Pith. sign in

REVIEW 4 major objections 4 minor 49 references

Blind Sparse Estimation of Intermittent Sources over Unknown Fading Channels

T0 review · 4 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read A two-stage algorithm—smooth dictionary learning followed by per-source stochastic filtering—separates intermittent radio sources from fewer sensors over unknown flat-fading channels.

desk verdict A useful practical combination of smooth dictionary learning and per-source HMM post-filtering for intermittent sources, with a load-bearing ADMM derivation that needs fixing and a BAC model that deserves validation. read the letter →

arxiv 1908.01822 v1 pith:RK6BZ4IP submitted 2019-08-05 eess.SP

classification eess.SP
keywords blindsourceseparationdictionarylearninghiddenMarkovmodelsparserecoveryflatfadingchannelssmoothLASSOforward-backwardalgorithmexpectationmaximization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that intermittent radio sources can be separated and detected at a fusion center even when there are more sources than sensors, no channel state information, and unknown burst statistics, by combining dictionary learning with per-source stochastic filtering. It argues that source memory—bursts lasting many time samples—is itself a usable separability cue: smooth LASSO discourages short on/off flickers during dictionary learning, and a forward-backward filter then cleans the activity decisions by treating them as outputs of a binary asymmetric channel. The paper claims this two-stage design raises detection probability substantially (from 0.9 to 0.96 in its representative setup at 0.1 false-alarm probability and 30 dB SNR) and cuts signal estimation error by about 40 percent, while remaining robust as sources become denser, where conventional dictionary-learning baselines degrade sharply.

What carries the argument

The load-bearing mechanism is the pairing of a smoothness prior with a noise model on activity estimates. Smooth LASSO adds the penalty term μ Σ_t ||x(t) − x(t−1)||_2 to the dictionary-learning signal update, encoding the prior that bursts last many samples, with SL-SEQ and SL-ADMM as low-complexity approximations. The per-source stochastic filtering stage then quantizes each source's estimated signal into a binary activity sequence and models that sequence as the output of a binary asymmetric channel with flip probabilities p'_n and q'_n, so the forward-backward algorithm can compute MAP activity states; EM updates both HMM transition probabilities and flip probabilities from the data. The binary-asymmetric-channel model is what converts dictionary-learning mistakes into a tractable filtering problem.

What would settle it

Generate the same N=30, M=20, T=1000 scenario but replace the independent binary-asymmetric-channel flips in the PSF model with a Markov chain of flips, or with flip probabilities that depend on the transmitted symbol, then run the proposed algorithm and check whether the detection probability at 0.1 false-alarm probability and 30 dB SNR remains near 0.96; a drop toward the no-PSF level would confirm that the memoryless-BAC premise is carrying the result.

Watch

Extended reading notes

Core claim

The paper claims that in the model Y = HX + Z with more sources than sensors, X column-sparse with hidden-Markov-driven on/off states and H constant over the observation window, the fusion center can estimate H, X, and the activity states blindly by alternating channel and signal updates, with the signal step using smooth LASSO to penalize short transmissions, followed by a per-source filtering stage. The filtering stage thresholds each source's estimated signal into a binary activity sequence, models that sequence as the output of a binary asymmetric channel with flip probabilities, and runs forward-backward MAP estimation, with EM estimating both the HMM transition probabilities and the flip probabilities when they are unknown. In simulation with 30 potential sources, 20 sensors, an average of 3 active sources per time sample, 1000 time samples, and 30 dB SNR, per-source filtering raises the probability of detection from 0.9 to 0.96 at 0.1 false-alarm probability and reduces the error vector magnitude by about 40 percent; with EM-estimated parameters the gain remains, though somewhat reduced.

Load-bearing premise

The whole per-source filtering gain rests on the assumption in Sec. V-A that the dictionary-learning stage's activity mistakes are independent, per-time flips with fixed probabilities; the paper does not justify this memoryless binary-asymmetric-channel model, so if the mistakes are correlated over time or across sources, the reported detection gain is not guaranteed.

