I am going to talk about my personal experience in life, the opportunities I have had, and create for others to achieve their full potential. As a mother and professional, I believe there should be room for each person and especially women to achieve their goal life and brake barriers, especially if they have the strength and the will.
Modern robots come equipped with a variety of sensors to obtain data about the world. They can transmit this data wirelessly over the internet to other devices. Any device that has these properties is an IoT (Internet of Things) device. Pepper is a social humanoid robot intended to interact with elderly dementia patients. In order to safely interact with patients and move around a space, Pepper must be able to process the data it receives from its cameras, lasers, sonars and microphones to recognize users and avoid obstacles such as doors and tables. My research is working towards developing the algorithms and techniques to make Pepper a more intelligent and responsive companion.
Medication adherence is one of the most pressing issues in the field of healthcare. Especially for elderly people and people affected with Dementia, this a pressing problem that can appear as a serious threat with regards to their well being. With the advancement of robotics and autonomous technologies, numerous approaches have been taken to shift this load of keeping track and maintaining the proper dosage of correct medicines and hand them over to a more reliable source. The automation of this process will not only help the patients themselves in improving their medication adherence but will also help their caregiver's proper insight in aiding and monitoring their conditions and to take necessary and quick actions in any situation- if required.
We are studying random numerical semigroups. A numerical semigroup is a subset of the natural numbers that has a finite complement and is closed under addition. The model we are using to generate random numerical semigroups takes a fixed positive integer N and probability p. We then generate all semigroups with generators between 2 and N, and randomly select semigroups using probability p. We create a single random numerical semigroup by intersecting the selected semigroups. In this research project, we change the value for N and study how it alters two characteristics of numerical semigroups, the Frobenius number and the number of minimal generators. We use the program SageMath to compute and graph these randomly generated semigroups, and to datafit curves to our gathered data. We find that as N increases, both the average Frobenius number and the average number of minimal generators increase quadratically. Furthermore, as p changes, the curves for the minimal number of generators diverge from each other but the Frobenius number curves tend to stick together.
Autism spectrum disorder(ASD) is a disability that affects children’s ability to interact with others and get involved in social interaction. Children with ASD face difficulty expressing or understanding emotion. With the recent advancement in au-autonomous technology, humanoid robots are used to have a better understanding of the behavior and attitude of the children with ASD. Following this line of thought,monitoring and analyzing child-robot interaction has become a promising aspect in healthcare and robotics. In this thesis, the interaction between a humanoid robot: Nao and children with autism are investigated and analyzed with different sensors. Children with ASD often lack social skills. They are often very reluctant to express their emotions. So therapists or healthcare professionals often find it hard to under-stand their emotional aspects by observing them with their eyes. In such a scenario,sensors might play a vital role to understand what’s happening inside their body and brain. Instead of guessing abstract ideas about their emotions, real numbers from the sensors might help the healthcare professionals understand what excites them, what agitates them, what concerns them and so on.
Understanding the reaction to robots and robotic gender in different age categories.
The growing popularity of social media has made the proliferation of fake news easily possible. Hyperpartisan News reports events with extreme bias for a certain party. The proposed models participated in Task 4 of SemEval 2019 where it was ranked 23rd out of the 42 participating teams, the objective of the task being to classify an article as Hyperpartisan or not.
Recent research reveals that patients with dementia (PWD) face numerous behavioral and psychological problems and they struggle to express their emotions as easily as healthy individuals. There is also evidence that playing music can help PWD increase positive emotions and help them feel better. Additionally, after listening to the music, aggressive behaviors signs’ decline in PWD. The main objective of this study is to determine if nursery rhymes impact human emotion and brain signals, and determine whether emotional responses to music varies between genders.
In the field of computer security, one of the most effective tools to teach adversarial thinking to students is capture the flag games, or CTFs. In practice, CTFs are time consuming to create and after one playthrough, the student knows all the tricks and nuances of that specific game. This project, Procedural Content Generation for Capture the Flag games (PCGCTF), is a framework that aims to drastically reduce setup times and enhance the uniqueness of CTF games. At the click of a button, a batch of unique CTF games can be generated for a classroom of students. Each generated CTF is a network of computers, with each computer containing one or more exploitable vulnerabilities. The objective of the game is to exploit each vulnerability in order, which leads the player to the 'flag' at the end of the game. The PCGCTF tool creates the network topology, assigns which computers get which vulnerabilities, then configures each vulnerability uniquely. In the future, developers can add their own vulnerabilities to the library that the game chooses from, drastically increasing the possible number of games.
Cyber security plays a crucial role in both protecting sensitive information and preventing attacks that destruct business operations. The ability to think like a hacker is essential to people who provide cyber protections. This research project aims to develop students’ adversarial thinking or computer security skills through special challenges in the form of playing games. This research project contains two main parts. One task is to introduce potential cheating flaws that players can take advantage of. We are currently working on a game that is similar to Pac-Man and our progress includes several different cheating behaviors such as teleporting the player when entering a certain spot in the game. The other task is running experiments with users in order to see if finding cheating behaviors in a game can help translate to finding flaws in other computer programs. We plan to invite two groups of students to play this game, but only one group of students will be exposed to those potential cheating flaws. We are exceedingly interested in knowing whether doing this kind of exercise can better help students recognize dangerous flaws not only in the form of game but also in real-life situations.
