAPE) are error metrics used to evaluate forecast capability of the models. The Diebold –Mariano (DM) test is employed in the study to check forecast accuracy in order to distinguish the forecasting performance between the two models (ARIMA and SETAR). The results indicate that both models perform well when modelling and forecasting the exchange rates, but SETAR seemed to outperform ARIMA.

Keywords: ARIMA, error metrices, model selection, SETAR

Procedia PDF Downloads 244
6804 A Trend Based Forecasting Framework of the ATA Method and Its Performance on the M3-Competition Data

Authors: H. Taylan Selamlar, I. Yavuz, G. Yapar

Abstract:

It is difficult to make predictions especially about the future and making accurate predictions is not always easy. However, better predictions remain the foundation of all science therefore the development of accurate, robust and reliable forecasting methods is very important. Numerous number of forecasting methods have been proposed and studied in the literature. There are still two dominant major forecasting methods: Box-Jenkins ARIMA and Exponential Smoothing (ES), and still new methods are derived or inspired from them. After more than 50 years of widespread use, exponential smoothing is still one of the most practically relevant forecasting methods available due to their simplicity, robustness and accuracy as automatic forecasting procedures especially in the famous M-Competitions. Despite its success and widespread use in many areas, ES models have some shortcomings that negatively affect the accuracy of forecasts. Therefore, a new forecasting method in this study will be proposed to cope with these shortcomings and it will be called ATA method. This new method is obtained from traditional ES models by modifying the smoothing parameters therefore both methods have similar structural forms and ATA can be easily adapted to all of the individual ES models however ATA has many advantages due to its innovative new weighting scheme. In this paper, the focus is on modeling the trend component and handling seasonality patterns by utilizing classical decomposition. Therefore, ATA method is expanded to higher order ES methods for additive, multiplicative, additive damped and multiplicative damped trend components. The proposed models are called ATA trended models and their predictive performances are compared to their counter ES models on the M3 competition data set since it is still the most recent and comprehensive time-series data collection available. It is shown that the models outperform their counters on almost all settings and when a model selection is carried out amongst these trended models ATA outperforms all of the competitors in the M3- competition for both short term and long term forecasting horizons when the models’ forecasting accuracies are compared based on popular error metrics.

Keywords: accuracy, exponential smoothing, forecasting, initial value

Procedia PDF Downloads 177
6803 Advancing Communication Theory in the Age of Digital Technology: Bridging the Gap Between Traditional Models and Emerging Platforms

Authors: Sidique Fofanah

Abstract:

This paper explores the intersection of traditional communication theories and modern digital technologies, analyzing how established models adapt to contemporary communication platforms. It examines the evolving nature of interpersonal, group, and mass communication within digital environments, emphasizing the role of social media, AI-driven communication tools, and virtual reality in reshaping communication paradigms. The paper also discusses the implications for future research and practice in communication studies, proposing an integrated framework that accommodates both classical and emerging theories.

Keywords: communication, traditional models, emerging platforms, digital media

Procedia PDF Downloads 26
6802 Mathematical Modeling of Carotenoids and Polyphenols Content of Faba Beans (Vicia faba L.) during Microwave Treatments

Authors: Ridha Fethi Mechlouch, Ahlem Ayadi, Ammar Ben Brahim

Abstract:

Given the importance of the preservation of polyphenols and carotenoids during thermal processing, we attempted in this study to investigate the variation of these two parameters in faba beans during microwave treatment using different power densities (1; 2; and 3W/g), then to perform a mathematical modeling by using non-linear regression analysis to evaluate the models constants. The variation of the carotenoids and polyphenols ratio of faba beans and the models are tested to validate the experimental results. Exponential models were found to be suitable to describe the variation of caratenoid ratio (R²= 0.945, 0.927 and 0.946) for power densities (1; 2; and 3W/g) respectively, and polyphenol ratio (R²= 0.931, 0.989 and 0.982) for power densities (1; 2; and 3W/g) respectively. The effect of microwave power density Pd(W/g) on the coefficient k of models were also investigated. The coefficient is highly correlated (R² = 1) and can be expressed as a polynomial function.

