7264 research outputs found
Sort by
AUTISM SPECTRUM DISORDER DETECTION USING MACHINE LEARNING
No full textThis article examines the visual preferences of autistic children in order to identify specific patterns, such as repetitive behavior, and focus on certain elements of the visual content, such as geometric shapes, etc. To analyze visual preferences, the research team collected the experimental data of two groups of children: those diagnosed with Autism Spectrum Disorders and typically developing children. Based on the received data, a model was trained to detect autism with the usage of machine learning. In addition, the machine was safely tested on children and showed the possibility of detecting Autism Spectrum Disorders in 40% of children with autism. The study was conducted on a web platform specially designed for the young audience, which allows them to track the direction of their gaze. The obtained results also indicate that children with autism give visual preference to geometric shapes with dynamic scene changes. The implementation of this system will be useful for early detection of Autism Spectrum Disorders due to the wide accessibility of this web platform and its beneficence as a reliable screening tool. The aim of the research is to create an innovative software that will provide an opportunity to identify Autism Spectrum Disorder using machine learning
A COST-EFFICIENT APPROACH FOR EV ENERGY MANAGEMENT IN A MICROGRID
No full textThis thesis introduces a particle swarm optimization (PSO)-based energy management system (EMS) integrated into a microgrid (MG) infrastructure for charging electric vehicles (EV). The primary objective of this study is to enhance the efficiency of integrating locally produced renewable energy, power from the grid, and battery energy storage systems (BESS) to achieve cost-effectiveness and optimize the usage of sustainable resources within a community context. The design of the MG incorporates several components, including wind turbines (WT), solar panels (PV), and EV charging stations. The demand and energy generation profiles are modeled and used as MG inputs for analyzing 16 different scenarios with four seasons. The results demonstrate the versatility and effectiveness of the PSO-based EMS in attaining a financially viable MG system that satisfies the hourly energy demand criteria. This study aims to thoroughly comprehend the influence of the optimized EMS on EV charging schedules and MG performance. The findings of this research provide valuable insights for the development of sustainable and economically feasible EMS in community settings. Overall, the use of PSO in the proposed EMS for EV charging and battery charging/discharging is based on electricity price. Additionally, the results show that the implementation of PSO-based optimization results in a cost reduction of 21% and 16% in winter, 17% and 8% in spring, 12% and 9% in autumn, 21% and 14% in summer with and without EV consideration, respectively. These findings highlight the effectiveness of the PSO optimization technique in efficiently adjusting to various energy demand patterns
QAZAQ SIGN LANGUAGE: BRIDGING COMMUNICATION BARRIERS
No full textThe project is aimed to address the issue of communication barriers encountered by individuals with hearing disabilities, given the scarcity of resources in sign language for people in Kazakhstan. The project tackles the issue by developing a web application that provides the translation of spoken language (Kazakh, English, and Russian) to Kazakh-Russian signed language. The project aims to facilitate accessibility and communication of the deaf community by utilizing Language Technology, specifically the Sign Language Synthesis, Sign Language Translation, and Natural Language Processing. Additionally, the web application expands educational opportunities of the individuals with hearing disabilities by providing a book library, which could be also translated to sign language
BALANCING WORK AND STUDY AT NEO-LIBERAL UNIVERSITY: MOTIVATIONS AND SOCIAL EFFECTS OF PART-TIME EMPLOYMENT AMONG NU UNDERGRADUATE STUDENTS
No full textCapstone project investigates the motivations of Nazarbayev University (NU) students for engaging in part-time work with their studies, within the consideration of the university as a neo-liberal institution. The study explores the influence of neoliberal policies shaped by Kazakhstan's post-Soviet political climate under President Nursultan Nazarbayev, which emphasize privatization, meritocracy, individualism, self-reliance, and competition. It assesses the effects of these policies on the educational system overall and thus students' academic performance, personal life, health, and career opportunities through seven semi-structured interviews. The project employs three main theoretical perspectives: Self-Determination Theory to evaluate personal motivations for work; The New Spirit of Capitalism to understand corporate expectations and self-exploitation; and Ecological Systems Theory to analyze socio-ecological influences on students' decisions. The findings aim to provide insights into how a neoliberal educational environment like at NU, which prioritizes corporate competitiveness and professional flexibility, motivates students to work and impacts their lives and career paths
HOMELAND REEMERGED: EXPLORING CHANGES IN UZBEK’S ENGAGEMENT WITH KIN STATE POLITICS AND HOST STATE LOYALTY IN SAIRAM
