Numerical Analysis and Scientific Computing Commons^™

A Bert-Based Two-Stage Model For Chinese Chengyu Recommendation, Minghuan Tan, Jing Jiang, Bingtian Dai

Research Collection School Of Computing and Information Systems

In Chinese, Chengyu are fixed phrases consisting of four characters. As a type of idioms, their meanings usually cannot be derived from their component characters. In this paper, we study the task of recommending a Chengyu given a textual context. Observing some of the limitations with existing work, we propose a two-stage model, where during the first stage we re-train a Chinese BERT model by masking out Chengyu from a large Chinese corpus with a wide coverage of Chengyu. During the second stage, we fine-tune the retrained, Chengyu-oriented BERT on a specific Chengyu recommendation dataset. We evaluate this method on ...

High-Order Flexible Multirate Integrators For Multiphysics Applications, Rujeko Chinomona

Mathematics Theses and Dissertations

Traditionally, time integration methods within multiphysics simulations have been chosen to cater to the most restrictive dynamics, sometimes at a great computational cost. Multirate integrators accurately and efficiently solve systems of ordinary differential equations that exhibit different time scales using two or more time steps. In this thesis, we explore three classes of time integrators that can be classified as one-step multi-stage multirate methods for which the slow dynamics are evolved using a traditional one step scheme and the fast dynamics are solved through a sequence of modified initial value problems. Practically, the fast dynamics are subcycled using a small ...

A Comparison Of Word Embedding Techniques For Similarity Analysis, Tyler Gerth

Computer Science and Computer Engineering Undergraduate Honors Theses

There have been a multitude of word embedding techniques developed that allow a computer to process natural language and compare the relationships between different words programmatically. In this paper, similarity analysis, or the testing of words for synonymic relations, is used to compare several of these techniques to see which performs the best. The techniques being compared all utilize the method of creating word vectors, reducing words down into a single vector of numerical values that denote how the word relates to other words that appear around it. In order to get a holistic comparison, multiple analyses were made, with ...

Using Deep Learning For Children Brain Image Analysis, Rafael Toche Pizano

Computer Science and Computer Engineering Undergraduate Honors Theses

Analyzing the correlation between brain volumetric/morphometry features and cognition/behavior in children is important in the field of pediatrics as identifying such relationships can help identify children who may be at risk for illnesses. Understanding these relationships can not only help identify children who may be at risk of illnesses, but it can also help evaluate strategies that promote brain development in children. Currently, one way to do this is to use traditional statistical methods such as a correlation analysis, but such an approach does not make it easy to generalize and predict how brain volumetric/morphometry will impact ...

Using Large Pre-Trained Language Models To Track Emotions Of Cancer Patients On Twitter, Will Baker

Computer Science and Computer Engineering Undergraduate Honors Theses

Twitter is a microblogging website where any user can publicly release a message, called a tweet, expressing their feelings about current events or their own lives. This candid, unfiltered feedback is valuable in the spaces of healthcare and public health communications, where it may be difficult for cancer patients to divulge personal information to healthcare teams, and randomly selected patients may decline participation in surveys about their experiences. In this thesis, BERTweet, a state-of-the-art natural language processing (NLP) model, was used to predict sentiment and emotion labels for cancer-related tweets collected in 2019 and 2020. In longitudinal plots, trends in ...

Trunctrimmer: A First Step Towards Automating Standard Bioinformatic Analysis, Z. Gunner Lawless, Dana Dittoe, Dale R. Thompson, Steven C. Ricke

Computer Science and Computer Engineering Undergraduate Honors Theses

Bioinformatic analysis is a time-consuming process for labs performing research on various microbiomes. Researchers use tools like Qiime2 to help standardize the bioinformatic analysis methods, but even large, extensible platforms like Qiime2 have drawbacks due to the attention required by researchers. In this project, we propose to automate additional standard lab bioinformatic procedures by eliminating the existing manual process of determining the trim and truncate locations for paired end 2 sequences. We introduce a new Qiime2 plugin called TruncTrimmer to automate the process that usually requires the researcher to make a decision on where to trim and truncate manually after ...

Improving Bayesian Graph Convolutional Networks Using Markov Chain Monte Carlo Graph Sampling, Aneesh Komanduri

Computer Science and Computer Engineering Undergraduate Honors Theses

In the modern age of social media and networks, graph representations of real-world phenomena have become incredibly crucial. Often, we are interested in understanding how entities in a graph are interconnected. Graph Neural Networks (GNNs) have proven to be a very useful tool in a variety of graph learning tasks including node classification, link prediction, and edge classification. However, in most of these tasks, the graph data we are working with may be noisy and may contain spurious edges. That is, there is a lot of uncertainty associated with the underlying graph structure. Recent approaches to modeling uncertainty have been ...

