Search results

Title: SUPPORT VECTOR MACHINE BASED CLASSIFICATION FOR TRAFFIC SIGNS AND ULTRASONIC FLAW DETECTION
Creator: Virupakshappa, Kushal
Date: 2015, 2015-12
Description: The use of machine learning techniques for the advanced signal and image processing applications is gaining importance due to performance...
Show moreThe use of machine learning techniques for the advanced signal and image processing applications is gaining importance due to performance increases in accuracy and robustness. Support Vector Machine (SVM) is a machine learning method used for classification and regression analysis of complex real-world problems that may be difficult to analyze theoretically. In this dissertation, the use of SVM for the application of ultrasonic flaw detection and traffic sign classification has been investigated and new methods are introduced. For traffic sign detection, Bag of visual Words technique has been implemented on Speeded Up Robust Feature (SURF) descriptors of the traffic signs and later the sturdy classifier SVM is used to categorize the traffic signs to its respective groups. Experimental results demonstrate that the proposed method of implementation can reach an accuracy of 95.2 % . For ultrasonic aw detection, subband decomposition filters are used to generate the necessary feature vectors for the SVM classifier. Experimental results, using A-scan data measurements from a steel block, show that a very high classification accuracy can be achieved. Robust performance of the classifier is due to proper selection of frequency-diverse feature vectors and successful training. SVM has also been used for regression analysis to locate and amplify the aw by suppressing the clutter noise. The results show that the use of SVM is reliable and achievable for both the applications.
M.S. in Electrical Engineering, December 2015
Show less

Title: ANALYZING THE LINGUISTIC CHARACTERISTICS OF MARIJUANA USE BY INCOME USING SOCIAL MEDIA
Creator: Zeinali, Sahand
Date: 2018, 2018-05
Description: Marijuana use and legality has been a widely-discussed topic in the recent years. Knowing that marijuana has different effects on health, mood...
Show moreMarijuana use and legality has been a widely-discussed topic in the recent years. Knowing that marijuana has different effects on health, mood and behavior after its use, it is important to understand what the underlying causes for marijuana use also are. As marijuana use is becoming more prevalent every day, it is crucial to know what the motives behind the users' tendencies are for smoking marijuana. To be able to identify the words/patterns associated with marijuana use prior to its use, we will need a real-time method to understand the problem on a deeper level with a better method than surveying users. In our study, we aim to understand the different linguistic characteristics of marijuana users based on their income. Social media's provision of data into understanding and tracking people's behavior can be very beneficial in understanding the contrast between the different social classes prior to marijuana use and understand what the underlying causes are for their marijuana use. In our experiment, we use social media to analyze the patterns and characteristics of marijuana use based on income class. By collecting data on Twitter, we then proceed to classify users based on their income. Using this method, we predict the income of each user by utilizing the user's Twitter activity and their linguistic characteristics based on the tweets associated with them. Through the experiment, we can identify patterns amongst the marijuana users in two different income classes and predict what class a user will be placed in based on their recent Twitter activity with a good accuracy.
M.S. in Computer Science, May 2018
Show less

Title: PRIVACY PRESERVING BAG PREPARATION FOR LEARNING FROM LABEL PROPORTION
Creator: Yan, Xinzhou
Date: 2018
Description: We apply Privacy-preserving data mining standards (PPDM) to the Learning from label proportion (LLP) model to create the Private-preserving...
Show moreWe apply Privacy-preserving data mining standards (PPDM) to the Learning from label proportion (LLP) model to create the Private-preserving machine learning framework. We design the data preparation step for the LLP framework to meet the PPDM standards. In the data preparation step, we develop a bag selection method to boost the accuracy of the LLP model by more than 7%. Besides that, we propose three K- anonymous aggregation methods for the datasets which have almost zero accuracy loss and very robust. After the K-anonymous step, we apply Differential privacy to the LLP model and ensure a low accuracy loss for the LLP modelBecause of the LLP model’s special loss function, not only it is possible to replace all the feature vectors with the mean feature vector within each bag, but also the accuracy loss caused by Differential privacy can be bounded by a small number. The loss function ensures low accuracy loss when training LLP model on PPDM dataset. We evaluate the PPDM LLP model on two datasets, one is the Adult dataset and the other is the Instagram comment dataset. Both of them give empirical evidence of the low accuracy loss after applying the PPDM LLP model.
Show less

