Projection approximation subspace tracking

By Bin Yang. 1995

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

The Ideal Continual Learner: An Agent That Never Forgets (2023, background): “To do so, a classic idea (e.g., see Yang (1995; 1996)) is to define a forgetting factor β with 0 β ≤ 1, and minimize the following objective in a streaming fashion: min U T∑ t=1 βT−t‖Yt −UU>Yt‖2F. (32) Here, the forgetting factor β discounts the previous samples, accounting for the fact that St is…”
GSVD-based optimal filtering for single and multimicrophone speech enhancement (2002, methodology): “Rezayee and Gazor [14] have reduced the computational complexity of the signal subspacebased speech enhancement techniques by using an adaptive KLT tracking algorithm, namely, the projection approximation subspace tracking (PAST) with deflation [17].”
Robust Subspace Learning: Robust PCA, Robust Subspace Tracking, and Robust Subspace Recovery (2017, background): “Many algorithms exist, e.g., see [38], [39], [40], however all existing guarantees are asymptotic results for the statistically stationary setting of data being generated from a single unknown subspace; and most are partial guarantees.”
Stochastic subspace identification for output‐only modal analysis: application to super high‐rise tower under abnormal loading condition (2013, methodology): “2013; 42:477–498 DOI: 10.1002/eqe approach called PAST was initially developed by Yang [12], which takes advantage of a mathematical lemma to find the required column subspace as an unconstrained optimization problem.”
A Hebbian/Anti-Hebbian Neural Network for Linear Subspace Learning: A Derivation from Multidimensional Scaling of Streaming Data (2015, methodology): “A natural way to track an evolving subspace is to “forget” the contribution of older data points (Yang, 1995).”
Streaming Pattern Discovery in Multiple Time-Series (2005, methodology): “We use an algorithm based on adaptive filtering techniques [37, 24], which have been tried and tested in practice, performing well in a variety of settings and applications (e.”
A Consolidated Perspective on Multimicrophone Speech Enhancement and Source Separation (2017, unknown): “A tracking mechanism for the RTFs of the desired and interference sources, based on subspace tracking [375], is described in [198].”
Adaptive blind signal and image processing (2002, methodology): “∆W(k) = ē(k)z(k) PAST Algorithm z(k) = 1 α Q(k − 1)y(k) Yang (1995) [1318] y(k) = W(k − 1)x(k) γ(k) = [1 + yH(k)z(k)]−1 ē(k) = γ(k)[x̄(k)−W(k − 1)y(k)]”
Multi-Pitch Estimation (2009, background): “Some examples of subspace trackers that track only an arbitrary basis are PAST [179], NIC [118], FAPI [8], and FDPM [45], while the PASTd [179], SWASVD3 [10], LORAF3 [165], and FST [138] are examples of subspace trackers that track the eigenvectors and are thus applicable to the problem of determini”
A feasible method for optimization with orthogonality constraints (2012, background): “Computing the principle invariant subspace of a symmetric or Hermitian matrix arises in adaptive filtering, direction finding, and other signal processing problems [45, 58].”
Unsupervised Neural Learning on Lie Group (2002, background): “42 ; 58 We propose here a quite di(cid:11)erent perspective, well suited for both supervised and unsupervised learning tasks.”
Provable Dynamic Robust PCA or Robust Subspace Tracking (2017, unknown): “The subspace tracking (ST) problem (without outliers), and with or without missing data, has been extensively studied [23]–[28]; however, all existing guarantees are asymptotic results for the statistically stationary setting of data being generated from a single unknown subspace.”
Signal enhancement using beamforming and nonstationarity with applications to speech (2001, background): “A subspace tracking solution [30] has been proposed.”
Streaming PCA and Subspace Tracking: The Missing Data Case (2018, methodology): “The Projection Approximation Subspace Tracking (PAST) is proposed by Yang [25], [28] for subspace tracking with full data, which is described in Algorithm 5.”