No full textBeing cut off from Uzbekistan due to a lack of welcoming attitude and restrictive policies that overlooked the question of co-ethnics outside its borders, Uzbeks in Sairam district evolved and developed its identity separately from it during the first decades of independence. This relationship dynamic was stable until the election of the second president of Uzbekistan, Mirziyoyev, who drastically changed the state’s approach to dealing with co-ethnics abroad. Decisive steps in this direction involved the introduction of the Decree on Compatriots, the first decree defining Uzbekistan’s position towards co-ethnics abroad. This decree aimed to foster ties with Uzbeks living abroad, it redefined Uzbekistan’s political boundaries in a more flexible and inclusive definition of it, including co-ethnics living abroad. This re-emergence of national homeland and its effect on the stance of Uzbeks in Sairam is explored through Brubaker’s (1995) theoretical framework of triadic nexus which emphasized the dynamic nature of this triangular relationship between kin state, host state, and minority ethnic group. Analyzing the recent changes in Uzbekistan’s stance towards co-ethnics through this model in the context of the re-emergence of ethnic homeland after the straining of ties, it was revealed that activists in the community noticed the positive changes that resulted from the adoption of this decree. However, these changes are not part of the discussion among Uzbeks of this community, who are largely unaware of this decree and Uzbekistan’s change of stance. Instead, since the introduction of the decree went hand in hand with tightening cooperation between Uzbekistan and Kazakhstan, the changes were largely seen as the outcome of this friendly relationship between the two states. Therefore, this community's participation in joining cultural activities and intensified cooperation were perceived not as part of building ties with the kin state, but as supporting the host state’s endeavors in establishing a friendlier relationship with a neighboring state. Therefore, the re-
iv
emergence of the national homeland did not affect the existing interaction model between Uzbek’s in Kazakhstan and Kazakhstan’s government
SYNTHESIS AND CHARACTERIZATION OF NATURAL SURFACTANT FOR ENHANCED OIL RECOVERY APPLICATION
No full textEnhanced Oil Recovery (EOR) techniques rely on chemical agents, but the excessive utilization of synthetic chemicals presents environmental and economic issues. Natural surfactants have become known as a viable option for Chemically Enhanced Oil Recovery (CEOR) processes. These natural agents are being acknowledged for being eco-friendly, less harmful, and cost-efficient in comparison to conventional synthetic surfactants. However, ongoing research and investigation are necessary to determine the extent to which natural surfactants can be effectively utilized across a range of conditions.
The study seeks to illustrate that surfactants derived from neem powder possess the capability to function as a viable replacement for conventional surfactants in oil recovery procedures. The research commences by creating a surfactant sourced from sustainable materials, mainly neem powder and olive oil, in conjunction with sodium hydroxide (NaOH). The synthesis procedure entails an infusion of neem powder into olive oil, subsequently followed by the saponification reaction with NaOH. The resultant surfactant presents the benefit of being environmentally friendly and economically feasible.
In this research, the efficacy of the natural surfactant was evaluated through experimental assessments, including measurements of interfacial tension (IFT) and core flooding experiments. After confirming the natural surfactant's properties through FTIR, SEM, and EDS analyses, solutions of the surfactant were formulated at different concentrations. These concentrations ranged from 1 to 6 wt% in deionized water (DIW) and from 0,5 to 5 wt% in brine. The IFT findings indicated that the optimal critical micelle concentration (CMC) for the natural surfactant was determined to be 4,0 wt% in deionized water (DIW) and 0,9 wt% in brine. The brine-based surfactant solution reached a minimum IFT of 1,4 mN/m at the CMC point. The analysis investigated the influence of temperatures ranging from 35 to 55°C on the surfactant's efficacy, showcasing its consistent performance across the temperature range.
This study offers a valuable understanding of the possibilities of utilizing natural surfactants in EOR and underscores the necessity for additional research to enhance their effectiveness in practical applications
THE PERCEPTIONS OF UNDERGRADUATE STUDENTS REGARDING ACADEMIC STUDENT SUPPORT CENTERS
No full textAcademic student support is one of the essential services provided by higher education institutions (HEIs) in order to meet students’ academic needs and requirements. During their studies, students might encounter various academic obstacles or misunderstandings that require institutional support and help which, in turn, might shape students’ satisfaction with their university. Even though the phenomenon of providing academic support has been practicing among universities for years in different corners of the world, Kazakhstan has started to be exposed to it not long ago. Due to this, academic support and advising were poorly introduced and integrated in the context of Kazakhstan. Therefore, the study aims to explore undergraduate students’ perceptions of academic support centers and their services as a result of utilizing them. The study also identifies the main reasons for academic help-seeking and ways of improving these centers.