The ability to visualize a two-dimensional (2D) map from a three-dimensional (3D) surface, or a three-dimensional surface from a two-dimensional map, is a learned skill. Principles of topographic maps and contours, and such concepts as watershed delineation, and the overland flow path of surface water runoff are difficult for many students to master. With the vast research growing in the areas of science, technology, engineering, and math (STEM), The College of St. Scholastica is looking for innovative ways to apply augmented reality (AR) as a 2D-3D visualization tool, thus St. Scholastica’s Department of Computer Information Systems has built an AR sandbox as an educational tool in which students can potentially learn these concepts by creating topographic models of watersheds out of sand, upon which a color-coded contour map is projected in real time. This on-going research project is exploring ways to engage students at a tactile level, where students are able to see, touch, and view the reaction of user input simultaneously; the students can simulate rainfall, and observe hydrologic phenomena such as watershed boundaries, surface runoff that flows down the elevation gradient, and the purposes of levees and floodplain storage. Furthermore, this AR sandbox research project will also explore possible applications in the areas of STEM teaching and learning, curriculum development and assessment, special education, and learning games.
Hydration is a balancing act that everyone deals with everyday. Not enough water, and your body will start to malfunction. Conversely, too much water can cause harm to a body. Elderly individuals are even more at risk to the effects of dehydration. Dehydration related incidents in elderly nursing home patients causes a significant strain on resources, which has sparked research into methods of tracking hydration. Even so, there are only a few methods being researched and many do it indirectly, which can be susceptible to error such as spills. This research project explores the option of directly measuring hydration using an EDA sensor and then incorporating that into an Internet of Things system to track hydration in real-time.
Modern prostheses commonly use signals collected from muscle tissue to infer the motion a user wishes to make with the prosthetic. For some amputees, however, the muscle tissues of the remaining limb aren't enough to allow high-accuracy movement classifications. In this case, one possible approach is to use signals obtained from cortical (brain) activity. We examine the use of a hybrid-data collection model, in which minimal myoelectric signals and cortical activity are collected and used in a classification model to predict which specific type of grasp (cylindrical, tip, palmar, spherical, lateral or hook) the user intends.
Care robotics have recently gained much momentum in the healthcare industry. With the hopes of less physical labor for healthcare professionals and a shorter recovery period for patients, we are using Baxter, a research robot with strength capabilities, as a means to help with tasks that can be difficult for human caregivers, such as lifting patients out of bed. In this study, we are focusing on a control mechanism that adjusts to the patients physical metrics with minimal human intervention as well as using an object recognition technique as a way to identify the position of the patient.
Autism spectrum disorder(ASD) is a disability that affects children’s ability to interact with others and get involved in social interaction. Children with ASD facedifficulty expressing or understanding emotion. With the recent advancement in au-tonomous technology, humanoid robots are used to have a better understanding ofthe behavior and attitude of the children with ASD. Following this line of thought,monitoring and analyzing child-robot interaction has become a promising aspect inhealthcare and robotics. In this thesis, the interaction between a humanoid robot:Nao and children with autism are investigated and analyzed with different sensors.Children with ASD often lack social skills. They are often very reluctant to expresstheir emotions. So therapists or healthcare professionals often find it hard to under-stand their emotional aspects by observing them with their eyes. In such a scenario,sensors might play a vital role to understand what’s happening inside their body andbrain. Instead of guessing abstract ideas about their emotions, real numbers from thesensors might help the healthcare professionals understand what excites them, whatagitates them, what concerns them and so on.
Voting algorithm is a simple evolutionary algorithm in which population is produced by tournament selection and uses majority voting mechanism over the population to find the optimal solution.This algorithm effectively solves OneMax ,Jump and any monotonic functions with the best running time so far. In our research we extended the analysis of how the Voting algorithm performs in other nonlinear and non-monotonic functions and also the presence of spin-flip symmetry in the algorithm . We found that majority vote technique does not solve the spin-flip symmetry functions with reasonable probability.It fails to optimize functions because the fitness-distance is mirrored.To mitigate this, we introduced a symmetry breaking technique which provides a bias towards global optimum or its complement. We experimented the symmetry breaking voting algorithm on generalized TwoMax and families of constructed 3-NAE-SAT and 2-XOR-SAT formulas and proved that this small modification results in O(n2 logn) performance.We also proved that this technique fails on one dimensional Ising model.
Genetic Algorithms are inspired by the idea of evolution, it's no wonder that they take time to converge or find the solution. Evolution itself is a very slow process and takes millions of years to evolve species. There is plenty of room for optimization for computation time, especially in terms of parallelization. Recently, in July 2019, a new algorithm has been published that has a lot of potential for parallelization [1]. The algorithm is named “Mixing Genetic Algorithm”. The algorithm has shown promises for "travelling salesman problem". Now we are testing it for "Ising model for trees" and "OneMax" problems to further evaluate the algorithm. If it works for Ising model then it will be helpful to create another option to study Ising model for trees. For the parallel computation graphical processing units (GPUs) are a very good option, as lately GPUs have been developed for general purpose programming as well. In this thesis we attempt to analyze the benefits of parallelization using GPUs for Mixing Genetic algorithm.
Recent research reveals that patients with dementia (PWD) face numerous behavioral and psychological problems and they struggle to express their emotions as easily as healthy individuals. There is also evidence that playing music can help PWD increase positive emotions and help them feel better. Additionally, after listening to the music, aggressive behaviors signs’ decline in PWD. The main objective of this study is to determine if nursery rhymes impact human emotion and brain signals, and determine whether emotional responses to music varies between genders.