Music Splash is an app designed to help students develop their musical ability and musicality. The app uses machine learning to account for and identify the players own elements of musicality and help to enhance them. The app gets input which is initially stored as a .wav file. This .wav file is then put through a process, parsed by the computer, to convert it into a file format that can be used. Once the file has been converted, it is compared to a base file that has been pre-recorded by a teacher. This analysis, due to the complexity of music, involves using machine learning to compare the two files.Based on the given comparison, the program then chooses, from a pre-set list of outputs, what feedback needs to be given to the user to improve their performance.
Artificial Intelligence System Object and Place Recognizer for Blind People and Development of Data Set
In many areas, such as finance, health and safety, and in the process of facilitating the lives of persons with disabilities, artificial intelligence systems operate. Applications developed using artificial intelligence technologies are able to identify visually impaired objects, while location recognition practices are inadequate. With a mobile application that works on deep learning as we develop for our project and does not require an Internet connection, visually impaired individuals can identify places near the external environment. Individuals with no visual impairment can also be part of the solution by taking pictures of the surrounding sites, thereby contributing to the growth of the data set.
In my project Blockchain Decrypted I developed a method how the complex fundamentals of Blockchain can be explained in an intelligible way. The functionality and mathematical backgrounds of the future-oriented Blockchain are presented in an understandable way. Furthermore, an exciting application of the Blockchain, called Smart Contract, is described. An own Smart Contract, which implements a TicTacToe game, has been developed and an own crypto currency has been created and integrated. Thus the most exciting aspects of the Blockchain are demonstrated, the possibilities of the technology are clearly recognizable. Thanks to a webpage-interface, the game is easy to use. My work offers a great opportunity to get to know the dimensions of Blockchain on the basis of a specific example.
eSports are a rapidly growing industry, and thus, there is a lot of attention drawn to it lately. One hot button issue is cheating. As with other more traditional sports, doping is also present in eSports, however with the sport moving onto electronic devices comes a new way – cheating using digital programs aimed to give the player an unfair advantage. In my project I have researched the various cheating methods, and created various solutions to detect and prevent players from using such things to their advantage.
For decades, mankind has been dreaming of colonizing the Moon and Mars. This idea raises various scientific, technological and economic challenges. In 2009, lunar probe discovered moon cave, which may be optimal base for a habitable base protected from radiation. First exploration stages may include robotic mission. In this work authors designed algorithms that implement automatic robot rescue if communication was lost. A new method of measuring the efficiency of the rescue algorithm was developed — a complex parameter that includes the time to search for a communication site and the amount of energy consumed. The algorithms were tested and verified using computer simulation and using the physical robot developed in this work.
Quantum computers may tackle problems beyond the capabilities of current computers. However, only small scale quantum devices are currently available. This introduces a need for fast and accurate simulation methods and tools.In this work, a series of tools for simulating quantum computers are developed. Existing techniques are built upon, and new algorithms are developed. A classical preprocessing step is introduced, allowing for optimizations throughout the simulation process. These developments create a coherent approach towards the simulation of quantum circuits, that can be used by any researcher to improve the simulation process for any quantum circuit, allowing more qubits, more quantum gates and faster development times.
The ultraviolet part of the sunlight can damage the human skin by sunburn depending upon the radiation intensity. For people who are heavily exposed to sunlight (beach tourists, sailors, mountain hikers), it would be useful to know whether the current UV radiation they are exposed to can cause sunburn. Ideally, these people would carry a small measuring device with them that gives them continuous information about the intensity of UV radiation. It can be shown that open source hardware and software, 3D printing and a smartphone app are already sufficient to realize such a UVI measuring device in the form of a wristwatch.
PlantPlant is a mobile doctor crop in your pocket. Our project aims to help small farmers in developing countries that had a limited access to experts. The only thing that the user needs is a smartphone. Wherever the problem lies, a smartphone picture is enough and in seconds you will receive a diagnosis and the appropriate treatment tips. Take a picture of your arable crop by using a simple smartphone, Our application analysis it within the blink of an eye and reports detailed information about the plant’s disease.
The oral problems of young children are particularly serious because they are at a critical moment in developing good oral hygiene habits. How to help them maintain oral cleaning through sensors and artificial intelligence technology has become a research point of intelligent Internet of Things. We present an easy accessible monitoring system for evaluating the tooth brushing with smart devices. The system captures the users’ brushing behaviors (e.g., the motion of hands and the acoustic signals during tooth brushing) through the two build-in sensors. Then the collected data is transmitted to the smartphones and evaluated through a designed machine learning-based model. Finally, a DNN model is adopted to significantly improve the accuracy of detecting tooth brush tasks by up to 97.7%.
Creating efficient deep neural networks involves repetitive manual optimization of the topology and the hyperparameters. This human intervention significantly inhibits the process.Neural Architecture Search (NAS) algorithms can effectively automate this work and achieve results that surpass the best human-designed models.This research proposes a novel blockchain network protocol that incentivises independent computing nodes to run NAS algorithms and compete in finding better neural network models for a particular task. If implemented, such network can be an autonomous and self-improving source of machine learning models, significantly lowering the cost and access to accurate Machine Learning solutions.