Information Retrieval Perspective to Nonlinear Dimensionality Reduction for Data Visualization

By Jarkko Venna, Jaakko Peltonen, Kristian Nybo, Helena Aidos, Samuel Kaski. 2010

407 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Retrieval & Search, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Information Retrieval & Search

In Representation & Self-Supervised Learning

Citations in the corpus

UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction (2018, methodology): “Algorithms such as PCA [8], MDS [9], and Sammon mapping [20] fall into the former category while t-SNE [15] [26], Isomap [24], LargeVis [22], Laplacian eigenmaps [1] [2], diffusion maps [4], NeRV [28], and JSE [12] all fall into the latter category.”
TriMap: Large-scale Dimensionality Reduction Using Triplets (2019, background): “Due to the high computational complexity for calculating the trustworthiness-continuity and AUC scores for large data sets, we use nearest-neighbors accuracy as the local measure of performance henceforth.”
Barnes-Hut-SNE (2013, background): “A plethora of embedding techniques have been proposed over the last decade, e.g. , [5, 15, 20, 23, 25, 26].”
Stochastic Neighbor Embedding under f-divergences (2018, methodology): “KL-SNE (i.e. f taken as KL-divergence, cf. Table 1) for instance penalizes pairs of nearby points in the original space getting mapped far away in the embedded space more heavily than faraway points being mapped nearby (since the corresponding pij qij ≈ 0).”
Minimum curvilinearity to enhance topological prediction of protein interactions by network embedding (2013, background): “Details on the characteristics of the networks are provided in Supplementary Section I and Table S1.”
Inductive Hashing on Manifolds (2013, background): “Moreover, as stated in [30], the cost function of t-SNE in fact maximizes the smoothed recall [30] of query points and their neighbors.”
Accelerating t-SNE using tree-based algorithms (2014, background): “In particular, a family of techniques that preserves small pairwise distances via stochastic neighbor embedding (SNE; Hinton and Roweis (2003)) has recently gained popularity (Carreira-Perpiñán, 2010; van der Maaten and Hinton, 2008; Venna et al., 2010).”
Generalized Autoencoder: A Neural Network Framework for Dimensionality Reduction (2014, methodology): “Various methods of dimensionality reduction have been proposed to discover the underlying manifold structure, which plays an important role in many tasks, such as pattern classification [9] and information visualization [16].”
Shift-invariant similarities circumvent distance concentration in stochastic neighbor embedding and variants (2011, methodology): “In terms of results, t-SNE and NeRV significantly outperform most of the older methods, as shown in [21, 22, 23, 24] for instance.”
A Unifying Perspective on Neighbor Embeddings along the Attraction-Repulsion Spectrum (2020, background): “Related work Various trade-offs in SNE and t-SNE generalizations have been studied previously (Yang et al., 2009; Kobak et al., 2020; Venna et al., 2010; Amid et al., 2015; Amid and Warmuth, 2019; Narayan et al., 2015; Im et al., 2018), but our work is the first to study the exaggeration-induced tra”
Linear discriminant analysis: A detailed tutorial (2017, neutral): “Dimensionality reduction techniques are important in many applications related to machine learning [15], data mining [6,33], Bioinformatics [47], biometric [61] and information retrieval [73].”
Graph abstraction reconciles clustering with trajectory inference through a topology preserving map of single cells (2017, result): “We note that this is similar to the procedure introduced by [62].”
Perplexity-free t-SNE and twice Student tt-SNE (2018, background): “As to DR quality, some studies developed indicators of the HD neighbourhoods preservation in the LD space [18, 14], becoming generally adopted in several publications [7, 14].”
Hierarchical Stochastic Neighbor Embedding (2016, methodology): “We base our quantitative assessement of the quality of the embedding on the Nearest-Neighbor Preservation metric (NNP) as proposed by Venna et al. [VPN ∗ 10] and implemented by In-gram and Munzner [IM15].”
Information-Geometric Dimensionality Reduction (2011, methodology): “In particular, when the data samples are probability distributions, use of an information divergence such as KullbackLeibler (KL) instead of Euclidean distance leads to a class o f information geometric algorithms for dimensionality reduction [6], [7].”
Type 1 and 2 mixtures of Kullback-Leibler divergences as cost functions in dimensionality reduction based on similarity preservation (2013, background): “This means that the force between xi and xj vanishes only if sij ¼ sij and sji ¼ sji.”