Large Language Models (LLMs) have shown excellent generalization capabilities that have led to the development of numerous models. These models propose various new architectures, tweaking existing architectures with refined training strategies, increasing context length, using high-quality training data, and increasing training time to outperform baselines. Analyzing new developments is crucial for identifying changes that enhance training stability and improve generalization in LLMs. This survey paper comprehensively analyses the LLMs architectures and their categorization, training strategies, training datasets, and performance evaluations and discusses future research directions. Moreover, the paper also discusses the basic building blocks and concepts behind LLMs, followed by a complete overview of LLMs, including their important features and functions. Finally, the paper summarizes significant findings from LLM research and consolidates essential architectural and training strategies for developing advanced LLMs. Given the continuous advancements in LLMs, we intend to regularly update this paper by incorporating new sections and featuring the latest LLM models.

pdf/a-comprehensive-overview-of-large-language-models-3hxe0yb7.pdf

A Comprehensive Overview of Large Language Models

Optimal planning and operation of grid-connected PV/CHP/battery energy system considering demand response and electric vehicles for a multi-residential complex building

With improvements in synthetic aperture radar (SAR) image resolutions and multitemporal SAR interferometry (MTInSAR) algorithms, it becomes feasible to generate 3D point clouds with millimeter deformation accuracy to monitor urban infrastructure health in detail. This article reviews the progress and challenges of MT interferometric SAR (InSAR) techniques for infrastructure deformation monitoring, with an emphasis on refinement performance. After a brief overview of literature statistics, we examine characteristics of the main MTInSAR algorithms. We demonstrate performance refinement in terms of point density, height accuracy, and deformation accuracy, using a comparative study of the reclaimed Hong Kong International Airport (HKIA). A series of typical infrastructure dynamic behaviors in urban areas revealed by MTInSAR is then illustrated. This article also describes key issues that remain to be addressed to facilitate operational fine surveillance. Finally, we go beyond MTInSAR to illustrate intelligent infrastructure diagnosis systems applicable to city-scale and even infrastructural component-scale analysis. It is our expectation that this review will provide a reference and guidance for future research and applications of MTInSAR in urban infrastructure monitoring.

Toward Fine Surveillance: A review of multitemporal interferometric synthetic aperture radar for infrastructure health monitoring

Solar and wind energy sources hold significant potential to meet the escalating energy demand in Saudi Arabia sustainably. This research aims to assess the feasibility and prospects of deploying solar photovoltaic (PV) and wind energy systems in Saudi Arabia (SA). The study adopts a comprehensive approach, encompassing spatial analysis of land suitability, techno-economic feasibility studies, and prediction of renewable energy (RE) resource variability and trends using machine learning techniques. Geographical Information System (GIS) spatial analysis is employed to identify suitable areas for establishing solar PV and wind farms based on multi-criteria evaluation methods. Techno-socioeconomic feasibility analysis of solar PV and wind energy systems, under various configurations and scenarios, is conducted using the System Advisor Model (SAM) tool. Furthermore, a Support Vector Machine (SVM) machine learning algorithm is developed and deployed to forecast Global Horizontal Irradiation (GHI), wind speed, and influential weather parameters at different locations in SA. The projections of these variables derived from global climate models are utilized to analyze prospects and variabilities. The land suitability analysis identified four optimal locations for large-scale solar PV fields in Tabuk, Al Madinah, Makkah, and Riyadh provinces. Additionally, four promising wind farms were found in Al Madinah, Makkah, Riyadh, and Eastern provinces. The analysis also identified one location in Al Jouf province suitable for hybrid systems combining solar PV and wind energy. The techno-economic assessment revealed that wind farms performed best overall, achieving a capacity factor of 42.6 % in Al Madinah province. Although current tariffs render projects economically unviable, solar PV, wind energy, and hybrid solar PV-wind technologies are economically feasible in SA at Power Purchase Agreement (PPA) rates above $32.8/MWh, $26.1/MWh, and $50.6/MWh, respectively. The social development analysis estimated potential job creation from solar PV and wind energy deployment under different scenarios from 2020 to 2060, indicating that more ambitious climate targets could translate to millions of renewable energy jobs. The SVM model predicted solar irradiance with an R-squared value of 0.893. The CMIP6 model was then used to project the GHI for SA in 2049, suggesting an increase of approximately 19 %. Wind speed is also expected to rise roughly 5 % over the same period. The integration of GIS spatial analysis, techno-economic modeling, and machine learning-based forecasting provides comprehensive insights into harnessing solar and wind energy in SA. This study facilitates evidence-based planning and risk assessment, crucial for a sustainable energy transition. The insights and roadmap derived from this research can inform policy frameworks, support the United Nations Sustainable Development Goals (UNSDGs), and attract private investments for RE development in SA. Consequently, this study establishes solar and wind energy as viable and promising solutions for meeting SA's growing energy demands sustainably, while minimizing the associated environmental impact. 