Editorial extensions

If this is right

  • In IoT-style settings with bursty transmitters, a fusion center can monitor more sources than antennas and still detect which are active, provided the activity patterns have memory and the channel is roughly constant over the observation block.
  • The smooth-LASSO front end alone already outperforms memory-agnostic OMP and LASSO baselines in the paper's simulations, and per-source filtering adds a further gain on top.
  • When HMM transition probabilities are unknown, EM-based per-source filtering recovers most of the detection gain, so the method does not require prior knowledge of burst statistics.
  • The reported gain is robust to the sparsity level for the smooth-LASSO plus filtering scheme, while the baseline schemes degrade sharply as the average number of active sources increases.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • An unstated corollary is that the method's success depends on bursts being long relative to the dictionary-learning window; if transmission durations shrink toward single symbols, the smoothness penalty has nothing to exploit and the per-source filter loses its edge.
  • If the binary-asymmetric-channel independence assumption fails, for instance when a missed detection makes nearby misses more likely, a natural extension is to model the flip process itself as Markov or to filter all sources jointly, which could preserve the gain in correlated-error regimes.
  • Because the two stages communicate only one-way, feeding the cleaned activity states back into the dictionary-learning objective as a warm start could further sharpen channel and signal estimates; the paper does not explore this closed-loop variant.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper considers blind estimation of N sparse, intermittently active sources from M<N sensors over unknown slow flat-fading channels, with source activity modeled as a two-state HMM. It proposes a two-stage algorithm: a dictionary learning stage with a smooth LASSO penalty on temporal activity changes (implemented as SL-SEQ and SL-ADMM), followed by per-source stochastic filtering that thresholds the DL activity estimates and passes them through a forward-backward HMM filter (with EM when HMM/BAC parameters are unknown). Numerical experiments with N=30, M=20, T=1000 report that the proposed SL+PSF method outperforms OMP and LASSO baselines in detection and EVM, including robustness to sparsity.

Significance. If validated, the architecture is a sensible combination of sparse dictionary learning and HMM-based post-filtering that addresses an underdetermined, bursty-source scenario of practical IoT interest. The paper's reported gains—e.g., Pd from 0.9 to 0.96 at Pfa=0.1 with unknown parameters, and a 40% EVM reduction—are nontrivial. The simulation study is fairly extensive, covering ROC curves, SNR, sparsity level, known/unknown parameters, and multiple baselines; the two-stage idea is clearly presented. However, the support is empirical and rests on several model assumptions that need to be checked before the central claim is fully established.

major comments (4)
  1. [Sec. IV-B, Eqs. (9)-(12)] There is a mismatch between the constraint and the dual update in the SL-ADMM derivation. Problem (9) imposes x'(t)=x(t) for t=1,...,T-1 and the augmented Lagrangian in (10) contains alpha^T(t)(x(t)-x'(t)), but the dual update in (12) uses x'(i)(t-1) rather than x'(i)(t). Moreover, (10) adds a rho-term on ||x(t)-x'(t-1)||2 while the constraint is on x(t)-x'(t); this corresponds to a different consensus problem. The text should correct the formulation and confirm that the implemented algorithm actually solves (7), or derive the correct ADMM updates.
  2. [Sec. V-A, Eq. (13)] With alpha_t defined as the filtered posterior Pr(s_n(t)=1|s_tilde(1:t)) and beta_t as the backward likelihood Pr(s_tilde(t+1:T)|s_n(t)=1), the equality Pr(s_n(t)=1|s_tilde(1:T)) = alpha_t beta_t is not correct; the right side is not normalized and can exceed one. The standard forward-backward posterior requires alpha_t to be the joint forward variable, or the final product must be normalized by the total observation likelihood. Please correct the definition or equation and ensure the MAP threshold comparison uses a proper probability.
  3. [Sec. V-A, Fig. 4 and Sec. VI] The PSF gain in the Conclusions depends on modeling the thresholded DL activity estimates as outputs of a memoryless binary asymmetric channel with fixed flip probabilities p'_n=0.02 and q'_n=0.27. The paper does not validate that the DL error process is memoryless, time-independent, or source-independent; errors produced by iterative dictionary learning are functions of the entire burst and the learned dictionary and can be correlated. Please add an empirical analysis of the activity-error process (e.g., autocorrelation of s_tilde_n(t)-s_n(t) and pairwise cross-correlation across sources), and report sensitivity of the Pd/Pfa curves to p' and q' or to using EM-estimated values instead of fixed ones.
  4. [Sec. VI-A] The statement that lambda and mu are optimized to satisfy given constants on Pd or Pfa is ambiguous: if this optimization uses the true activity labels, the reported ROC curves depend on oracle tuning, and it is unclear whether the OMP and LASSO baselines receive equally favorable tuning. Please specify the tuning protocol (grid, criterion, ground-truth dependence) and apply the same protocol to all compared algorithms.
minor comments (4)
  1. [Sec. II] Near the top of Sec. II, 'the fusion center has access to N receive antennas' should read 'M receive antennas', since the model uses M sensors.
  2. [Sec. II] The definition of the transition probabilities has a subscript typo: p_n = Pr(s_n(t)=1|s_{n-1}(t)=0) should be Pr(s_n(t)=1|s_n(t-1)=0), and similarly for q_n.
  3. [Sec. V-A] The MAP decision uses the same symbol s_hat_n(t) for both the quantized DL activity estimate and the final MAP estimate; using distinct notation would avoid confusion.
  4. [Sec. VI-C and Sec. VII] The EVM results condition on the set of correctly detected active time samples, which can favor methods with conservative detection; this conditioning should be stated more prominently. Also, the Conclusion contains a typo, 'decrease the EVM from by 40%'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the PSF detection gain is a simulation result measured against ground-truth activity and independent baselines, not a quantity constructed from the algorithm's own fitted outputs.