Title: A Complete Machine Learning Approach for Predicting Lithium-Ion Cell Combustion
Creator: Almagro Yravedra, Fernando
Date: 2020
Description: The object of the herein thesis work document is to develop a functional predictive model, able to predict the combustion of a US18650 Sony...
Show moreThe object of the herein thesis work document is to develop a functional predictive model, able to predict the combustion of a US18650 Sony Lithium-Ion cell given its current and previous states. In order to build the model, a realistic electro-thermal model of the cell under study is developed in Matlab Simulink, being used to recreate the cell's behavior under a set of real operating conditions. The data generated by the electro-thermal model is used to train a recurrent neural network, which returns the chance of future combustion of the US18650 Sony Lithium-Ion cell. Independently obtained data is used to test and validate the developed recurrent neural network using advanced metrics.
Show less

Title: Towards Assisting Human-Human Conversations
Creator: Nanaware, Tejas Suryakant
Date: 2021
Description: The idea of the research is to understand the open-topic conversations and ways to provide assistance to humans who face difficulties in...
Show moreThe idea of the research is to understand the open-topic conversations and ways to provide assistance to humans who face difficulties in initiating conversations and overcome social anxiety so as to be able to talk and have successful conversations. By providing humans with assistive conversational support, we can augment the conversation that can be carried out. The AdvisorBot can also help to reduce the time taken to type and convey the message if the AdvisorBot is context aware and capable of providing good responses.There has been a significant research for creating conversational chatbots in open-domain conversations that have claimed to have passed the Turing Test and can converse with humans while not seeming like a bot. However, if these chatbots can converse like humans, can they provide actual assistance in human conversations? This research study observes and improves the advanced open-domain conversational chatbots that are put in practice for providing conversational assistance.While performing this thesis research, the chatbots were deployed to provide conversational assistance and a human study was performed to identify and improve the ways to tackle social anxiety by connecting strangers to perform conversations that would be aided by AdvisorBot. Through the questionnaires that the research subjects filled during their participation, and by performing linguistic analysis, the quality of the AdvisorBot can be improved so that humans can achieve better conversational skills and are able to clearly convey their message while conversing. The results were further enhanced by using transfer learning techniques and quickly improve the quality of the AdvisorBot.
Show less

Title: Effect of Pre-Processing Data on Fairness and Fairness Debugging using GOPHER
Creator: Sarkar, Mousam
Date: 2023
Description: At present, Artificial intelligence has been contributing to the decision-making process heavily. Bias in machine learning models has existed...
Show moreAt present, Artificial intelligence has been contributing to the decision-making process heavily. Bias in machine learning models has existed throughout and present studies’ direct usage of eXplainable Artificial Intelligence (XAI) approaches to identify and study bias. To solve the problem of locating bias and then mitigating it has been achieved by Gopher [1]. It generates interpretable top-k explanations for the unfairness of the model and it also identifies subsets of training data that are the root cause of this unfair behavior. We utilize this system to study the effect of pre-processing on bias through provenance. The concept of data lineage through tagging of data points during and after the pre-processing stage is implemented. Our methodology and results provide a useful point of reference for studying the relation of pre-processing data with the unfairness of the machine learning model.
Show less

Title: Data-Driven Methods for Soft Robot Control and Turbulent Flow Models
Creator: Lopez, Esteban Fernando
Date: 2022
Description: The world today has seen an exponential increase in its usage of computers for communication and measurement. Thanks to recent technologies,...
Show moreThe world today has seen an exponential increase in its usage of computers for communication and measurement. Thanks to recent technologies, we are now able to collect more data than ever before. This has dawned a new age of data-driven methods which can describe systems and behaviors with increasing accuracy. Whereas before we relied on the expertise of a few professionals with domain-specific knowledge developed over years of rigorous study, we are now able to rely on collected data to reveal patterns, develop novel ideas, and offer solutions to the world’s engineering problems. No domain is safe. Within the engineering realm, data-driven methods have seen vast usage in the areas of control and system identification. In this thesis we explore two areas of data-driven methods, namely reinforcement learning and data-driven causality. Reinforcement learning is a method by which an agent learns to increase its selection of ideal actions and behaviors which result in an increasing reward. This method was applied to a soft-robotic concept called the JAMoEBA to solve various tasks of interest in the robotics community, specifically tunnel navigation, obstacle field navigation, and object manipulation. A validation study was conducted to show the complications that arise when applying reinforcement learning to such a complex system. Nevertheless, it was shown that reinforcement learning is capable of solving three key tasks (static tunnel navigation, obstacle field navigation, and object manipulation) using specific simulation and learning hyperparameters. Data-driven causality encompasses a range of metrics and methods which attempt to uncover causal relationships between variables in a system. Several information theoretic causal metrics were developed and applied to nine mode turbulent flow data set which represents the Moehlis model. It was shown that careful consideration into the method used was required to identify significant causal relationships. Causal relationships were shown to converge over several hundred realizations of the turbulent model. Furthermore, these results match the expected causal relationships given known information of self-sustaining processes in turbulence, validating the method’s ability to identify causal relationships in turbulence.
Show less