Potentials and opportunities of solar PV and wind energy sources in Saudi Arabia: Land suitability, techno-socio-economic feasibility, and future variability

Integrated GIS-AHP-based approach for off-river pumped hydro energy storage site selection

Geospatial-assisted multi-criterion analysis of solar and wind power geographical-technical-economic potential assessment

The refugee crisis forces various security, economic and natural weight on host nations. To house refugees, camps are considered a preferred option by humanitarian agencies and governments. The siting of refugee camps is an inherently complex and multidisciplinary task and can result in further displacement and long-term risks and distress to people of concern. This study proposes a geospatial decision-making approach combining geographic information systems (GIS) technology and fuzzy-analytical hierarchy process (F-AHP) method to assess land suitability and identify the optimal sites of refugee camps in Kenya. Four main factors and 19 evaluation criteria relevant to geographical, environmental, infrastructure, and social aspects are thoroughly investigated. The F-AHP is applied to determine the weights of the adopted criteria in consultation with three experts. The geospatial analysis is established using ArcGIS Pro software to derive the suitability map of refugee camps. The results concluded that Kenyan territory is classified into four ranks, low suitable represents 6.5% of the total surface area, 10.6% moderate suitable, 10.3% suitable, and 7.5% high suitable. Further, the analysis showed that the existing camps are located in unsuitable areas, whether north or east. Thus, the current study recommends moving these camps to areas far away from the flooded risk lands. In addition, eighteen new camps sites were suggested to host current and future refugees as well. The majority of these sites are located in Northeast, Northwest, and Northern Kenya, near the Kenyan borders with refugees' countries of origin. The findings revealed that the proposed new locations have superlative m2/refugee and approved the minimum area for a refugee (45 m2) defined by UNHCR. 

Spatial suitability analysis for site selection of refugee camps using hybrid GIS and fuzzy AHP approach: The case of Kenya

Urban areas are subject to subsidence due to varying natural and anthropogenic causes. Often, subsidence is interpreted and correlated to a single causal factor; however, subsidence is usually more complex. In this study, we adopt a new model to distinguish different causes of subsidence in urban areas based on complexity. Ascending and descending Sentinel-1 data were analyzed using permanent scatterer interferometry (PS-InSAR) and decomposed to estimate vertical velocity. The estimated velocity is correlated to potential causes of subsidence, and modeled using different weights, to extract the model with the highest correlations among subsidence. The model was tested in Alexandria City, Egypt, based on three potential causes of subsidence: rock type, former lakes and lagoons dewatering (FLLD), and built-up load (BL). Results of experiments on the tested area reveal singular patterns of causal factors of subsidence distributed across the northeast, northwest, central south, and parts of the city center, reflecting the rock type of those areas. Dual causes of subsidence are found in the southwest and some parts of the southeast as a contribution of rock type and FLLD, whereas the most complex causes of subsidence are found in the southeast of the city, as the newly built-up areas interact with the rock type and FLLD to form a complex subsidence regime. Those areas also show the highest subsidence values among all other parts of the city. The accuracy of the final model was confirmed using linear regression analysis, with an R2 value of 0.88.