full rationale

The paper's derivation chain is self-contained and contains no step in which a 'prediction' reduces by construction to a fitted input or to a self-citation. The DL stage solves a smooth-LASSO problem (Eq. (7)) whose parameters lambda and mu are tuned numerically, and the PSF stage assumes a memoryless binary asymmetric channel for the quantized DL activity estimates (Fig. 4 and Eq. (13)). The BAC flip probabilities are either fixed simulation parameters (p' = 0.02, q' = 0.27) or estimated from the DL output via EM (Eqs. (17)-(18)); they are not fitted to the ground-truth detection probability that the paper claims to improve. The reported gain from 0.9 to 0.96 detection probability is obtained by comparing SL-ADMM with and without PSF on synthetic data generated from a known HMM, where the baselines OMP and LASSO are independent of the PSF model. No uniqueness theorem is imported from the authors' prior work, and the cited references are external standard methods. The main weakness, namely that the BAC assumption is not validated against temporally correlated DL errors, is a model-misspecification or robustness concern, not a circularity: the paper does not define the detection improvement in terms of the BAC posterior it assumes. Thus the appropriate finding is no significant circularity.

Assumptions & free parameters 7 free parameters · 5 assumptions · 1 invented entities

The method rests on a set of modeling assumptions and tuned numerical parameters. The central empirical claim is supported only by simulations, not by a theorem, so the burden falls on assumptions about sparsity, Markov activity, the BAC error structure, and convergence of nonconvex dictionary learning.

free parameters (7)
  • lambda (sparsity penalty) = lambda = 1/SNR (rule of thumb, numerically tuned)
    Sparsity penalty in Eq. (7) and Eq. (8); tuned per operating point in Sec. VI-A to meet target Pd/Pfa values.
  • mu (temporal smoothness penalty) = mu = 0.1/q_n, with q_n the true transition probability
    Penalty on changes between consecutive time samples; depends on source parameters and is tuned to match target detection performance.
  • gamma (state quantization threshold) = 0.5
    Threshold used in Sec. V-A to map |x_tilde_n(t)| to a binary activity state; fixed without sensitivity analysis.
  • rho (ADMM penalty parameter) = 0.1
    Parameter in the augmented Lagrangian and dual update of SL-ADMM, set without a stated selection rule.
  • K (SL-ADMM iterations) = 30
    Number of ADMM iterations fixed in Sec. VI; no convergence check is reported.
  • BAC flip probabilities in known-parameter PSF = p'_n = 0.02, q'_n = 0.27
    Fixed as known inputs in the PSF simulations (Sec. VI); these are privileged values rather than estimated quantities.
  • EM initialization and stopping tolerance = p_n = 0.5, q_n = 0.5, p'_n = 0.1, q'_n = 0.2; epsilon unspecified
    Initialization values for the EM stage (Sec. VI) and the stopping threshold in Sec. V-B are not justified or fully specified.
assumptions (5)
  • domain assumption The channel H is constant over T symbol periods, noise is iid zero-mean complex Gaussian, sensors are time-synchronized, and backhaul to the fusion center is ideal.
    Invoked in Sec. II, Eq. (1) and the synchronization discussion; if the channel varies within the block or sensors are not synchronized, the linear mixture model and the algorithm premise fail.
  • domain assumption Each source's activity s_n(t) follows a two-state Markov chain with transition probabilities p_n and q_n, and active transmissions are iid from a known distribution such as Gaussian or binary.
    Introduced in Sec. II and Fig. 2; the sparsity and temporal memory of this model are what make the underdetermined BSS problem tractable and drive both the smoothness penalty and the PSF stage.
  • domain assumption The alternating dictionary-learning minimization over (H, X) converges to a useful dictionary and sparse coefficient matrix despite being nonconvex and underdetermined.
    The paper uses iterative MOD/MDU and sparse coding (Sec. III) without convergence or identifiability guarantees; if the DL stage settles in a poor local optimum, the whole pipeline inherits the error.
  • ad hoc to paper The quantized DL state estimates are related to the true states through a memoryless binary asymmetric channel with flip probabilities p'_n and q'_n.
    Assumed in Sec. V-A and Fig. 4 to justify forward-backward filtering; no evidence is given that DL state errors are independent per time or across sources, making this the most fragile modeling step.
  • standard math Standard algorithms (forward-backward, EM, ADMM) converge to the claimed estimates in this setting.
    Forward-backward and EM on well-specified HMMs have standard guarantees, but the ADMM variant in Sec. IV-B is not derived consistently and its convergence for the nonconvex DL problem is not established.
invented entities (1)
  • Memoryless binary asymmetric channel (BAC) representing DL state errors
    purpose: Connects the true HMM state s_n(t) to the quantized DL estimate s_tilde_n(t) so forward-backward filtering can be applied.
    The BAC is introduced in Sec. V-A purely as a modeling convenience; the paper provides no test of whether actual DL errors are memoryless or independent, so this construct has no falsifiable handle outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Blind Sparse Estimation of Intermittent Sources over Unknown Fading Channels." pith.science (2026). https://pith.science/paper/RK6BZ4IP

