20 skills found
Aastha2104 / Parkinson Disease PredictionIntroduction Parkinson’s Disease is the second most prevalent neurodegenerative disorder after Alzheimer’s, affecting more than 10 million people worldwide. Parkinson’s is characterized primarily by the deterioration of motor and cognitive ability. There is no single test which can be administered for diagnosis. Instead, doctors must perform a careful clinical analysis of the patient’s medical history. Unfortunately, this method of diagnosis is highly inaccurate. A study from the National Institute of Neurological Disorders finds that early diagnosis (having symptoms for 5 years or less) is only 53% accurate. This is not much better than random guessing, but an early diagnosis is critical to effective treatment. Because of these difficulties, I investigate a machine learning approach to accurately diagnose Parkinson’s, using a dataset of various speech features (a non-invasive yet characteristic tool) from the University of Oxford. Why speech features? Speech is very predictive and characteristic of Parkinson’s disease; almost every Parkinson’s patient experiences severe vocal degradation (inability to produce sustained phonations, tremor, hoarseness), so it makes sense to use voice to diagnose the disease. Voice analysis gives the added benefit of being non-invasive, inexpensive, and very easy to extract clinically. Background Parkinson's Disease Parkinson’s is a progressive neurodegenerative condition resulting from the death of the dopamine containing cells of the substantia nigra (which plays an important role in movement). Symptoms include: “frozen” facial features, bradykinesia (slowness of movement), akinesia (impairment of voluntary movement), tremor, and voice impairment. Typically, by the time the disease is diagnosed, 60% of nigrostriatal neurons have degenerated, and 80% of striatal dopamine have been depleted. Performance Metrics TP = true positive, FP = false positive, TN = true negative, FN = false negative Accuracy: (TP+TN)/(P+N) Matthews Correlation Coefficient: 1=perfect, 0=random, -1=completely inaccurate Algorithms Employed Logistic Regression (LR): Uses the sigmoid logistic equation with weights (coefficient values) and biases (constants) to model the probability of a certain class for binary classification. An output of 1 represents one class, and an output of 0 represents the other. Training the model will learn the optimal weights and biases. Linear Discriminant Analysis (LDA): Assumes that the data is Gaussian and each feature has the same variance. LDA estimates the mean and variance for each class from the training data, and then uses properties of statistics (Bayes theorem , Gaussian distribution, etc) to compute the probability of a particular instance belonging to a given class. The class with the largest probability is the prediction. k Nearest Neighbors (KNN): Makes predictions about the validation set using the entire training set. KNN makes a prediction about a new instance by searching through the entire set to find the k “closest” instances. “Closeness” is determined using a proximity measurement (Euclidean) across all features. The class that the majority of the k closest instances belong to is the class that the model predicts the new instance to be. Decision Tree (DT): Represented by a binary tree, where each root node represents an input variable and a split point, and each leaf node contains an output used to make a prediction. Neural Network (NN): Models the way the human brain makes decisions. Each neuron takes in 1+ inputs, and then uses an activation function to process the input with weights and biases to produce an output. Neurons can be arranged into layers, and multiple layers can form a network to model complex decisions. Training the network involves using the training instances to optimize the weights and biases. Naive Bayes (NB): Simplifies the calculation of probabilities by assuming that all features are independent of one another (a strong but effective assumption). Employs Bayes Theorem to calculate the probabilities that the instance to be predicted is in each class, then finds the class with the highest probability. Gradient Boost (GB): Generally used