A Model for Complex Subsidence Causality Interpretation Based on PS-InSAR Cross-Heading Orbits Analysis

Extracting informative representations from videos is fundamental for the effective learning of various downstream tasks. Inspired by classical works on saliency, we present a novel information-theoretic approach to discover meaningful representations from videos in an unsupervised fashion. We argue that local entropy of pixel neighborhoods and its evolution in a video stream is a valuable intrinsic supervisory signal for learning to attend to salient features. We, thus, abstract visual features into a concise representation of keypoints that serve as dynamic information transporters . We discover in an unsupervised fashion spatio-temporally consistent keypoint representations that carry the prominent information across video frames, thanks to two original information-theoretic losses. First, a loss that maximizes the information covered by the keypoints in a frame. Second, a loss that encourages optimized keypoint transportation over time, thus, imposing consistency of the information ﬂow. We evaluate our keypoint-based representation compared to state-of-the-art baselines in different downstream tasks such as learning object dynamics. To evaluate the expressivity and consistency of the keypoints, we propose a new set of metrics. Our empirical results showcase the superior performance of our information-driven keypoints that resolve challenges like attendance to both static and dynamic objects, and to objects abruptly entering and leaving the scene.

An information-theoretic approach to unsupervised keypoint representation learning

Extracting informative representations from videos is fundamental for effectively learning various downstream tasks. We present a novel approach for unsupervised learning of meaningful representations from videos, leveraging the concept of image spatial entropy (ISE) that quantifies the per-pixel information in an image. We argue that \textit{local entropy} of pixel neighborhoods and their temporal evolution create valuable intrinsic supervisory signals for learning prominent features. Building on this idea, we abstract visual features into a concise representation of keypoints that act as dynamic information transmitters, and design a deep learning model that learns, purely unsupervised, spatially and temporally consistent representations \textit{directly} from video frames. Two original information-theoretic losses, computed from local entropy, guide our model to discover consistent keypoint representations; a loss that maximizes the spatial information covered by the keypoints and a loss that optimizes the keypoints' information transportation over time. We compare our keypoint representation to strong baselines for various downstream tasks, \eg, learning object dynamics. Our empirical results show superior performance for our information-driven keypoints that resolve challenges like attendance to static and dynamic objects or objects abruptly entering and leaving the scene.

Entropy-driven Unsupervised Keypoint Representation Learning in Videos

Handle everyday research tasks with reliable, citation-backed results

Your personal Research Agent to handle research tasks with citation-backed results

Popular Tasks used by Researchers

How can I help with your research?

Meet SciSpace

Get more enhanced response by uploading the PDFs you want me to reference.

No relevant PDFs in your library

SciSpace is the AI research assistant for academics. Run systematic literature reviews on 280M+ papers, and write papers with cited sources. Trusted by 1M+ students, PhDs & researchers.

SciSpace | AI for Research

Analyze PDFs

Code & Manuscripts

Funding & Grants

Literature & Patents

Medical & Clinical Data

Systematic Review

Visualize & Present

Web & Data

Build a Google Scholar-like website for your research.

Build a website

Create charts and images for your research

Create a Chart

Write a paper for submission to a journal

Draft a manuscript

Patent Search

Design eye-catching scientific posters in minutes.

Scientific Poster Generation

Systematic Literature Review

One task is running at the moment. Your messages will be shown right after.

Drag and drop or click here to browse

Loved by <highlight>1 million+</highlight> researchers

Extract a list of specific topics and their sources from unstructured text

Topics

Compare and analyze relevant papers that matches with your search

Papers

Get insights from PDFs and bookmarked papers from your library

My library

Recent searches

Try searching for:

Catch AI-generated content in scholarly and non-scholarly content

{ai} Detector

Ai Writer

Get PDF Summaries, highlighted text explanations 

Chat with PDF

Effortlessly create in-text citations and bibliographies in APA and 2,500 other formats

Citation generator

Get explanations, summaries, and answers on academic papers

Ease up your research workflow with {scispace}'s cohort of exciting AI tools

Elevate your academic writing skills and convey your ideas the way you want

Paraphraser

Explore our range of reading and writing tools

Your file is being prepared and should be ready in a few minutes. If it's a large file, it might take a bit longer. You can close this window, and we'll email you the file when it's done.

You have reached a maximum limit of <strong>{limit}</strong> columns in the table. Remove at least <strong>1</strong> column to add or create another one.

Ali Younes

Author Tools

Chat about Author