Call for PapersThe REU workshop seeks original papers that focus on the implementation of various machine learning and cryptography tasks. This workshop allows undergraduates the opportunity to showcase their work over the past ten weeks, in these various areas. We accept submissions of original research papers discussing these subjects. Topics of Interest include(but are not limited to):
Review Process The research papers will be accepted by Jordyn Blakey and Brandon Haakenson and reviewed by our peers and mentors in the REU summer program. Papers will be accepted based on originality, relevance to their respective fields, and reviewers' comments.
Paper FormatPapers should adhere to the IEEE manuscript standard and be submitted as PDFs.
Melissa Gaines and Gunner Lawless
Kevin Gonzalez and Alex Chen
Technical Program Chairs:
Jordyn Blakey and Brandon Haakenson
Syed Zahidi and Javier Campos
Deep learning models for graphs have shown promising achievements in recent years for the tasks of node and graph classification. Despite these recent advancements, very little research has been done to challenge the robustness of these models against adversarial attacks. In this work, we test the robustness of the GAM graph classification model against various adversarial perturbations. The GAM model is a deep learning model designed by  for graph classification using attention. We begin our experimentation by generating two different types of adversarial data sets: one with randomly generated perturbations and another with perturbations generated using the NETTACK algorithm. The NETTACK algorithm is a poisoning strategy introduced by  to lower accuracy of deep learning models designed for node classification. After generating these data sets, we analyze how these perturbed graphs affect the GAM model’s accuracy in a poison attack setting. We further propose ideas as to why neural networks designed for graph classification behave in the way that they do and call for further investigation.
This paper implements and augments a hierarchical role-based access control (RBAC) model and a dynamic and efficient key management scheme. In such a scheme, different classes of personnel are granted data access based on a hierarchical RBAC model. A corresponding dynamic key management scheme defines key generation and distribution protocols which enforce the data access policies defined by the RBAC model. Through the combined use of the RBAC model and dynamic key management scheme, data access and management can be controlled with fine-tuned precision. This paper proposes a novel improvement upon the key management scheme proposed by  through the implementation of an Access Control Polynomial (ACP)  within the independent groups that make up the key-management scheme, allowing for privilege designation on a user level.
Domain adaptation has attracted great attentions to facilitate the sparsely labeled or unlabeled target learning by leveraging previously well-established source domain through knowledge transfer. Recent activities on domain adaptation attempt to build deep architectures to decrease cross-domain divergences by extracting more effective features. However, its generalizability would decrease significantly due to the domain mismatch that enlarges particularly at the top layers. In this work, we develop a novel Dual Classifiers assisted Domain Adaptation framework (DCDA) to solve the domain mismatch across source and target domains. Specifically, we explore the maximize mean discrepancy (MMD) by incorporating the pseudo labels of target samples to measure the domain difference better. Moreover, dual different types of classifiers are jointly trained to optimize the prediction on the target samples to maximally enhance the prediction ability of both classifiers. .
In an everyday traffic environment, pedestrians and drivers move independently of one another allowing for free movement and turns in various directions, which may make a prediction’s actions unpredictable. For autonomous and driver assistance systems to comprehend pedestrian’s activities, visual data must be collected using a camera attached to a vehicle. The turning and moving of the vehicle creates continuous varying camera angle changes of pedestrians in sight. The changing camera angles make it more difficult to comprehend a pedestrian's movements and as a result more difficult to infer intention of the pedestrian. In this paper, we implement code from an earlier work , to create 2D and 3D human pose estimations from raw RGB image sequences from our dataset. The 2D and 3D pose sequences are used to observe the stride of a pedestrian, which should be consistent but when observed from a constantly changing angle it may be distorted in 2D but not in 3D observations.
Key phrase extraction is an important task in natural language processing. However, traditional methods may not fulfill this task efficiently. Therefore, deep learning techniques are applied to speed up the process. For example, neural machine translation (NMT) models are used to efficiently work with sequential data like text. However, there has been little work exploring the use of sequential models like NMT for key phrase extraction. One part of this paper examines the usefulness of neural machine translation, a sequence-to-sequence model, for key phrase extraction. First, we created an automated sequential labeling system to extract keywords from tweet sequences, and we examine possible methods to sequentially analyze the key phrases of each tweet to create an overall sentiment polarity for each tweet. Another language representation model we used was BERT: Bidirectional Encoder Representations from Transformer. When using BERT for sentiment analysis, we added a layer to capture the attention of words in determining the polarity of a sentence. By examining these attention values, we could find the important phrases in each sentence. In order to analyze these methods, we worked with a dataset of 600 tweets about the HPV vaccine. In this paper, we hope to explore both methods as a means to extract important phrases from the text.