@misc{pith2026190801822,
  author       = {Pith},
  title        = {Pith review of: Blind Sparse Estimation of Intermittent Sources over Unknown Fading Channels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RK6BZ4IP}},
  note         = {Machine review of arXiv:1908.01822}
}
read the original abstract

Radio frequency sources are observed at a fusion center via sensor measurements made over slow flat-fading channels. The number of sources may be larger than the number of sensors, but their activity is sparse and intermittent with bursty transmission patterns. To account for this, sources are modeled as hidden Markov models with known or unknown parameters. The problem of blind source estimation in the absence of channel state information is tackled via a novel algorithm, consisting of a dictionary learning (DL) stage and a per-source stochastic filtering (PSF) stage. The two stages work in tandem, with the latter operating on the output produced by the former. Both stages are designed so as to account for the sparsity and memory of the sources. To this end, smooth LASSO is integrated with DL, while the forward-backward algorithm and Expectation Maximization (EM) algorithm are leveraged for PSF. It is shown that the proposed algorithm can enhance the detection and the estimation performance of the sources, and that it is robust to the sparsity level.

Figures

Figures reproduced from arXiv: 1908.01822 by the authors.

Figure 1
Figure 1. Wireless Sources transmit sporadically over a flat fading channel. The fusion [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Hidden Markov Model (HMM) for a source n. signals y(t), for t = 1, . . . , T, are collected at the fusion center. Based on the received signals Y, the goal of the fusion center is to detect the sources’ activity and to recover the signals x(t), for t = 1, . . . , T, or equivalently the matrix X, in the absence of information about the channel matrix H. For each source n, we define the activation pattern as a binary … view at source ↗
Figure 3
Figure 3. Block diagram of the proposed algorithm with Per-Source Filtering under the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Binary Asymmetric Channel (BAC) describing the relationship between the true [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Probability of detection Pd versus probability of false alarm Pfa for the considered algorithms (N = 30, M = 20, T = 1000, SNR= 30 dB, pn = 0.0022, qn = 0.02) [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Probability of detection Pd when the probability of false alarm is Pfa = 0.1 versus λ for the SL-ADMM algorithm (N = 30, M = 20, T = 1000, pn = 0.0022, qn = 0.02) [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Probability of detection Pd when the probability of false alarm is Pfa = 0.1 versus µ for the SL-ADMM algorithm (N = 30, M = 20, T = 1000, SNR= 30 dB [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Probability of detection Pd when the probability of false alarm is Pfa = 0.1 versus SNR for the considered algorithms (N = 30, M = 20, T = 1000, pn = 0.0022, qn = 0.02) [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: Probability of detection Pd when the probability of false alarm is Pfa = 0.1 versus average number of active sources (N = 30, M = 20, T = 1000, SNR= 30 dB, qn = 0.02) [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: EVM versus SNR (N = 30, M = 20, T = 1000, pn = 0.0022, qn = 0.02). the figure, the estimate is seen to be close to the real value. The parameters pn and qn in the HMM have similar estimation accuracy as p 0 n and q 0 n [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: Estimation performance of p 0 n and q 0 n for BAC in the EM algorithm (N = 30, M = 20, T = 1000, SNR= 30 dB, pn = 0.0022, qn = 0.02) [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: Probability of detection Pd versus probability of false alarm Pfa for the considered algorithms with EM algorithm (N = 30, M = 20, T = 1000, SNR= 30 dB, pn = 0.0022, qn = 0.02) [PITH_FULL_IMAGE:figures/full_fig_p013_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 47 canonical work pages

