Next time you cross a crowded plaza, crosswalk, or airport concourse, take note of the pedestrian flow. Are people walking in orderly lanes, single-file, to their respective destinations? Or is it a haphazard tangle of…
Category: 7. Maths
-

Bio-inspired neural networks with central pattern generators for learning multi-skill locomotion
Robot platform and simulation setup
The Unitree Go1 quadruped robot is used as the robot platform in simulation and real-world experiments. The Go1 robots weighs 12 kg with a dimension of around 645 \(\times\) 280 \(\times\) 400 mm while…
Continue Reading
-

Enhanced hybrid CNN and transformer network for remote sensing image change detection
To comprehensively validate the superiority of our proposed EHCTNet model, we have conducted extensive comparisons with SOTA change detection methods on two large-scale, high spatial resolution remote sensing (RS) image datasets, namely, LEVIR-CD
Continue Reading
-

A deep learning-based hybrid method for PM2.5 prediction in central and western China
The proposed PSO-Transformer-LSTM model is a hybrid architecture that integrates the advantages of the transformer model, the LSTM network and the particle swarm optimization (PSO) algorithm. This section describes the main components of the…
Continue Reading
-

Data harmonization for the analysis of personalized treatment of psychosis with metacognitive training
Data harmonization is a process of conciliating various types, levels, and sources of data into compatible and comparable formats to ensure that the data can be effectively utilized for better decision-making47. It is a complex process involving…
Continue Reading
-

Two-tier nature inspired optimization-driven ensemble of deep learning models for effective autism spectrum disorder diagnosis in disabled persons
In this paper, a novel T2MEDL-EASDDP model is developed. The main aim of the presented T2MEDL-EASDDP model is to analyze and diagnose the different stages of ASD in disabled individuals. The T2MEDL-EASDDP model has data normalization,…
Continue Reading
-

A hybrid zero-reference and dehazing network for joint low-light underground image enhancement
In this section, the experiment configurations and results using the Z-DCE-DNet for Underground Mine image enhancement are presented. Firstly, the PyTorch deep learning framework was used, running on an NVIDIA GeForce RTX 1650 GPU, applied to…
Continue Reading
-

Image segmentation and coverage estimation of deep-sea polymetallic nodules based on lightweight deep learning model
Dataset building
The dataset utilized in this research is primarily sourced from the Chinese Ocean Polymetallic Nodule Survey, conducted in the Clarion-Clipperton Fracture Zone (CCZ) of the equatorial northeastern Pacific. This dataset encompasses…
Continue Reading
-

Efficient self-attention with smart pruning for sustainable large language models
LLMs, being deep neural networks, have several layers, including self-attention mechanisms, feedforward networks, and embedding layers, all of which require significant amounts of memory and computation. The self-attention mechanism, for example,…
Continue Reading
-

Machine learning insights for sustainable hydroponic cultivation and growth monitoring of allium cepa using smart hydro kit
Agriculture plays a vital role in enhancing several nations’ economic growth and food security worldwide. In India, onion (Allium Cepa) is one of the most widely cultivated and consumed vegetable and cash crops with indispensable commercial,…
Continue Reading