Keywords: microwave treatment, power density, carotenoid, polyphenol, modeling

Procedia PDF Downloads 259
6801 Exchange Rate Forecasting by Econometric Models

Authors: Zahid Ahmad, Nosheen Imran, Nauman Ali, Farah Amir

Abstract:

The objective of the study is to forecast the US Dollar and Pak Rupee exchange rate by using time series models. For this purpose, daily exchange rates of US and Pakistan for the period of January 01, 2007 - June 2, 2017, are employed. The data set is divided into in sample and out of sample data set where in-sample data are used to estimate as well as forecast the models, whereas out-of-sample data set is exercised to forecast the exchange rate. The ADF test and PP test are used to make the time series stationary. To forecast the exchange rate ARIMA model and GARCH model are applied. Among the different Autoregressive Integrated Moving Average (ARIMA) models best model is selected on the basis of selection criteria. Due to the volatility clustering and ARCH effect the GARCH (1, 1) is also applied. Results of analysis showed that ARIMA (0, 1, 1 ) and GARCH (1, 1) are the most suitable models to forecast the future exchange rate. Further the GARCH (1,1) model provided the volatility with non-constant conditional variance in the exchange rate with good forecasting performance. This study is very useful for researchers, policymakers, and businesses for making decisions through accurate and timely forecasting of the exchange rate and helps them in devising their policies.

Keywords: exchange rate, ARIMA, GARCH, PAK/USD

Procedia PDF Downloads 562
6800 Study on Flexible Diaphragm In-Plane Model of Irregular Multi-Storey Industrial Plant

Authors: Cheng-Hao Jiang, Mu-Xuan Tao

Abstract:

The rigid diaphragm model may cause errors in the calculation of internal forces due to neglecting the in-plane deformation of the diaphragm. This paper thus studies the effects of different diaphragm in-plane models (including in-plane rigid model and in-plane flexible model) on the seismic performance of structures. Taking an actual industrial plant as an example, the seismic performance of the structure is predicted using different floor diaphragm models, and the analysis errors caused by different diaphragm in-plane models including deformation error and internal force error are calculated. Furthermore, the influence of the aspect ratio on the analysis errors is investigated. Finally, the code rationality is evaluated by assessing the analysis errors of the structure models whose floors were determined as rigid according to the code’s criterion. It is found that different floor models may cause great differences in the distribution of structural internal forces, and the current code may underestimate the influence of the floor in-plane effect.

Keywords: industrial plant, diaphragm, calculating error, code rationality

Procedia PDF Downloads 140
6799 Probing Language Models for Multiple Linguistic Information

Authors: Bowen Ding, Yihao Kuang

Abstract:

In recent years, large-scale pre-trained language models have achieved state-of-the-art performance on a variety of natural language processing tasks. The word vectors produced by these language models can be viewed as dense encoded presentations of natural language that in text form. However, it is unknown how much linguistic information is encoded and how. In this paper, we construct several corresponding probing tasks for multiple linguistic information to clarify the encoding capabilities of different language models and performed a visual display. We firstly obtain word presentations in vector form from different language models, including BERT, ELMo, RoBERTa and GPT. Classifiers with a small scale of parameters and unsupervised tasks are then applied on these word vectors to discriminate their capability to encode corresponding linguistic information. The constructed probe tasks contain both semantic and syntactic aspects. The semantic aspect includes the ability of the model to understand semantic entities such as numbers, time, and characters, and the grammatical aspect includes the ability of the language model to understand grammatical structures such as dependency relationships and reference relationships. We also compare encoding capabilities of different layers in the same language model to infer how linguistic information is encoded in the model.

Keywords: language models, probing task, text presentation, linguistic information

Procedia PDF Downloads 114
6798 Application Difference between Cox and Logistic Regression Models

Authors: Idrissa Kayijuka

Abstract:

The logistic regression and Cox regression models (proportional hazard model) at present are being employed in the analysis of prospective epidemiologic research looking into risk factors in their application on chronic diseases. However, a theoretical relationship between the two models has been studied. By definition, Cox regression model also called Cox proportional hazard model is a procedure that is used in modeling data regarding time leading up to an event where censored cases exist. Whereas the Logistic regression model is mostly applicable in cases where the independent variables consist of numerical as well as nominal values while the resultant variable is binary (dichotomous). Arguments and findings of many researchers focused on the overview of Cox and Logistic regression models and their different applications in different areas. In this work, the analysis is done on secondary data whose source is SPSS exercise data on BREAST CANCER with a sample size of 1121 women where the main objective is to show the application difference between Cox regression model and logistic regression model based on factors that cause women to die due to breast cancer. Thus we did some analysis manually i.e. on lymph nodes status, and SPSS software helped to analyze the mentioned data. This study found out that there is an application difference between Cox and Logistic regression models which is Cox regression model is used if one wishes to analyze data which also include the follow-up time whereas Logistic regression model analyzes data without follow-up-time. Also, they have measurements of association which is different: hazard ratio and odds ratio for Cox and logistic regression models respectively. A similarity between the two models is that they are both applicable in the prediction of the upshot of a categorical variable i.e. a variable that can accommodate only a restricted number of categories. In conclusion, Cox regression model differs from logistic regression by assessing a rate instead of proportion. The two models can be applied in many other researches since they are suitable methods for analyzing data but the more recommended is the Cox, regression model.