when seeking a model with very high predictive performance. Used to reduce bias and variance (“error”) by combining multiple “weak learners” (not very good models) to create a “strong learner” (high performance model). Involves 3 elements: a loss function (error function) to be optimized, a weak learner (decision tree) to make predictions, and an additive model to add trees to minimize the loss function. Gradient descent is used to minimize error after adding each tree (one by one). Engineering Goal Produce a machine learning model to diagnose Parkinson’s disease given various features of a patient’s speech with at least 90% accuracy and/or a Matthews Correlation Coefficient of at least 0.9. Compare various algorithms and parameters to determine the best model for predicting Parkinson’s. Dataset Description Source: the University of Oxford 195 instances (147 subjects with Parkinson’s, 48 without Parkinson’s) 22 features (elements that are possibly characteristic of Parkinson’s, such as frequency, pitch, amplitude / period of the sound wave) 1 label (1 for Parkinson’s, 0 for no Parkinson’s) Project Pipeline pipeline Summary of Procedure Split the Oxford Parkinson’s Dataset into two parts: one for training, one for validation (evaluate how well the model performs) Train each of the following algorithms with the training set: Logistic Regression, Linear Discriminant Analysis, k Nearest Neighbors, Decision Tree, Neural Network, Naive Bayes, Gradient Boost Evaluate results using the validation set Repeat for the following training set to validation set splits: 80% training / 20% validation, 75% / 25%, and 70% / 30% Repeat for a rescaled version of the dataset (scale all the numbers in the dataset to a range from 0 to 1: this helps to reduce the effect of outliers) Conduct 5 trials and average the results Data a_o a_r m_o m_r Data Analysis In general, the models tended to perform the best (both in terms of accuracy and Matthews Correlation Coefficient) on the rescaled dataset with a 75-25 train-test split. The two highest performing algorithms, k Nearest Neighbors and the Neural Network, both achieved an accuracy of 98%. The NN achieved a MCC of 0.96, while KNN achieved a MCC of 0.94. These figures outperform most existing literature and significantly outperform current methods of diagnosis. Conclusion and Significance These robust results suggest that a machine learning approach can indeed be implemented to significantly improve diagnosis methods of Parkinson’s disease. Given the necessity of early diagnosis for effective treatment, my machine learning models provide a very promising alternative to the current, rather ineffective method of diagnosis. Current methods of early diagnosis are only 53% accurate, while my machine learning model produces 98% accuracy. This 45% increase is critical because an accurate, early diagnosis is needed to effectively treat the disease. Typically, by the time the disease is diagnosed, 60% of nigrostriatal neurons have degenerated, and 80% of striatal dopamine have been depleted. With an earlier diagnosis, much of this degradation could have been slowed or treated. My results are very significant because Parkinson’s affects over 10 million people worldwide who could benefit greatly from an early, accurate diagnosis. Not only is my machine learning approach more accurate in terms of diagnostic accuracy, it is also more scalable, less expensive, and therefore more accessible to people who might not have access to established medical facilities and professionals. The diagnosis is also much simpler, requiring only a 10-15 second voice recording and producing an immediate diagnosis. Future Research Given more time and resources, I would investigate the following: Create a mobile application which would allow the user to record his/her voice, extract the necessary vocal features, and feed it into my machine learning model to diagnose Parkinson’s. Use larger datasets in conjunction with the University of Oxford dataset. Tune and improve my models even further to achieve even better results. Investigate different structures and types of neural networks. Construct a novel algorithm specifically suited for the prediction of Parkinson’s. Generalize my findings and algorithms for all types of dementia disorders, such as Alzheimer’s. References Bind, Shubham. "A Survey of Machine Learning Based Approaches for Parkinson Disease Prediction." International Journal of Computer Science and Information Technologies 6 (2015): n. pag. International Journal of Computer Science and Information Technologies. 