AI News

RSS

Crack the Code: Advanced Method Enhances Train Rivet Inspection

This study proposes an innovative method for detecting cracks in train rivets using fluorescent magnetic particle detection (FMPFD) and instance segmentation, achieving high accuracy and recall. By enhancing the YOLOv5 algorithm and developing a single coil non-contact magnetization device, the researchers achieved significant improvements in crack detection.

17 May 2024

Mapping ML in Seismic Engineering

A recent scientometric review highlighted the transformative impact of machine learning (ML) in seismic engineering, showcasing advancements in material performance prediction and seismic resistance. The study, published in the journal Buildings, analyzed 3189 papers using the Scopus database, identifying key research trends and fostering collaboration within the field.

16 May 2024

AI Boosts Manufacturing Particle Detection

A novel framework combining deep learning and preprocessing algorithms significantly improved particle detection in manufacturing, addressing challenges posed by heterogeneous backgrounds. The framework, validated through extensive experimentation, enhanced in-situ process monitoring, offering robust, real-time solutions for diverse industrial applications.

16 May 2024

Machine Learning Tames Chaos on Edge Devices

This study highlights the use of a next-generation reservoir computing (NG-RC) algorithm to control chaotic dynamics with high sensitivity and efficiency. Implemented on an FPGA, the NG-RC-based controller outperformed conventional chaos control techniques by stabilizing various unstable states and trajectories with minimal power consumption, suggesting promising applications in diverse fields like autonomous systems and biological control.

16 May 2024

ChemCrow: AI-Driven Chemistry Breakthrough

In a Nature Machine Intelligence paper, researchers unveiled ChemCrow, an advanced LLM chemistry agent that autonomously tackles complex tasks in organic synthesis and materials design. By integrating GPT-4 with 18 expert tools, ChemCrow excels in chemical reasoning, planning syntheses, and guiding drug discovery, outperforming traditional LLMs and showcasing its potential to transform scientific research.

16 May 2024

Silent Speech Interface Using Graphene-Based Textile Strain Sensors and AI

Researchers introduced a groundbreaking silent speech interface (SSI) leveraging few-layer graphene (FLG) strain sensing technology and AI-based self-adaptation. Embedded into a biocompatible smart choker, the sensor achieved high accuracy and computational efficiency, revolutionizing communication in challenging environments.

15 May 2024

DPA-1: A Deep Potential Model for Molecular Simulation

Researchers introduced DPA-1, a deep potential model with a gated attention mechanism, for representing atomic system conformation and chemical spaces. DPA-1 demonstrated superior performance in learning potential energy surfaces (PES) compared to existing benchmarks, offering efficiency and interpretability.

15 May 2024

Machine Learning-based Predictive Modeling for Traffic Congestion

Researchers investigated the predictability of vehicle travel time and traffic status on Al-Madina Al-Monawara St in Amman, Jordan, using machine learning algorithms. Results showed high accuracy in predicting travel time and traffic status six hours ahead, with AdaBoost demonstrating exceptional performance. The study suggests integrating predictive models into navigation apps and leveraging recent traffic data for effective congestion identification and traffic management in urban areas.

15 May 2024

JARVIS-Leaderboard: Advancing Materials Design Benchmarking

The article introduces JARVIS-Leaderboard, an open-source platform facilitating materials design benchmarking across various categories like AI, electronic structure, force-field, quantum computation, and experiments. Integrated with NIST-JARVIS infrastructure, it offers a dynamic framework for comparing methods and datasets, fostering reproducibility and collaboration in materials science research.

15 May 2024

Better Structural Monitoring Using LSTM Neural Networks

Researchers utilized long-short-term memory (LSTM) neural networks to address sensor maintenance issues in structural monitoring systems, particularly during grid structure jacking construction. Their LSTM-based approach effectively recovered missing stress data by analyzing data autocorrelation and spatial correlations, showcasing superior accuracy compared to traditional methods.

14 May 2024

Smart Traffic Surveillance: Faster R-CNN for Vehicle Segmentation

Researchers introduced a novel deep learning approach based on faster R-CNN for segmenting vehicles in traffic videos, addressing challenges like occlusions and varying traffic densities. Through adaptive background modeling and topological active nets, the method achieved superior segmentation accuracy, showcasing its potential to enhance real-world traffic surveillance and management systems.

14 May 2024

SiTHv2: A Model for Global Evapotranspiration and Soil Moisture Dynamics

Researchers introduced SiTHv2, a terrestrial hydrosphere model, in Scientific Data, offering high-resolution daily estimates of evapotranspiration (ET) components and soil moisture dynamics. Utilizing satellite observations and reanalysis data, SiTHv2 was validated against in situ measurements, demonstrating accuracy and reliability for water cycle dynamics, drought assessment, and ecosystem health studies.

14 May 2024

Novel Metrics for Assessing Prediction Difficulty in Neural Networks

Researchers introduced innovative metrics to gauge the prediction difficulty of individual cases within datasets using neural networks (NN). These metrics, designed for fully connected feedforward NNs, offer a unique approach to understanding data complexity by considering factors like model complexity, double modeling, and predictive uncertainty.

14 May 2024

Neuromorphic Computing: Harnessing Spintronics' Potential

A recent paper in Npj Spintronics explores the frontier of neuromorphic computing, delving into the fusion of spintronics and magnetic materials to forge hardware-based neural networks. By dissecting spintronic synapses, neurons, and neural networks, the study reveals strides towards efficient and scalable neuromorphic systems, highlighting the reservoir computing paradigm's prowess.

13 May 2024

Snail-Inspired Robotic Swarms for Navigating Unstructured Environments

Researchers unveiled a terrestrial robotic swarm system inspired by land snails, featuring a unique two-mode connection mechanism for maneuvering in unstructured outdoor environments. By harnessing magnet-embedded tracks and vacuum suckers with directional polymer stalks, the system showcased individual robot capabilities and collective synergy through outdoor experiments, heralding a new era in real-world applications of robotic swarms.

13 May 2024

Advancing Sports Performance Using Motion Capture Technology

In a recent article published in Sensors, researchers conducted a thorough review of motion capture technology (MCT) in sports, comparing and evaluating various systems including cinematography capture, electromagnetic capture, computer vision capture, and multimodal capture.

13 May 2024

Smart Sensing and Predictive Analytics in Geotechnical Investigations

In their study published in the journal Smart Cities, researchers employed smart sensing and predictive analytics to address challenges in Japan's urban development and infrastructure resilience. Focusing on Setagaya, Tokyo, the research produced predictive models accurately determining critical bearing layer depths, crucial for government plans and construction risk assessments.

13 May 2024

Sustainable Smart Glasses with Text Mining, QFD and TRIZ

This article introduces an innovative methodology combining quality function deployment (QFD), text mining, and the theory of inventive problem solving (TRIZ) for sustainable product design, demonstrated through the design of smart glasses for augmented reality (AR) technology.

10 May 2024

Efficient Training Data Generation for Machine-Learned Interatomic Potentials

This article presents a novel approach, uncertainty-biased molecular dynamics (MD) simulations, for efficiently creating comprehensive datasets to train machine-learned interatomic potentials (MLIPs). By leveraging bias stress and automatic differentiation, this method captures rare events and extrapolative regions in configurational space, improving accuracy while reducing computational costs.

10 May 2024

Forecasting Metal Futures Using Machine Learning and Deep Learning

This paper investigates the prediction of metal commodity futures in financial markets through machine learning (ML) and deep learning (DL) models, analyzing multiple metals simultaneously. Despite promising results, variations in model performance across metals, input periods, and time frames underscore the challenges in consistently outperforming the market.

10 May 2024