269 skills found · Page 9 of 9
BIT-DA / LSG[NeurIPS 2023] Official Implementation of Language Semantic Graph Guided Data-Efficient Learning
MIvanovska / Y GANOfficial implementation of the paper "Y-GAN: Learning Dual Data Representations for Efficient Anomaly Detection"
wwyalice / FAVAL ARANetMachine Learning to Promote Efficient Screening of Low-Contact Electrode for Two-Dimensional Semiconductor Transistor Under Limited Data
shbz80 / Model LearningCode for the paper Data-efficient model learning and prediction for contact-rich manipulation tasks, RA-L, 2020
Fadhaa / AutoMixRegRAutoMixRegR is a machine learning framework for automated mixed-effects regression, enabling efficient modeling of complex hierarchical and grouped data in R.
RiturajSaha / Diabetes Predictor ApplicationAn efficient disease detection application with web based fronted and machine learning backend which detects if a patient is diabetic or normal from essential patient data in real time.
nahin91 / SoilMoisturePrediction MLThe purpose of this project is to predict soil moisture and finding an efficient Machine Learning Model that fits well with the subjected dataset. The predictions are made on 4 different fields by sampling data from the dataset.
yash98gupta / Safest RouteThe project aims to predict the crime type in a locality based on latitude, longitude, time of day(morning, afternoon, evening, night), and other relevant features using the Los Angeles crime dataset. Finally, a safest path algorithm is designed to find the safest path between a source and destination similar to shortest fastest path algorithm in Google maps. Machine learning models such as K-nearest neighbors (K-NN), Logistic Regression, convolutional neural networks (CNN), and Random Forests Classification are capable to predict crimes efficiently. The following strategy entails using effective tools and technologies to predict crimes, classify patterns, and visualize data. The use of historical crime data trends allows us to correlate factors that may aid in understanding the scope of future crimes.
stvsd1314 / PPGN Physics Preserved Graph NetworksThe increasing number of variable renewable energy (solar and wind power) causes power grids to have more abnormal conditions or faults. Faults may further trigger power blackouts or wildfires without timely monitoring and control strategy. Machine learning is a promising technology to accelerate the automation and intelligence of power grid monitoring systems. Unfortunately, the black-box machine learning methods are weak to the realistic challenges in power grids: low observation, insufficient labels, and stochastic environments. To overcome the vulnerability of black-box machine learning, we preserve the physics of power grids through graph networks to efficiently and accurately locate the faults even with limited observability and low label rates. We first calculate the graph embedding of power grid infrastructure by establishing a reduced graph network with the observed nodes, then efficiently locate the fault on the node level using the low-dimensional graph embedding. To augment the location accuracy at low label rates, we build another graph network representing the physical similarity of labeled and unlabeled data samples. Importantly, we provide the physical interpretations of the benefits of the graph design through a random walk equivalence. We conduct comprehensive numerical experiments in the IEEE 123-node. Our proposed method shows superior performance than three baseline classifiers for different fault types, label rates, and robustness to out-of-distribution (OOD) data. Additionally, we extend the proposed method to the IEEE 37-node benchmark system and validate the effectiveness of the proposed training strategy.
pandeyankit83 / Deep Learning Recommender SystemYou can train a neural network with user ratings or purchases, and use it to make recommendations; deep learning can be very good at recognizing patterns in a way similar to how our brain may do it. It's good at things like image recognition and predicting sequences of events.Neural networks are fundamentally matrix operations and there are already well-established matrix factorization techniques for recommender systems that fundamentally do something similar. In SVD for example, we find matrices that we multiply together using weights that are learned from stochastic gradient descent, it's almost the same thing, just thought of in a different way. So yeah, you could think of recommender systems as looking for patterns, just very complex ones based on the behavior of other people. So a matrix factorization can be modeled as a neural network. I think the main reason to experiment with applying neural networks to recommender systems is that it lets us take advantage of all the rapid advances in the fields of AI and deep learning. Amazon, for example, has open-sourced a system called DSSTNE, that's D-S-S-T-N-E, which allows you to run huge neural networks that deal with sparse data, on a cluster, efficiently. They claim to be using this internally for their own recommender systems. There are also ways to use TensorFlow in a cluster, and take advantage of a whole fleet of GPUs. And there's always research on new topologies for neural networks that can lead to fresh insights on how to make better recommendations using them. In some cases, approaches using neural networks have been shown to outperform SVD already, even if it's by a rather small margin. So, let's dive into some ways you can apply neural networks to the problem of making recommendations.
solgaardlab / PhotonicbackpropData and code for the paper "Experimentally realized \textit{in situ} backpropagation for deep learning in energy-efficient nanophotonic neural networks"
Rahm-no / MinatoLoaderArtifact for EuroSys'26 paper "MinatoLoader: Accelerating Machine Learning Training Through Efficient Data Preprocessing"
smartly-creators-program / AI MLThis project builds an intelligent AI-ML system using Neural Networks and Deep Learning to learn from complex data. It uses NLP for understanding text and Computer Vision for image analysis. MLOps practices are applied for efficient training, deployment, monitoring, scalability, and to ensure a reliable, production-ready solution for realworld use.
shaunak27 / Hepco FedHePCo introduces a parameter-efficient approach to continually adapt foundation models for federated learning. Further, a novel data-free approach is proposed to consolidate clients models at the server.
icon-lab / ProvoGANOfficial Implementation of Progressively Volumetrized Deep Generative Models for Data-Efficient Contextual Learning of MR Image Recovery
ArShx17 / Electricity Load ForecastingElectricity Load Forecasting using Machine Learning techniques to predict short-term power demand from historical time-series data for efficient energy planning and smart grid applications.
RahulRajGiri15 / Flight Fare Predictor Machine LearningFlight Fare Predictor is a machine learning project that predicts flight ticket prices using historical flight data. The model analyzes key travel features to estimate fares accurately, helping users plan trips more efficiently.
LynnFernandes23 / Diabetes Prediction System The Diabetes Prediction System is a comprehensive web application utilizing HTML for the frontend and Python's Flask framework for the backend. It leverages Pandas and NumPy for efficient data management, employing Scikit-Learn to develop robust machine learning models. Matplotlib is used for insightful data visualization.
FardinRastakhiz / QuickCharNetQuickCharNet is a deep learning project that leverages an efficient character-level Convolutional Neural Network (CNN) for URL classification, aimed at enhancing Search Engine Optimization (SEO). The project includes datasets, model evaluation notebooks, and visualization scripts. Key features include data preprocessing, detailed model architecture
emirkaanozdemr / Face Mask Detection Using ResNet50This project uses transfer learning with the ResNet50 model to develop an efficient face mask detection system. By leveraging pre-trained weights and data augmentation, the model accurately classifies images as masked or unmasked, contributing to public health and safety by ensuring mask-wearing compliance.
