On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,742 papers · 148 categories

Trend · papers per month

count % of corpus

4 results for “LargeVis”

TriMap: Large-scale Dimensionality Reduction Using Tripletscs.LG

TriMap improves data visualization by preserving global structure better than existing methods.

problem Visualizing high-dimensional data with preserved global structure.

method TriMap uses triplet constraints for dimensionality reduction.

result TriMap outperforms other methods in terms of runtime and quality of embedding.

dimensionality reduction triplets t-SNE LargeVis UMAP

2019-10-01abs ↗pdf ↗

Facilitate the Parametric Dimension Reduction by Gradient Clippingcs.LG

Parametric t-SNE improves generalization for streaming data.

problem Training neural networks for t-SNE objective function fails due to gradient exploding.

method Applied gradient clipping to solve gradient exploding problem.

result Parametric t-SNE achieves quality compatible with non-parametric t-SNE while generalizing to new data.

t-SNE dimension reduction gradient clipping neural networks streaming data

2020-09-30abs ↗pdf ↗

ShapeVis: High-dimensional Data Visualization at Scalecs.LG

ShapeVis visualizes high-dimensional data efficiently.

problem Scalability issues with Mapper in high-dimensional data.

method Landmarks, weighted witness-graph, induced maps, modularity-based pruning.

result ShapeVis scales to millions of data points while preserving visualization quality.

topological data analysis point cloud visualization high-dimensional data scalability Mapper

2020-01-15abs ↗pdf ↗

Statistical embedding: Beyond principal componentsstat.ML

Survey of embedding methods for high-dimensional and network data.

problem Embedding high-dimensional and nonlinear data structures in a lower-dimensional space.

method Survey of various embedding methods including principal curves, multidimensional scaling, graph-based methods, and topological embeddings.

result Discussion of the pros and cons of algorithmic machine learning and statistical modeling approaches.

embedding high-dimensional data network data principal curves multidimensional scaling

2021-06-03abs ↗pdf ↗