Keywords: logistic regression model, Cox regression model, survival analysis, hazard ratio

Procedia PDF Downloads 457
6797 Comparison of Wake Oscillator Models to Predict Vortex-Induced Vibration of Tall Chimneys

Authors: Saba Rahman, Arvind K. Jain, S. D. Bharti, T. K. Datta

Abstract:

The present study compares the semi-empirical wake-oscillator models that are used to predict vortex-induced vibration of structures. These models include those proposed by Facchinetti, Farshidian, and Dolatabadi, and Skop and Griffin. These models combine a wake oscillator model resembling the Van der Pol oscillator model and a single degree of freedom oscillation model. In order to use these models for estimating the top displacement of chimneys, the first mode vibration of the chimneys is only considered. The modal equation of the chimney constitutes the single degree of freedom model (SDOF). The equations of the wake oscillator model and the SDOF are simultaneously solved using an iterative procedure. The empirical parameters used in the wake-oscillator models are estimated using a newly developed approach, and response is compared with experimental data, which appeared comparable. For carrying out the iterative solution, the ode solver of MATLAB is used. To carry out the comparative study, a tall concrete chimney of height 210m has been chosen with the base diameter as 28m, top diameter as 20m, and thickness as 0.3m. The responses of the chimney are also determined using the linear model proposed by E. Simiu and the deterministic model given in Eurocode. It is observed from the comparative study that the responses predicted by the Facchinetti model and the model proposed by Skop and Griffin are nearly the same, while the model proposed by Fashidian and Dolatabadi predicts a higher response. The linear model without considering the aero-elastic phenomenon provides a less response as compared to the non-linear models. Further, for large damping, the prediction of the response by the Euro code is relatively well compared to those of non-linear models.

Keywords: chimney, deterministic model, van der pol, vortex-induced vibration

Procedia PDF Downloads 221
6796 Analysis of Moving Loads on Bridges Using Surrogate Models

Authors: Susmita Panda, Arnab Banerjee, Ajinkya Baxy, Bappaditya Manna

Abstract:

The design of short to medium-span high-speed bridges in critical locations is an essential aspect of vehicle-bridge interaction. Due to dynamic interaction between moving load and bridge, mathematical models or finite element modeling computations become time-consuming. Thus, to reduce the computational effort, a universal approximator using an artificial neural network (ANN) has been used to evaluate the dynamic response of the bridge. The data set generation and training of surrogate models have been conducted over the results obtained from mathematical modeling. Further, the robustness of the surrogate model has been investigated, which showed an error percentage of less than 10% with conventional methods. Additionally, the dependency of the dynamic response of the bridge on various load and bridge parameters has been highlighted through a parametric study.

Keywords: artificial neural network, mode superposition method, moving load analysis, surrogate models

Procedia PDF Downloads 100
6795 Effects of a School-based Mindfulness Intervention on Stress Levels and Emotion Regulation of Adolescent Students Enrolled in an Independent School

Authors: Tracie Catlett

Abstract:

Students enrolled in high-achieving schools are under tremendous pressure to perform at high levels inside and outside the classroom. Achievement pressure is a prevalent source of stress for students enrolled in high-achieving schools, and female students, in particular, experience a higher frequency and higher levels of stress compared to their male peers. The practice of mindfulness in a school setting is one tool that has been linked to improved self-regulation of emotions, increased positive emotions, and stress reduction. A mixed methods randomized pretest-posttest no-treatment control trial evaluated the effects of a six-session mindfulness intervention taught during a regularly scheduled life skills period in an independent day school, one type of high-achieving school. Twenty-nine students in Grades 10 and 11 were randomized by class, where Grade 11 students were in the intervention group (n = 14) and Grade 10 students were in the control group (n = 15). Findings from the study produced mixed results. There was no evidence that the mindfulness program reduced participants’ stress levels and negative emotions. In fact, contrary to what was expected, students enrolled in the intervention group experienced higher levels of stress and increased negative emotions at posttreatment when compared to pretreatment. Neither the within-group nor the between-groups changes in stress level were statistically significant, p > .05, and the between-groups effect size was small, d = .2. The study found evidence that the mindfulness program may have had a positive impact on students’ ability to regulate their emotions. The within-group comparison and the between-groups comparison at posttreatment found that students in the mindfulness course experienced statistically significant improvement in the in their ability to regulate their emotions at posttreatment, p = .009 < .05 and p =. 034 < .05, respectively. The between-groups effect size was medium, d =.7, suggesting that the positive differences in emotion regulation difficulties were substantial and have practical implications. The analysis of gender differences, as they relate to stress and emotions, revealed that female students perceive higher levels of stress and report experiencing stress more often than males. There were no gender differences when analyzing sources of stress experienced by the student participants. Both females and males experience regular achievement pressures related to their school performance and worry about their future, college acceptance, grades, and parental expectations. Females reported an increased awareness of their stress and actively engaged in practicing mindfulness to manage their stress. Students in the treatment group expressed that the practice of mindfulness resulted in feelings of relaxation and calmness.