2015. Web. 8 Mar. 2017. Brooks, Megan. "Diagnosing Parkinson's Disease Still Challenging." Medscape Medical News. National Institute of Neurological Disorders, 31 July 2014. Web. 20 Mar. 2017. Exploiting Nonlinear Recurrence and Fractal Scaling Properties for Voice Disorder Detection', Little MA, McSharry PE, Roberts SJ, Costello DAE, Moroz IM. BioMedical Engineering OnLine 2007, 6:23 (26 June 2007) Hashmi, Sumaiya F. "A Machine Learning Approach to Diagnosis of Parkinson’s Disease."Claremont Colleges Scholarship. Claremont College, 2013. Web. 10 Mar. 2017. Karplus, Abraham. "Machine Learning Algorithms for Cancer Diagnosis." Machine Learning Algorithms for Cancer Diagnosis (n.d.): n. pag. Mar. 2012. Web. 20 Mar. 2017. Little, Max. "Parkinsons Data Set." UCI Machine Learning Repository. University of Oxford, 26 June 2008. Web. 20 Feb. 2017. Ozcift, Akin, and Arif Gulten. "Classifier Ensemble Construction with Rotation Forest to Improve Medical Diagnosis Performance of Machine Learning Algorithms." Computer Methods and Programs in Biomedicine 104.3 (2011): 443-51. Semantic Scholar. 2011. Web. 15 Mar. 2017. "Parkinson’s Disease Dementia." UCI MIND. N.p., 19 Oct. 2015. Web. 17 Feb. 2017. Salvatore, C., A. Cerasa, I. Castiglioni, F. Gallivanone, A. Augimeri, M. Lopez, G. Arabia, M. Morelli, M.c. Gilardi, and A. Quattrone. "Machine Learning on Brain MRI Data for Differential Diagnosis of Parkinson's Disease and Progressive Supranuclear Palsy."Journal of Neuroscience Methods 222 (2014): 230-37. 2014. Web. 18 Mar. 2017. Shahbakhi, Mohammad, Danial Taheri Far, and Ehsan Tahami. "Speech Analysis for Diagnosis of Parkinson’s Disease Using Genetic Algorithm and Support Vector Machine."Journal of Biomedical Science and Engineering 07.04 (2014): 147-56. Scientific Research. July 2014. Web. 2 Mar. 2017. "Speech and Communication." Speech and Communication. Parkinson's Disease Foundation, n.d. Web. 22 Mar. 2017. Sriram, Tarigoppula V. S., M. Venkateswara Rao, G. V. Satya Narayana, and D. S. V. G. K. Kaladhar. "Diagnosis of Parkinson Disease Using Machine Learning and Data Mining Systems from Voice Dataset." SpringerLink. Springer, Cham, 01 Jan. 1970. Web. 17 Mar. 2017.
anupamkliv / FedERAFedERA is a modular and fully customizable open-source FL framework, aiming to address these issues by offering comprehensive support for heterogeneous edge devices and incorporating both standalone and distributed computing. It includes new software modules to enhance usability and promote environ- mental sustainability.
HewlettPackard / Dc RlSustainDC is a set of Python environments for Data Center simulation and control using Heterogeneous Multi Agent Reinforcement Learning. Includes customizable environments for workload scheduling, cooling optimization, and battery management, with integration into Gymnasium.
HongyuHe / Energat🔋🎯 Thread-level, NUMA-aware energy attribution model for multi-tenancy
Dogiye12 / AI Based SDG Tracker For Environmental GoalsThis project uses synthetic environmental data and AI to track progress toward Sustainable Development Goals (SDGs). It computes a composite score from multiple indicators, enabling data-driven insights for monitoring environmental sustainability and supporting informed decision-making.
intel / Cloud Native AI PipelineAI cloud native pipeline for confidential and sustainable computing
GaiaKnowledge / VerticalFarmingVertical Farming Wizard is an open-source project that centralises vertical farming (VF) and controlled environment agriculture (CEA) solutions. The objective of this decision support system is to compute economic viability and environmental sustainability in the absence of data. It uses imprecise probability techniques to achieve this.
EmileRolley / Sustainable ComputingAn attempt to centralized some resources about sustainable programming, green computing and frugal engineering.