Title: Large Language Model Based Machine Learning Techniques for Fake News Detection
Creator: Chen, Pin-Chien
Date: 2024
Description: With advanced technology, it’s widely recognized that everyone owns one or more personal devices. Consequently, people are evolving into...
Show moreWith advanced technology, it’s widely recognized that everyone owns one or more personal devices. Consequently, people are evolving into content creators on social media or the streaming platforms sharing their personal ideas regardless of their education or expertise level. Distinguishing fake news is becoming increasingly crucial. However, the recent research only presents comparisons of detecting fake news between one or more models across different datasets. In this work, we applied Natural Language Processing (NLP) techniques with Naïve Bayes and DistilBERT machine learning method combing and augmenting four datasets. The results show that the balanced accuracy is higher than the average in the recent studies. This suggests that our approach holds for improving fake news detection in the era of widespread content creation.
Show less

Title: Adaptive Learning Approach of a Domain-Aware CNN-Based Model Observer
Creator: Bogdanovic, Nebojsa
Date: 2023
Description: Application of convolutional neural networks (CNNs) for performing defect detection tasks and their use as model observers (MO) has become...
Show moreApplication of convolutional neural networks (CNNs) for performing defect detection tasks and their use as model observers (MO) has become increasingly popular in the medical imaging field. Building upon this use of CNN MOs, we have trained the CNNs to discern between the data it was trained on, and the previously unseen images. We termed this ability domain awareness. To achieve domain awareness, we are simultaneously training a new variation of U-Net CNN to perform defect detection task, as well as to reconstruct a noisy input image. We have shown that the values of the reconstruction mean squared error can be used as a good indicator of how well the algorithm performs in the defect localization task, making a big step towards developing a domain aware CNN MO. Additionally, we have proposed an adaptive learning approach for training these algorithms, and compared them to the non-adaptive learning approach. The main results that we achieved were for the ideal observers, but we also extended these results to human observer data. We have compared different architectures of CNNs with different numbers and sizes of layers, as well as introduced data augmentation to further improve upon our results. Finally, our results show that the proposed adaptive learning approach with introduced data augmentation drastically improves upon the results of a non-adaptive approach in both human and ideal observer cases.
Show less

Title: Measurement and Control of Beam Energy at the Fermilab 400 MeV Transfer Line
Creator: Mwaniki, Matilda W.
Date: 2023
Description: Linac is the first machine in the Accelerator chain at Fermilab where particles are accelerated from 35 keV to 400 MeV and travel to the...
Show moreLinac is the first machine in the Accelerator chain at Fermilab where particles are accelerated from 35 keV to 400 MeV and travel to the Booster where they are stripped of the extra electrons to become protons. Tuning Linac is performed using diagnostics to ensure stable intensity and energy while minimizing uncontrolled particle loss. I have been revisiting diagnostics in the Linac in order to understand their signals and to ensure their data is reliable. I revisited Beam Loss Monitors (BLMs) for the loss data confidence. For the confidence of energy data there were two approaches. The first approach was time-of-flight measurements using Beam Position Monitors (BPMs) and beam velocity stripline pick-up that provides beam phase data. The second approach used the relation between beam position data from BPMs and dispersion values from MAD-X simulation to calculate energy. Our goal after understanding the data from the Linac diagnostics and finding the data reliable is to control the Linac parameters using Machine Learning techniques to increase the reliability and quality of beam delivered from Linac.
Show less

repository.iit

Search the repository

Enabled Filters

Refine Results

Date

Department

Subject

Creator

Rights