Keywords: achievement pressure, adolescents, emotion regulation, emotions, high-achieving schools, independent schools, mindfulness, negative affect, positive affect, stress

Procedia PDF Downloads 63
6794 Applying Multiplicative Weight Update to Skin Cancer Classifiers

Authors: Animish Jain

Abstract:

This study deals with using Multiplicative Weight Update within artificial intelligence and machine learning to create models that can diagnose skin cancer using microscopic images of cancer samples. In this study, the multiplicative weight update method is used to take the predictions of multiple models to try and acquire more accurate results. Logistic Regression, Convolutional Neural Network (CNN), and Support Vector Machine Classifier (SVMC) models are employed within the Multiplicative Weight Update system. These models are trained on pictures of skin cancer from the ISIC-Archive, to look for patterns to label unseen scans as either benign or malignant. These models are utilized in a multiplicative weight update algorithm which takes into account the precision and accuracy of each model through each successive guess to apply weights to their guess. These guesses and weights are then analyzed together to try and obtain the correct predictions. The research hypothesis for this study stated that there would be a significant difference in the accuracy of the three models and the Multiplicative Weight Update system. The SVMC model had an accuracy of 77.88%. The CNN model had an accuracy of 85.30%. The Logistic Regression model had an accuracy of 79.09%. Using Multiplicative Weight Update, the algorithm received an accuracy of 72.27%. The final conclusion that was drawn was that there was a significant difference in the accuracy of the three models and the Multiplicative Weight Update system. The conclusion was made that using a CNN model would be the best option for this problem rather than a Multiplicative Weight Update system. This is due to the possibility that Multiplicative Weight Update is not effective in a binary setting where there are only two possible classifications. In a categorical setting with multiple classes and groupings, a Multiplicative Weight Update system might become more proficient as it takes into account the strengths of multiple different models to classify images into multiple categories rather than only two categories, as shown in this study. This experimentation and computer science project can help to create better algorithms and models for the future of artificial intelligence in the medical imaging field.

Keywords: artificial intelligence, machine learning, multiplicative weight update, skin cancer

Procedia PDF Downloads 80
6793 Chemometric Estimation of Inhibitory Activity of Benzimidazole Derivatives by Linear Least Squares and Artificial Neural Networks Modelling

Authors: Sanja O. Podunavac-Kuzmanović, Strahinja Z. Kovačević, Lidija R. Jevrić, Stela Jokić

Abstract:

The subject of this paper is to correlate antibacterial behavior of benzimidazole derivatives with their molecular characteristics using chemometric QSAR (Quantitative Structure–Activity Relationships) approach. QSAR analysis has been carried out on the inhibitory activity of benzimidazole derivatives against Staphylococcus aureus. The data were processed by linear least squares (LLS) and artificial neural network (ANN) procedures. The LLS mathematical models have been developed as a calibration models for prediction of the inhibitory activity. The quality of the models was validated by leave one out (LOO) technique and by using external data set. High agreement between experimental and predicted inhibitory acivities indicated the good quality of the derived models. These results are part of the CMST COST Action No. CM1306 "Understanding Movement and Mechanism in Molecular Machines".

Keywords: Antibacterial, benzimidazoles, chemometric, QSAR.

Procedia PDF Downloads 317
6792 Fusion of MOLA-based DEMs and HiRISE Images for Large-Scale Mars Mapping

Authors: Ahmed F. Elaksher, Islam Omar

Abstract:

In this project, we used MOLA-based DEMs to orthorectify HiRISE optical images. The MOLA data was interpolated using the kriging interpolation technique. Corresponding tie points were then digitized from both datasets. These points were employed in co-registering both datasets using GIS analysis tools. Different transformation models, including the affine and projective transformation models, were used with different sets and distributions of tie points. Additionally, we evaluated the use of the MOLA elevations in co-registering the MOLA and HiRISE datasets. The planimetric RMSEs achieved for each model are reported. Results suggested the use of 3D-2D transformation models.

Keywords: photogrammetry, Mars, MOLA, HiRISE

Procedia PDF Downloads 78