Organizational Unit:

Undergraduate Research Opportunities Program

Permanent Link

https://hdl.handle.net/1853/70802

Parent Organization

Organizational Unit

Office of Undergraduate Education

ArchiveSpace Name Record

https://finding-aids.library.gatech.edu/agents/corporate_entities/1180

Full item page

Publication Search Results

Now showing 1 - 10 of 121

Establishing a Home Sensing Platform in the Field of Technological Healthcare

(Georgia Institute of Technology, 2020-12) Link, Cooper

This thesis explores how home sensor platforms can be leveraged in the context of care for chronic conditions. In order to understand the needs of such a system, a platform has been developed and deployed at the Georgia Tech Aware Home to collect data in a research setting for the Emory and Georgia Tech joint Cognitive Empowerment Program. The goal of this program is to develop a personalized approach to the treatment of Mild Cognitive Impairments. This thesis will explore the design of the supporting home technology platform.
Analyzing and Learning Movement Through Human-Computer Co-Creative Improvisation and Data Visualization

(Georgia Institute of Technology, 2020-12) Liu, Lucas

Recent years have seen an incredible rise in the availability of household motion and video capture technologies, ranging from the humble webcam to the relatively sophisticated Kinect sensor. Naturally, this precipitated a rise in both the quantity and quality of motion capture data available on the internet. The wealth of data on the internet has caused a new interest in the field of motion data classification, the specific task of having a model classify and sort different clips of human motion. However, there is comparatively little work in the field of motion data clustering, which is an unsupervised field that may be more useful in the future as it allows for agents to recognize “categories” of motions without the need for user input or classified data. Systems that can cluster motion data focus more on “what type of motion data is this, and what is it similar to” rather than which motion is this. The LuminAI project, as described in this paper, is an example of a practical use for motion data clustering that allows the system to respond to user dance moves with a similar but different gesture. To analyze the efficacy and properties of this motion data clustering pipeline, we also propose a novel data visualization tool and the design considerations involved in its development.
Search and Rescue Dog Wearable and Mobile Interface

(Georgia Institute of Technology, 2020-12) Liu, Yunqi

Search and Rescue (SAR) dogs are important partners in SAR activities since their born talents in olfactory and auditory senses. Traditionally, the SAR dogs are usually released in the last known spot of the target person, doing a range searching, and returning to their handlers if they find any clues. The current process will cause a great waste on time and potential losing in track of the target while dogs returning to handlers. Our research project aims to replace this traditional working mode by developing an interactive computing system to enable a remote communication between the SAR dogs and the handlers. We provide a vest wearable for dog and a mobile application for handlers, and we allow distant data transmission between the vest and a mobile application. Our approach would prevent the returning step, and hence increase the efficiency and effectiveness of the SAR activities. After experiment and alpha test, we prove that our prototype has all desired functionalities with great precision and weight. Further training and testing with actual SAR dogs would be expected and conducted with K9 teams.
PeopleMap: NLP and Visualization Tool for Mapping Out Researchers

(Georgia Institute of Technology, 2020-12) Saad-Falcon, Jon

Discovering research expertise at universities can be a difficult task. Directories routinely become outdated, and few help in visually summarizing researchers' work or supporting the exploration of shared interests among researchers. This results in lost opportunities for both internal and external entities to discover new connections, nurture research collaboration, and explore the diversity of research. To address this problem, at Georgia Tech, we have been developing PeopleMap, an open-source interactive web-based tool that uses natural language processing (NLP) to create visual maps for researchers based on their research interests and publications. Requiring only the researchers' Google Scholar profiles as input, PeopleMap generates and visualizes embeddings for the researchers, significantly reducing the need for manual curation of publication information. To encourage and facilitate easy adoption and extension of PeopleMap, we have open-sourced it under the permissive MIT license at https://github.com/poloclub/people-map. PeopleMap has received positive feedback and enthusiasm for expanding its adoption across Georgia Tech.
Training Artificial Intelligence

(Georgia Institute of Technology, 2020-12) Gasser, Tarik

Training an Artificial Intelligence could be challenging in so many ways. In our research we are building a strong AI that has the ability to make decision on its own without given explicit answers. We are using Reinforcement learning with Imitating learning to train the Artificial Intelligent offline before putting him through the environment.
Using Language Models in Causal Story Generation

