Bringing UMAP Closer to the Speed of Light with GPU Acceleration

Nolet, Corey; Lafargue, Victor; Raff, Edward; Nanditale, Thejaswi; Oates, Tim; Zedlewski, John; Patterson, Joshua T.

doi:10.48550/arxiv.2008.00325

Cited by 4 publications

(3 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The approximate NN search makes the algorithm fast, but it also includes the possibility of small mistakes in the determination of nearest neighbors. Even though exact solutions exist, for example, the graphics processing unit (GPU)-accelerated UMAP implementation provides such an option (Nolet et al, 2020), and approximate NN-search within the original UMAP-learn python implementation (McInnes et al, 2018) provides enough accuracy for the typical fs-LIMS tasks. Furthermore, in the second stage, the algorithm weighs the nearest neighbors and forms the weighted NN-graph using smoothing kernels that adapt to the local neighborhood.…”

Section: Sample and Methodsmentioning

confidence: 99%

High Mass Resolution fs-LIMS Imaging and Manifold Learning Reveal Insight Into Chemical Diversity of the 1.88 Ga Gunflint Chert

Lukmanov

Koning

Schmidt

et al. 2022

Front. Space Technol.

View full text Add to dashboard Cite

Extraction of useful information from unstructured, large and complex mass spectrometric signals is a challenge in many application fields of mass spectrometry. Therefore, new data analysis approaches are required to help uncover the complexity of such signals. In this contribution, we examined the chemical composition of the 1.88 Ga Gunflint chert using the newly developed high mass resolution laser ionization mass spectrometer (fs-LIMS-GT). We report results on the following: 1) mass-spectrometric multi-element imaging of the Gunflint chert sample; and 2) identification of multiple chemical entities from spatial mass spectrometric data utilizing nonlinear dimensionality reduction and spectral similarity networks. The analysis of 40′000 mass spectra reveals the presence of chemical heterogeneity (seven minor compounds) and two large clusters of spectra registered from the organic material and inorganic host mineral. Our results show the utility of fs-LIMS imaging in combination with manifold learning methods in studying chemically diverse samples.

show abstract

Section: Sample and Methodsmentioning

confidence: 99%

High Mass Resolution fs-LIMS Imaging and Manifold Learning Reveal Insight Into Chemical Diversity of the 1.88 Ga Gunflint Chert

Lukmanov

Koning

Schmidt

et al. 2022

Front. Space Technol.

View full text Add to dashboard Cite

show abstract

“…For these reasons, a depth of three was chosen for all RAC descriptors used in this analysis. The Uniform Manifold Approximation and Projection (UMAP) technique 65 as implemented in the RAPIDS cuML Python library 66 was used to reduce the high dimensional data for each type of descriptor to only two dimensions for visualization purposes (e.g., the 20-dimensional metal-centre RAC descriptors were reduced to two dimensions using UMAP). Hyperparameters used for the generation of these plots are given in the supporting material.…”

Section: Diversity Analysismentioning

confidence: 99%

ARC-MOF: A Diverse Database of Metal-Organic Frameworks with DFT-Derived Partial Atomic Charges and Descriptors for Machine Learning

Burner

Luo

White

et al. 2022

Preprint

View full text Add to dashboard Cite

Metal-organic frameworks (MOFs) are a class of crystalline materials composed of metal nodes or clusters connected via semi-rigid organic linkers. Owing to their high surface area, porosity, and tunability, MOFs have received significant attention for numerous applications such as gas separation and storage. Atomistic simulations and data-driven methods (e.g., machine learning) have been successfully employed to screen large databases and successfully develop new experimentally synthesized and validated MOFs for CO2 capture. To enable data-driven materials discovery for any application, the first (and arguably most crucial) step is database curation. This work introduces the ab initio REPEAT charge MOF (ARC-MOF) database. This is a database of ~280,000 MOFs which have been either experimentally characterized or computationally generated, spanning all publicly available MOF databases. A key feature of ARC-MOF is that it contains DFT-derived electrostatic potential fitted partial atomic charges for each MOF. Additionally, ARC-MOF contains pre-computed descriptors for out-of-the-box machine learning applications. An in-depth analysis of the diversity of ARC-MOF with respect to the currently mapped design space of MOFs was performed – a critical, yet commonly overlooked aspect of previously reported MOF databases. Using this analysis, balanced subsets from ARC-MOF for various machine learning purposes have been identified. Other chemical and geometric diversity analyses are presented, with an analysis on the effect of charge assignment method on atomistic simulation of gas uptake in MOFs.

show abstract

“…Another topic modeling project used a different data source to analyze the discussions of COVID-19. Ordun et al [20] utilized LDA, UMAP [15,18], and other approaches to visualize the COVID-19 related posts on Twitter.…”

Section: Related Workmentioning

confidence: 99%

COVID-19 Kaggle Literature Organization

Eren

Solovyev

Raff

et al. 2020

Proceedings of the ACM Symposium on Document Engineering 2020

Self Cite

View full text Add to dashboard Cite

The world has faced the devastating outbreak of Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2), or COVID-19, in 2020. Research in the subject matter was fast-tracked to such a point that scientists were struggling to keep up with new findings. With this increase in the scientific literature, there arose a need for organizing those documents. We describe an approach to organize and visualize the scientific literature on or related to COVID-19 using machine learning techniques so that papers on similar topics are grouped together. By doing so, the navigation of topics and related papers is simplified. We implemented this approach using the widely recognized CORD-19 dataset to present a publicly available proof of concept. CCS CONCEPTS • Computing methodologies → Cluster analysis; Dimensionality reduction and manifold learning; • Human-centered computing → Visualization toolkits.

show abstract

Bringing UMAP Closer to the Speed of Light with GPU Acceleration

Cited by 4 publications

References 36 publications

High Mass Resolution fs-LIMS Imaging and Manifold Learning Reveal Insight Into Chemical Diversity of the 1.88 Ga Gunflint Chert

High Mass Resolution fs-LIMS Imaging and Manifold Learning Reveal Insight Into Chemical Diversity of the 1.88 Ga Gunflint Chert

ARC-MOF: A Diverse Database of Metal-Organic Frameworks with DFT-Derived Partial Atomic Charges and Descriptors for Machine Learning

COVID-19 Kaggle Literature Organization

Contact Info

Product

Resources

About