Tripdecoder: Study Travel Time Attributes And Route Preferences Of Metro Systems From Smart Card Data, Xiancai Tian, Baihua Zheng, Yazhe Wang, Hsao-Ting Huang, Chih-Cheng Hung

Research Collection School Of Computing and Information Systems

In this paper, we target at recovering the exact routes taken by commuters inside a metro system that are not captured by an Automated Fare Collection (AFC) system and hence remain unknown. We strategically propose two inference tasks to handle the recovering, one to infer the travel time of each travel link that contributes to the total duration of any trip inside a metro network and the other to infer the route preferences based on historical trip records and the travel time of each travel link inferred in the previous inference task. As these two inference tasks have interrelationship, most ...

Enconter: Entity Constrained Progressive Sequence Generation Via Insertion-Based Transformer, Lee Hsun Hsieh, Yang Yin Lee, Ee-Peng Lim

Research Collection School Of Computing and Information Systems

Pretrained using large amount of data, autoregressive language models are able to generate high quality sequences. However, these models do not perform well under hard lexical constraints as they lack fine control of content generation process. Progressive insertion-based transformers can overcome the above limitation and efficiently generate a sequence in parallel given some input tokens as constraint. These transformers however may fail to support hard lexical constraints as their generation process is more likely to terminate prematurely. The paper analyses such early termination problems and proposes the ENtity-CONstrained insertion TransformER (ENCONTER), a new insertion transformer that addresses the above pitfall ...

Agent-Based Computational Economics: Overview And Brief History, Leigh Tesfatsion

Economics Working Papers

Scientists seek to understand how real-world systems work. Models devised for scientific purposes must always simplify reality. However, scientists should be permitted to tailor these simplifications to purposes at hand; they should not be forced to distort reality in specific predetermined ways in order to apply a modeling approach. Adherence to this modeling precept was a key goal motivating my development of Agent-Based Computational Economics (ACE), a variant of agent-based modeling characterized by seven specific modeling principles. This perspective provides an overview of ACE and a brief history of its development.

Parallel Arbitrary-Precision Integer Arithmetic, Davood Mohajerani

Electronic Thesis and Dissertation Repository

Arbitrary-precision integer arithmetic computations are driven by applications in solving systems of polynomial equations and public-key cryptography. Such computations arise when high precision is required (with large input values that fit into multiple machine words), or to avoid coefficient overflow due to intermediate expression swell. Meanwhile, the growing demand for faster computation alongside the recent advances in the hardware technology have led to the development of a vast array of many-core and multi-core processors, accelerators, programming models, and language extensions (e.g. CUDA, OpenCL, and OpenACC for GPUs, and OpenMP and Cilk for multi-core CPUs). The massive computational power of ...

Quantum Simulation Of Schrödinger's Equation, Mohamed Eltohfa

Capstone and Graduation Projects

Quantum computing is one of the promising active areas in physics research. This is because of the potential of quantum algorithms to outperform their classical counterparts. Grover’s search algorithm has a quadratic speed-up compared to the classical linear search. The quantum simulation of Schrödinger’s equation has an exponential memory save-up compared to the classical simulation. In this thesis, the ideas and tools of quantum computing are reviewed. Grover’s algorithm is studied and simulated as an example. Using the Qiskit quantum computing library, a code to simulate Schrödinger’s equation for a particle in one dimension is developed ...

Structurally Enriched Entity Mention Embedding From Semi-Structured Textual Content, Lee Hsun Hsieh, Yang Yin Lee, Ee-Peng Lim

Research Collection School Of Computing and Information Systems

In this research, we propose a novel and effective entity mention embedding framework that learns from semi-structured text corpus with annotated entity mentions without the aid of well-constructed knowledge graph or external semantic information other than the corpus itself. Based on the co-occurrence of words and entity mentions, we enrich the co-occurrence matrix with entity-entity, entity-word, and word-entity relationships as well as the simple structures within the documents. Experimentally, we show that our proposed entity mention embedding benefits from the structural information in link prediction task measured by mean reciprocal rank (MRR) and mean precision@K (MP@K) on two ...