(Georgia Institute of Technology, 2020-12) Li, Siyan

Story generation remains a challenge because it is still difficult to automatically generate logically coherent yet natural stories. In this thesis, we propose an approach to this problem by combining our previous pipeline for story generation and the GPT-2 language model. This new architecture involves filtering generation results from GPT-2, and it outperforms the unfiltered GPT-2 model on tasks such as maintaining a single plotline and having events occurring in a sensible order.
Learning Neural Networks That Can Sort

(Georgia Institute of Technology, 2020-12) Dey, Arnab

This thesis analyzes how neural networks can learn parallel sorting algorithms such as bitonic sorting networks. We discussed how neural networks perform at sorting when given no information or constraints about the allowable operations. We focused on analyzing how the architecture, training data, and length of the array impacted the neural network’s performance at sorting. After encountering challenges with using neural networks to sort, we analyzed how neural networks learn the building blocks for sorting (comparator and swap- ping operators). Once we saw that these basic operations cannot be learned, we framed parallel sorting as a Reinforcement Learning problem. Using Reinforcement Learning, we were able to learn parallel sorting algorithms for sequences of lengths 4 and 8 under certain conditions, specifically limiting the allowable actions. We concluded that using Deep Reinforcement Learning there is potential to learn parallel sorting algorithms without any constraints.
Investigating Sim-to-Real Transfer and Multi-Agent Learning in Assistive Gym

(Georgia Institute of Technology, 2020-12) Schaffer, Holden C.

As the world's population grows older on average and the number of available caregivers decreases, assistive robotics pose an opportunity for older adults or people with disabilities to continue receiving the care that they need. Recent work has shown tremendous progress in using deep reinforcement learning to teach robotic caregivers how to properly assist people in simulation, where robots can learn how to interact with humans in a safe, controlled manner. However, transferring what the robot has learned from simulation to reality continues to pose a challenge for assistive robotics, and a gap in the literature exists in finding techniques to overcome this challenge for this particular domain. The first part of this research uses an assistive simulation framework known as Assistive Gym and its simulated drinking environment to test various approaches to sim-to-real transfer for assistive robotics. The end result of this portion of the research is the identification of a series of baseline steps that are necessary to transfer the Drinking task in Assistive Gym to a physical PR2. Next, the avenues for future works are addressed by investigating a few potential modifications to the drinking task which could be implemented for a more successful transfer of policies. The second part of the research investigates how multi-agent learning could be implemented in Assistive Gym. This section implements multi-agent assistance for the bed-bathing environment, then tests the effectiveness of three different algorithms in order to gauge their effectiveness for solving this new multi-agent task. These algorithms include two variations of single-agent Proximal Policy Optimization modified for multi-agent use as well as Multi-Agent Deep Deterministic Policy Gradient. Finally, future works related to multi-agent assistance are discussed, namely choosing alternate implementations of MADDPG and investigating the dressing environment for its greater potential for cooperation between robots.
Wi-Fi Feature Engineering for Detection of Campus Social Dynamics and Academic Performance Prediction

(Georgia Institute of Technology, 2020-12) Patel, Devashru

Social interaction amongst students can often affect their university experience drastically. The manner in which students collaborate and socialize with their peers can also be used as a helpful metric to predict their academic performance. Traditional methods of quantifying social interaction like surveys suffer from reliability issues stemming from factors such as social desirability bias. In this research, our team investigates a method to predict academic performance that leverages Wi-Fi logs. The logs span a time period of 14 weeks and in their raw form can be used to estimate a student’s location, albeit with low spatial resolution. Analysis of these logs however can lead to fairly accurate inferences of collocations amongst students. We then found that a 0.75 rate of correlation exists between student performance predicted from these collocations and actual performance. These findings are significant in that they could demonstrate the utility of Wi-Fi data in applications such as well-being and mental health.
Trajectory Prediction for Nonuniform Geospatial Mobile Device Data

(Georgia Institute of Technology, 2020-08) Branham, Sara M.

GPS collection has become increasingly popular with the rise of mobile devices and applications. This data is collected for an incredibly wide range of reasons, including applications like Google Maps providing directions, weather applications providing weather predictions, Uber or Lyft providing transportation, or service providers seeking to better understand their users. While GPS data has many uses in our society, there exist enormous obstacles surrounding long term data collection, namely how to acquire uniformly sampled device data. We propose using Transformers and GRUs with added Attention to extract long term habits for individual users in the context of nonuniform GPS data. These models are traditionally used for Neural Machine Translation (NMT), so they are well equipped for nonuniform problem spaces.