caas-team / Caas Carbon FootprintSupport Sustainable Computing to provide customer with metrics for their carbon footprint workload
payMeQuiz / PayMe ProjectpayMe is a technical solution initiated by some concerned Nigerians aimed to catalyze the innate desire in humans to fairly compete in an intellectual learning exercise. payMe is structured to more than engage users to learn but to incentivize users to be compelled to strive to achieve desired results on set targets. That in fact is the excelling value of payMe over competitions. payMe is originally a web2 play-to-earn (P2E) gaming application upgraded to a hybrid platform by the adoption of the platform's native utility token for incentivizing success among the quizzers. Play-to-earn (P2E) games are online games that guarantee rewards with real-world value to players for completing given task in a contest with other players. It comes with different structure and rewarding system. In the blockchain ecosystem, these rewards can be in the form of in-game assets like crypto tokens, virtual land, as well as the game assets (weapons, tools etc.) and other NFTs. The advent of web3 and its decentralized nature made it possible for players to buy, transfer and sell these in-game assets, outside of the games's traditional platform in exchange for real money. payMe is designed to encourage knowledge development using incentivization of success. payMe serve as cognitive behavioral therapy (CBT) to users who are psychogically affected by the disturbing examination malpractice permitted in the system, with the learn to earn functionality that catalyzes users desire to aspire to be the best in a transparent and meritorious form of testing knowledgeability. The transparent, honest, and undisputed fairness in determination of examinations success, aim also to promote hardwork as valuable asset to success as against luck dependency, in a society that eligizes game-of-chance about competency test. The Service payMe is designed for every interested adult netizen. The service is web-based and a communal crowdfunding scheme. The game is a test of knowledge, interactive quizzing in a multi-choice questions type. payMe™ is designed to play on web browsers enabling desktop and mobile applications. payMe is intentionally created with meritorious rewarding functionality to differ from the original pattern of the gaming industry, which is famously circumscribed by randomly selection of winners in a game of chance that is luck dependent. It is technically structured as an anti-gambling game with superlative uniqueness that distinguished it among competing brands. payMe is an ongoing concern product that will continue to meet users demands that aligns with our believes, principles and goals. The Service's Aim and Objectives The service's core aim is to create economic opportunities using ethical functionalities in a democratized software. Other objectives are: to incentivise intellectual competence. to encourage healthy and fair competition in the field of learning. to promote edifying research habits among scholars using the platform. to provide alternative healthy empowerment platform for gamesters suffering from addiction The Economic Benefit There are so many economic benefits to be derived from the payMe™ product brand. payMe guarantees regular, sustainable fiscal empowerment to users. It is a healthy alternative means of rewarding users’ passion in games. payMe is designed with the ability to enhance valuable learning exercises. Engaging in the contest can help in the reduction of common crimes incidental to youths. The Playing System The console adopts the multiple-choice questions type to create an interactive quizzing format. It offers online learning capabilities that cover extensive information on various academic subjects and soccer (FIFA competitions, leagues, and clubs’ activities). The service features an interactive learning interface and an intuitive time-bound quiz contest amongst participants. The service is deliberately created to differ from the original pattern of the gaming industry, which is luck-dependent, to an intellectual development contest. By this, payMe™ is ethical. Though it entails the use of cash to gain access, however, it guarantees much value for the little Token expended on every entry. payMe™ is designed to play on web browsers enabling desktop and mobile applications. It is playable everywhere with internet access on PC, Laptop, mobile phones, and other devices, if supported. Its technicality and structure make it superlatively unique among competing brands. payMe™ is an ongoing-concerned revolutionary software. The Process payMe™ can be subscribed to online. It entails an initial free membership registration and thereafter funding of a personal wallet with the platform’s native utility asset – the payME Pay Token (payME) Contestants automatically qualify to either make use of the premium entry to the quiz contest or use the payment option. Any of the quiz contests is a set of Ten (10) objective intra-changeable questions. Contestants are expected to provide correct answers to the questions within the swiftest timeframe. Each contestant’s result is displayed after the last question is answered on the contestants’ quiz page and thereafter, updated on the general result page. Weekly participation is limitless for contestants entering as regular quizzers but limited to 10 entries for the premium contestants. Only the best result amongst a Quizzer’s several attempts is registered for the contestant