  1. [1]

    Comon and C

    P. Comon and C. Jutten, Handbook of Blind Source Separation: Independent Component analysis and Applications . Cambridge: Academic press, 2010

  2. [2]

    The fundamental limitation of frequency domain blind source separation for convolutive mixtures of speech,

    S. Araki, R. Mukai, S. Makino, T. Nishikawa, and H. Saruwatari, “The fundamental limitation of frequency domain blind source separation for convolutive mixtures of speech,” IEEE Transaction on Speech and Audio Processing , vol. 11, pp. 109–116, 2003

  3. [3]

    Blind separation and dereverberation of speech mixtures by joint optimization,

    T. Yoshioka, T. Nakatani, M. Miyoshi, and H. G. Okuno, “Blind separation and dereverberation of speech mixtures by joint optimization,” IEEE Trans. on Audio, Speech, and Language Processing , vol. 19, pp. 69–84, 2011. 13 Fig. 11: Estimation performance of p′ n and q′ n for BAC in the EM algorithm ( N = 30, M = 20, T = 1000, SNR= 30 dB, pn = 0.0022, qn = 0...

  4. [4]

    Cichocki and S.-i

    A. Cichocki and S.-i. Amari, Adaptive blind signal and image processing: learning algorithms and applications. Hoboken, NJ: John Wiley & Sons, 2002

  5. [5]

    Applications of neural blind separation to signal and image processing,

    J. Karhunen, A. Hyv ¨arinen, R. Vigario, J. Hurri, and E. Oja, “Applications of neural blind separation to signal and image processing,” Acoustics, Speech, and Signal Processing, 1997. ICASSP-97., 1997 IEEE International Coference on , vol. 5, pp. 131–134, 1997

  6. [6]

    Blind separation of disjoint orthogonal signals: demixing N sources from 2 mixtures,

    A. Jourjine, S. Rickard, and ¨O. Yilmaz, “Blind separation of disjoint orthogonal signals: demixing N sources from 2 mixtures,”Acoustics, Speech, and Signal Processing, 2000. ICASSP ’00. Proceedings. 2000 IEEE International Conference on, 2000

  7. [7]

    Synchronization of acoustic sensors for distributed ad-hoc audio networks and its use for blind source separation,

    S. Wehr, I. Kozintsev, R. Lienhart, and W. Kellermann, “Synchronization of acoustic sensors for distributed ad-hoc audio networks and its use for blind source separation,” Multimedia Software Engineering, 2004. Proceddings. IEEE Sixth International Symposium on , 2004

  8. [8]

    Independent component analysis: algorithms and applications,

    A. Hyv ¨arinen and E. Oja, “Independent component analysis: algorithms and applications,” Neural Networks, vol. 13, pp. 411–430, 2000

Show all 49 references
  1. [9]

    Principal component analysis,

    S. Wold, K. Esbensen, and P. Geladi, “Principal component analysis,” Chemometrics and Intelligent Laboratory Systems , vol. 2, pp. 37–52, 1987

  2. [10]

    Singular value decomposition and least squares solutions,

    G. H. Golub and C. H. Reinsch, “Singular value decomposition and least squares solutions,” Numerische Mathematik , vol. 14, pp. 403–420, 1970

  3. [11]

    Long-range communications in unlicensed bands: the rising stars in the IoT and smart city scenarios,

    M. Centenaro, L. Vangelista, A. Zanella, and M. Zorzi, “Long-range communications in unlicensed bands: the rising stars in the IoT and smart city scenarios,” IEEE Wireless Communications, vol. 23, pp. 60–67, 2016

  4. [12]

    New 3GPP standard for IoT,

    J. Gozalvez, “New 3GPP standard for IoT,” IEEE Vehicular Technology Magazine, vol. 11, pp. 14–20, 2016

  5. [13]

