Semisupervised alignment of manifolds

By Jihun Ham, Daniel D. Lee, Lawrence K. Saul. International Conference on Artificial Intelligence and Statistics, 2005

313 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/087a090a83e01b0c741b53ec76ffad95ab31e010

How researchers see this paper

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

Citations in the corpus

Learning Transferable Knowledge Through Embedding Spaces (2019, methodology): “This can be accomplished with distribution matching [164], inter-task mapping [178], or manifold alignment techniques [179; 180].”
Learning Shared Latent Structure for Image Synthesis and Robotic Imitation (2005, methodology): “The GP model is closely related to other kernel methods for solving CCA [3] and similar problems [2].”
Manifold alignment using Procrustes analysis (2008, methodology): “Our approach differs from semi-supervised alignment (Ham et al., 2005) in that it results in a mapping that is defined everywhere rather than just on the known data points (provided a suitable dimensionality”
Learning Multi-View Neighborhood Preserving Projections (2011, methodology): “Ham et al. (2005) use pairwise correspondences between two views to perform manifold alignment based on a correlation-like measure. The resulting alignment, however, does not extend to out-of-sample examples. Jeon et al. (2003) and Monay & Gatica-Perez (2007) learn joint probabilistic models between”
Nonlinear Dimensionality Reduction with Local Spline Embedding (2009, background): “A parallel work related to out-of-sample extension for manifold learning is the semisupervised manifold learning [57], [23], [19].”
Parallel field alignment for cross media retrieval (2013, methodology): “And in the same period, manifold alignment methods [10, 29, 31] are proposed and shown that they are appropriate joint models for pair matching between heterogeneous data sources.”
Indoor localization in multi-floor environments with reduced effort (2010, methodology): “Given the 2-D location labels Y trn, we can use the 2D floor-plan as the common low-dimensional embedding space, and the image of each RSS vector in Xtrn is constrained by its 2-D location [19] (as given in Y trn).”
Optimal Transport for Domain Adaptation (2014, background): “Several works based on extraction of common features under pairwise constraints have also been introduced as domain adaptation strategies [26], [46], [51].”
Proto-transfer Learning in Markov Decision Processes Using Spectral Methods (2006, methodology): “For more complicated domains, such as two-room grids we are exploring manifold alignment techniques to allow appropriate mappings from the source to target domain (Ham et al., 2005).”
An Overview of Cross-Media Retrieval: Concepts, Methodologies, Benchmarks, and Challenges (2017, background): “ifolds that provide the common low-dimensional representations of two different high-dimensional datasets [88] (for”
Data Fusion and Multicue Data Matching by Diffusion Maps (2006, background): “This issue was overlooked in previous works based on spectral embeddings [24], [25], [26], [27], although spectral and ISOMAPS embeddings are highly sensitive to the way the data points were originally sampled.”
Using Task Descriptions in Lifelong Machine Learning for Improved Performance and Zero-Shot Transfer (2017, methodology): “This can be accomplished with distribution matching (Bickel et al., 2009), inter-task mapping (Taylor, Stone, & Liu, 2007), or manifold alignment techniques (Wang & Mahadevan, 2009; Ham, Lee, & Saul, 2005).”
Machine Learning Paradigms for Speech Recognition: An Overview (2013, unknown): “common latent space [216], [217]:”
Multimodal Manifold Analysis by Simultaneous Diagonalization of Laplacians (2015, unknown): “In Section 4, we show that many recent works on multi-view clustering [25], [26], [27], [28], [29] and manifold alignment [30], [31] can be considered as particular instances of our framework.”
Heterogeneous Domain Adaptation Using Manifold Alignment (2011, background): “Recently, a new approach to transfer learning based on manifold alignment was proposed to address this problem [Ham, Lee, and Saul, 2005; Wang and Mahadevan, 2009].”
A Survey of Multi-View Representation Learning (2016, unknown): “[24] also introduce manifold alignment with pairwise correspondence, which builds connections between multiple data sets by aligning their underlying manifolds.”
Transferring Localization Models across Space (2008, unknown): “In practice, we use˜L = L T L instead of L (Ham, Lee, and Saul 2005).”
Manifold Relevance Determination (2012, methodology): “…assume the existence of a low-dimensional representation encoding the underlying phenomenon, but the variance contained in an observation space does not necessarily need to be governed by the 2005)’s original model is shown, a single latent variable X is used to represent the observed data Y .”