The description of electron correlation in quantum chemistry often relies on multi-index quantities. Here, we examine a compressed representation of the long-range part of electron correlation that is associated with dispersion interactions. For this purpose, we perform coupled-cluster singles and doubles (CCSD) computations on localized orbitals, and then extract the portion of CCSD amplitudes corresponding to dispersion energies. Using singular value decomposition, we uncover that a very compressed representation of the amplitudes is possible in terms of occupied-virtual geminal pairs located on each monomer. These geminals provide an accurate description of dispersion energies at medium and long distances. The corresponding virtual orbitals are examined by further singular value decompositions of the geminals. We connect each component of the virtual space to the multipole expansion of dispersion energies. Our results are robust with respect to basis set change and hold for systems as large as the benzene-methane dimer. This compressed representation of dispersion energies paves the way to practical and accurate approximations for dispersion, for example, in local correlation methods.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1063/1.4997186 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Estimation of Compressive Strength of Basalt Fiber-Reinforced Kaolin Clay Mixture Using Extreme Learning Machine.
Materials (Basel)
January 2025
Department of Geological Engineering, Firat University, Elazığ 23119, Türkiye.
Background: In this study, the unconfined compressive strength (q) of a mixture consisting of clay reinforced with 24 mm-long basalt fiber was estimated using extreme learning machine (ELM). The aim of this study is to estimate the results closest to the data obtained through experimental studies without the need for experimental studies. The literature review reveals that the ELM technique has not been applied to predict the compressive strength of basalt fiber-reinforced clay, and this study aims to provide a novel contribution in this area.
View Article and Find Full Text PDFSemantic segmentation using synthetic images of underwater marine-growth.
Front Robot AI
January 2025
AAU Energy, Aalborg University, Esbjerg, Denmark.
Introduction: Subsea applications recently received increasing attention due to the global expansion of offshore energy, seabed infrastructure, and maritime activities; complex inspection, maintenance, and repair tasks in this domain are regularly solved with pilot-controlled, tethered remote-operated vehicles to reduce the use of human divers. However, collecting and precisely labeling submerged data is challenging due to uncontrollable and harsh environmental factors. As an alternative, synthetic environments offer cost-effective, controlled alternatives to real-world operations, with access to detailed ground-truth data.
View Article and Find Full Text PDFDoes amplitude compression help or hinder attentional neural speech tracking?
J Neurosci
January 2025
Department of Psychology, University of Lübeck, Lübeck, Germany.
Amplitude compression is an indispensable feature of contemporary audio production and especially relevant in modern hearing aids. The cortical fate of amplitude-compressed speech signals is not well-studied, however, and may yield undesired side effects: We hypothesize that compressing the amplitude envelope of continuous speech reduces neural tracking. Yet, leveraging such a 'compression side effect' on unwanted, distracting sounds could potentially support attentive listening if effectively reducing their neural tracking.
View Article and Find Full Text PDFPEDRA-EFB0: colorectal cancer prognostication using deep learning with patch embeddings and dual residual attention.
Med Biol Eng Comput
January 2025
Radiol Dept, Jiangnan Univ, Affiliated Hosp, Wuxi, 214122, Jiangsu, People's Republic of China.
In computer-aided diagnosis systems, precise feature extraction from CT scans of colorectal cancer using deep learning is essential for effective prognosis. However, existing convolutional neural networks struggle to capture long-range dependencies and contextual information, resulting in incomplete CT feature extraction. To address this, the PEDRA-EFB0 architecture integrates patch embeddings and a dual residual attention mechanism for enhanced feature extraction and survival prediction in colorectal cancer CT scans.
View Article and Find Full Text PDFOptimal sparsity in autoencoder memory models of the hippocampus.
bioRxiv
January 2025
Center for Theoretical Neuroscience, Zuckerman Mind Brain Behavior Institute, Columbia University, New York, NY.
Storing complex correlated memories is significantly more efficient when memories are recoded to obtain compressed representations. Previous work has shown that compression can be implemented in a simple neural circuit, which can be described as a sparse autoencoder. The activity of the encoding units in these models recapitulates the activity of hippocampal neurons recorded in multiple experiments.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!