Subspace identification for predictive state representation by nuclear norm minimization

Glaude, Hadrien; Pietquin, Olivier; Enderli, Cyrille

doi:10.1109/adprl.2014.7010609

Cited by 6 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this case, a bigger subspace that maps both the system dynamics and a part of the noise will achieve better performance because the noise can be next reduced during regression by means of regularization. The trade off between approximation and estimation errors is handled by solving a rank minimization problem [18].…”

Section: The Learning Algorithmmentioning

confidence: 99%

Learning of scanning strategies for electronic support using predictive state representations

Claude

Enderli

Grandin

et al. 2015

2015 IEEE 25th International Workshop on Machine Learning for Signal Processing (MLSP)

Self Cite

View full text Add to dashboard Cite

In Electronic Support, a receiver must monitor a wide frequency spectrum in which threatening emitters operate. A common approach is to use sensors with high sensitivity but a narrow bandwidth. To maintain surveillance over the whole spectrum, the sensor has to sweep between frequency bands but requires a scanning strategy. Search strategies are usually designed prior to the mission using an approximate knowledge of illumination patterns. This often results in open-loop policies that cannot take advantage of previous observations. As pointed out in past researches, these strategies lack of robustness to the prior. We propose a new closed loop search strategy that learns a stochastic model of each radar using predictive state representations. The learning algorithm benefits from the recent advances in spectral learning and rank minimization using nuclear norm penalization.

show abstract

Section: The Learning Algorithmmentioning

confidence: 99%

Learning of scanning strategies for electronic support using predictive state representations

Claude

Enderli

Grandin

et al. 2015

2015 IEEE 25th International Workshop on Machine Learning for Signal Processing (MLSP)

Self Cite

View full text Add to dashboard Cite

show abstract

“…These advantages of PSR are helpful in solving the problems in HMM‐based methods. The most common PSR learning algorithm is the transformed PSR (TPSR) [8] as well as the improved PSR models [9–13], which uses assumed tests to construct the dynamic matrix and reduces its dimension via singular value decomposition (SVD). However, these algorithms were for applications in artificial intelligence and machine learning.…”

Section: Introductionmentioning

confidence: 99%

Method for operating mode identification of multi‐function radars based on predictive state representations

Yong-guang²,

Zhao

et al. 2017

IET Radar, Sonar & Navigation

View full text Add to dashboard Cite

“…MoM-based algorithms are often consistent with theoretical guarantees in form of finite-sample bounds. In addition, these algorithms are able to learn a large variety of models [1], [3], [15], [11]. In a recent work, [15] showed that numerous models are encompassed into the common framework of linear Sequential Systems (SSs) or equivalently Multiplicity Automata (MA).…”

Section: Introductionmentioning

confidence: 99%

Non-negative Spectral Learning for Linear Sequential Systems

Glaude

Enderli

Pietquin

2015

Neural Information Processing

Self Cite

View full text Add to dashboard Cite

Abstract. Method of moments (MoM) has recently become an appealing alternative to standard iterative approaches like Expectation Maximization (EM) to learn latent variable models. In addition, MoM-based algorithms come with global convergence guarantees in the form of finite sample bounds. However, given enough computation time, by using restarts and heuristics to avoid local optima, iterative approaches often achieve better performance. We believe that this performance gap is in part due to the fact that MoM-based algorithms can output negative probabilities. By constraining the search space, we propose a non-negative spectral algorithm (NNSpectral) avoiding computing negative probabilities by design. NNSpectral is compared to other MoM-based algorithms and EM on synthetic problems of the PAutomaC challenge. Not only, NNSpectral outperforms other MoM-based algorithms, but also, achieves very competitive results in comparison to EM.

show abstract

Subspace identification for predictive state representation by nuclear norm minimization

Cited by 6 publications

References 11 publications

Learning of scanning strategies for electronic support using predictive state representations

Learning of scanning strategies for electronic support using predictive state representations

Method for operating mode identification of multi‐function radars based on predictive state representations

Non-negative Spectral Learning for Linear Sequential Systems

Contact Info

Product

Resources

About