despite when it’s played each week. The first ranked 5% of the weekly contestants based on the most correct answers provided within the swiftest timeframe wins. The Playing Schedule Contests start every Monday at 12 am and end Saturday at 11.59 pm. From 12 am to 11.59 pm every Sunday, results are automatically displayed on the Result Page. The Web Portal Interface and Functions The web/mobile app has interactive interfaces and modules that help Quizzers easily glide through their activities. Some of these modules are described below: a. Wallet: A participant is expected to link his personal blockchain wallet after registration and fund it to enable him gain access to games. Only a decentralized crypto wallet is accepted. b. Quiz: This is 10 revolving questions, each having 4 objectives with 1 possible answer. When entering as a regular quizzer, once the play (ACE or COS) quiz button is clicked, $0.50 worth of payME Token will be debited from the quizzer’s wallet and credited to the platform’s wallet. Next, intra-changeable questions from the Question Bank will be appearing in non sequence routine. Quizzers are expected to provide the correct answers to each of the 10 questions as fast as they could and within 18 seconds. Immediately, after answering the 10th question, the quizzer’s result will display automatically on both the quizzing page and the general result dashboard. c. Tutorial Quiz: As the name implies, this is a free gaming zone created to enable holders of payMe token who may not want to engage in the contest, but want to improve their intellect through quizzing on the platform. However, the user must have payMe token worth $50 USD in his/her personal connected wallet to gain free access at anytime. d. Result Dashboard: this is a general result center for all Contestants. It updates automatically after each game is played during the play period, and according to the most correct answers within the swiftest time frame. That means that if a million quizzers scored 100/100, the system will display their result according to the fastest to answer the entire question using nanoseconds (an SI unit of time equal to one billionth of a second) in computing. So, be rest assured that it is impractical and impossible for 2 Quizzers to tally in the result. This makes payMe unique in the way winners are determined – fastest finger first! Updating of the dashboard however is programmed to freeze once it is 12 am every Sunday to determine the winners by publishing the result of the past week until 12 am on Monday when it continues its routine update. Available Rewards PayMe weekly quiz contest will commence by 12:00AM on Mondays and close by 11:59PM on Saturdays. Results are auto displayed on Sundays. the top most 5% of the participants are declared winners weekly and they are incentivized in ranges of: 1. The topmost 20% of the Winners earn 40% of the total revenue allocation to incentivizing pool in an equal share. 2. While the remaining 80% of the Winners earn 60% of the revenue share on an equal distribution rate. Practically, whatever revenue is generated weekly, 50% is automatically remitted to the incentivizing pool wallet and from there, the topmost 5% of the total participants are rewarded in the following ratio: 40% is equally shared to the topmost 20% while 60% is equally shared to the rest 80% of the winners. Winners claim button would automatically turn green by 12am every Sunday and they can claim their prizes themselves from their dashboard. Every Year, each of the 10 most intelligent Quizzers, drawn from the 52 weeks’ cumulative results of all participants in the ACE Quiz contests are rewarded with a Scholarship Award worth $1000 USDT in payME Token. Criteria is participating in every week of play. Cost of Play Cost of Game $0.50 worth of payME Token per game entry or with a single weekly subscription with payMe token worth $5. NB: The rewards are available for each quizzer's claim from Sunday at 12 am (according to the Nigerian calendar and time).
exalsius / Curtail LlmDistributed LLM pretraining during renewable curtailment windows 🌱
ImperialCollegeLondon / CarbonEstimates carbon emissions of compute jobs run on high-performance computing clusters.
jojacola / SCARF(HotCarbon '25) Code for SCARF: Not All Water Consumption Is Equal: A Water Stress Weighted Metric for Sustainable Computing
teto1992 / GreenICTerA logic programming solution to assess the ecological footprint of your ICT system
yahia997 / PavementEyeA sustainable, cost-effective solution for automating pavement assessment in Alexandria, Egypt. Using computer vision (YOLO v8), cloud computing, and OpenStreetMap integration, it provides a real-time dashboard to visualize crack severity and intelligently prioritize road repairs.
ozcanmiraay / Vidur EnergyEnergy and carbon-aware extension of Vidur for sustainable LLM inference simulation and analysis 💡🌳
DLR-VF / PtACA Python library to automatically compute walking access to public transport for the Sustainable Development Goal 11.2.1
LukeJYK / SC24 Eco LifeECOLIFE: Carbon-Aware Serverless Function Scheduling for Sustainable Computing
minhnhattrinh312 / Multiphase Active Contour Loss“A Semisupervised Deep Learning-based Approach with Multiphase Active Contour Loss for Left Ventricle Segmentation from CMR Images,” The third International Conference on Sustainable Computing (SUSCOM-2021), 2021.
djleamen / SUSTAINSUSTAIN is an environmentally-friendly, token-optimized AI wrapper designed to reduce compute costs and increase productivity.