The study was based on the qualitative research design with the phenomenological approach to focus on shared experiences. For the study, ten undergraduate students who used academic support centers were recruited, and semi-structured interviews were conducted to understand the students’ perspectives on these centers and services. The findings revealed that students at this particular university were mostly satisfied with the support services, even though there were several difficulties that arose during the process of using support services. It was also found that students had different reasons of seeking help like organizational, academic, and personal. The main recommendation made were increasing the availability of academic support and focusing more on providing support with their course materials. Considering all of the findings, there were some more suggestions made in improving the services of the academic support centers
STATISTICAL METHODS IN NATURAL LANGUAGE PROCESSING
Get PDFThis capstone project explores the application of statistical method
ologies to two distinct natural language processing (NLP) tasks: machine
translation between Ukrainian and Russian languages and the classifica
tion of comments for hate speech detection. The study shows that the
strategic integration of statistical approaches can improve performance
of the machine translation and text classification problems. The imple
mentation of linear regression with an added orthogonal constraint on
weight vectors has resulted in higher precision scores. For the classifi
cation of hate speech within textual comments, logistic regression with
TF-IDF features was identified as the the most effective model in terms
of AUC-ROC metric
SKELETON-BASED HUMAN ACTION RECOGNITION USING CONTRASTIVE LEARNING
No full textSkeletal-based action detection is gaining importance in the realm of human activity analysis because it provides a foundational understanding for applications ranging from surveillance to human-computer interaction. In light of the difficulties caused by the dearth of labeled datasets in this area, our research emphasizes the vital significance of self-supervised learning, emphasizing contrastive learning strategies in particular.By advocating for pretraining models with unlabeled data, we demon- strate how self-supervised learning can significantly enhance the model’s ability to identify human actions without the requirement for extensive annotated datasets. In contrastive learning, the encoder is essential because it converts raw input into repre- sentations that can discriminate between action sequences that are similar (positive) and different (negative). Our novel method incorporates Transformers into the MoCo (Momentum Contrast) contrastive learning framework as encoder. In addition to de- parting from conventional supervised learning techniques, this combination makes use of Transformers’ advantages to more effectively capture the intricate spatial-temporal correlations present in skeletal data. By including Transformer-based encoders into the MoCo framework, skeleton-based action detection has advanced significantly and new efficiency and accuracy benchmarks have been reached. Our results show the sig- nificant advantages of integrating sophisticated encoding methods with self-supervised learning, opening the door to more complex and less data-dependent assessments of human behavior
EXPLORING DATA DISTRIBUTION AND VALUE FUNCTION APPROXIMATION IMPACTS IN OFFLINE REINFORCEMENT LEARNING(RL): FROM GRIDWORLD ENVIRONMENTS
No full textIn the emerging landscape of off-policy reinforcement learning (RL), challenges arise
due to the significant costs and risks tied to data collection. To address these issues,
there is an alternative path for transitioning RL from off-policy to offline, which is
known for its fixed data collection practices. This stands in contrast to online algorithms, which are sensitive to changes in data during the learning phase. However,
the inherent challenge of offline RL lies in its limited interaction with the environment, resulting in inadequate data coverage. Hence, we underscore the convenient
application of offline RL, 1) starting from the collection and preprocessing of a static
dataset from online RL interactions, 2) followed by the training of offline RL models,
and 3) culminating with testing in the same environment as the off-policy RL algorithm. Specifically, the dataset collection involves the utilization of a uniform dataset
gathered systematically via non-arbitrary action selection, covering all possible states
of the environment. Furthermore, we incorporate Q-values into the static dataset,
representing the action distribution across the state-action space. This allows the offline RL model to directly update weights by comparing learned model Q-values with
collected Q-values. Utilizing the proposed approach, the Offline RL model employing
a Multi-Layer Perceptron (MLP) achieves a testing accuracy that falls within 1% of
the results obtained by the off-policy RL agent. Additionally, we provide a practical
guide with datasets, offering valuable tutorials on the application of Offline RL in a
Gridworld-based environment