All The Wiser: Fake News Intervention Using User Reading Preferences, Kuan Chieh Lo, Shih Chieh Dai, Aiping Xiong, Jing Jiang, Lun Wei Ku

Research Collection School Of Computing and Information Systems

To address the increasingly significant issue of fake news, we develop a news reading platform in which we propose an implicit approach to reduce people's belief in fake news. Specifically, we leverage reinforcement learning to learn an intervention module on top of a recommender system (RS) such that the module is activated to replace RS to recommend news toward the verification once users touch the fake news. To examine the effect of the proposed method, we conduct a comprehensive evaluation with 89 human subjects and check the effective rate of change in belief but without their other limitations. Moreover ...

A High-Precision Machine Learning Algorithm To Classify Left And Right Outflow Tract Ventricular Tachycardia, Jianwei Zhang, Guohua Fu, Islam Abudayyeh, Magdi Yacoub, Anthony Chang, William Feaster, Louis Ehwerhemuepha, Hesham El-Askary, Xianfeng Du, Bin He, Mingjun Feng, Yibo Yu, Binhao Wang, Jing Liu, Hai Yao, Hulmin Chu, Cyril Rakovski

Mathematics, Physics, and Computer Science Faculty Articles and Research

Introduction: Multiple algorithms based on 12-lead ECG measurements have been proposed to identify the right ventricular outflow tract (RVOT) and left ventricular outflow tract (LVOT) locations from which ventricular tachycardia (VT) and frequent premature ventricular complex (PVC) originate. However, a clinical-grade machine learning algorithm that automatically analyzes characteristics of 12-lead ECGs and predicts RVOT or LVOT origins of VT and PVC is not currently available. The effective ablation sites of RVOT and LVOT, confirmed by a successful ablation procedure, provide evidence to create RVOT and LVOT labels for the machine learning model.

Methods: We randomly sampled training, validation, and testing ...

The Introduction Of Big Data In Cloud Computing, Austin Gruenberg

Student Academic Conference

One of the fastest-growing technologies that many people are unaware of is the world of cloud computing. Having started in 2006, it is a relatively new technological advancement in the computer industry. The major branch of cloud computing that I decided to focus on was big data. I decided to research this topic to better understand what its current uses are, to see what the future holds for Big Data and cloud computing and because it is a growing, significant piece of technology being used in our society today. Big data and cloud computing are very important industries and have ...

Design Project: Smart Headband, John Michel, Jack Durkin, Noah Lewis

Williams Honors College, Honors Research Projects

Concussion in sports is a prevalent medical issue. It can be difficult for medical professionals to diagnose concussions. With the fast pace nature of many sports, and the damaging effects of concussions, it is important that any concussion risks are assessed immediately. There is a growing trend of wearable technology that collects data such as steps and provides the wearer with in-depth information regarding their performance. The Smart Headband project created a wearable that can record impact data and provide the wearer with a detailed analysis on their risk of sustaining a concussion. The Smart Headband uses accelerometers and gyroscopes ...

Football’S Future: An Analytical Interpretation Of The Premier League, Hunter Witeof

Williams Honors College, Honors Research Projects

This project looks to take the statistics of soccer players and run them through an algorithm to determine how well a player is performing. The system that will be designed in the project will look to accomplish 3 main goals: allow the user to enter new statistics, store the data for all 38 game weeks for all 20 teams, and compute a score for each player’s performance for each game as well as the average of all of the player's scores.

Predictive Maintenance Of Bearing Machinery Using Simulation- A Bibliometric Study, Karan Gulati Mr., Keshav Basandrai Mr., Shubham Tiwari Mr., Pooja Kamat Prof., Satish Kumar Dr.

Library Philosophy and Practice (e-journal)

Modelling is a way of constructing a virtual representation of software and hardware that involves a real-world device. We will discover the behaviour of the system if the software elements of this model are guided by mathematical relationships. For testing conditions that may be difficult to replicate with hardware prototypes alone, modelling and simulation are particularly useful, especially in the early phase of the design process when hardware might not be available. Model-based approach in MATLAB-Simulink can be useful for predictive maintenance of machines as it can reduce unplanned downtimes and maintenance costs when industrial equipment breaks. Through this bibliometric ...

Multigrid For The Nonlinear Power Flow Equations, Enrique Pereira Batista

Mathematics Theses and Dissertations

The continuously changing structure of power systems and the inclusion of renewable
energy sources are leading to changes in the dynamics of modern power grid,
which have brought renewed attention to the solution of the AC power flow equations.
In particular, development of fast and robust solvers for the power flow problem
continues to be actively investigated. A novel multigrid technique for coarse-graining
dynamic power grid models has been developed recently. This technique uses an
algebraic multigrid (AMG) coarsening strategy applied to the weighted
graph Laplacian that arises from the power network's topology for the construction
of coarse-grain approximations ...