    Distributed linear blind source separation over wireless sensor networks with arbitrary connectivity patterns,

    S. M. Alavi and W. B. Kleijn, “Distributed linear blind source separation over wireless sensor networks with arbitrary connectivity patterns,” Acoustics, Speech and Signal Processing, IEEE Transactions on , pp. 3171–3175, March 2016

  6. [14]

    A blind source separation technique for spectrum sensing in cognitive radio networks based on kurtosis metric,

    S. S. Ivrigh, S. M.-S. Sadough, and A. A. Ghorashi, “A blind source separation technique for spectrum sensing in cognitive radio networks based on kurtosis metric,” Computer and Knowledge Engineering (ICCKE), 2011 1st International eConference on, Oct. 2011

  7. [15]

    Independent component analysis for semi-blind signal separation in MIMO mobile frequency selective communication channels,

    D. Obradovic, N. Madhu, A. Szabo, and C. S. Wong, “Independent component analysis for semi-blind signal separation in MIMO mobile frequency selective communication channels,” Neural Networks, 2004. Proceedings. 2004 IEEE International Joint Conference on , July 2004

  8. [16]

    Blind signal separation in MIMO OFDM system using ICA and fractional sampling,

    S. R. Curnew and J. Ilow, “Blind signal separation in MIMO OFDM system using ICA and fractional sampling,” Signals, Systems and Electronics, 2007. ISSSE ’07. International Symposium on , Aug. 2007

  9. [17]

    Hyv ¨arinen, J

    A. Hyv ¨arinen, J. Karhunen, and E. Oja, Independent Component Analysis , 3rd ed. West Sussex, United Kingdom: John Wiley & Sons, 2004

  10. [18]

    Elad, Sparse and Redundant Representations: From Theory to Applications in Signal and Image Processing

    M. Elad, Sparse and Redundant Representations: From Theory to Applications in Signal and Image Processing . New York: Springer, 2010

  11. [19]

    Sparse signal reconstruction from limited data using FOCUSS: a re-weighted minimum norm algorithm,

    I. F. Gorodnitsky and B. D. Rao, “Sparse signal reconstruction from limited data using FOCUSS: a re-weighted minimum norm algorithm,” International Journal of Pattern Recognition and Artificial Intelligence , vol. 45, pp. 600–616, 1997

  12. [20]

    Optimized orthogonal matching pursuit approach,

    L. Rebollo-Neira and D. Lowe, “Optimized orthogonal matching pursuit approach,” IEEE Signal Processing Letters , vol. 9, pp. 137–140, 2002

  13. [21]

    Dictionaries for sparse representation modeling,

    R. Rubinstein, A. M. Bruckstein, and M. Elad, “Dictionaries for sparse representation modeling,” Proceedings of the IEEE, vol. 98, pp. 1045–1057, 2010

  14. [22]

    Dictionary learning,

    I. Tosic and P. Frossard, “Dictionary learning,” IEEE Signal Processing Magazine , vol. 28, pp. 27–38, 2011

  15. [23]

    Joint DOA Estimation and Array Calibration Using Multiple Parametric Dictionary Learning,

    H. Ghanbari, H. Zayyani, and E. Yazdian, “Joint DOA Estimation and Array Calibration Using Multiple Parametric Dictionary Learning,” ArXiv e-prints, Jul. 2017

  16. [24]

    A Stochastic Successive Minimization Method for Nonsmooth Nonconvex Optimization with Applications to Transceiver Design in Wireless Communication Networks,

    M. Razaviyayn, M. Sanjabi, and Z.-Q. Luo, “A Stochastic Successive Minimization Method for Nonsmooth Nonconvex Optimization with Applications to Transceiver Design in Wireless Communication Networks,” ArXiv e-prints, Jul. 2013

  17. [25]

    Cloud K-SVD: A collaborative dictionary learning algorithm for big, distributed data,

    H. Raja and W. U. Bajwa, “Cloud K-SVD: A collaborative dictionary learning algorithm for big, distributed data,” IEEE Transcations on Signal Processing , vol. 64, pp. 173–188, 2015

  18. [26]

    Dictionary Learning Based Sparse Channel Representation and Estimation for FDD Massive MIMO Systems,

    Y . Ding and B. D. Rao, “Dictionary Learning Based Sparse Channel Representation and Estimation for FDD Massive MIMO Systems,” ArXiv e-prints, Dec. 2016

  19. [27]

    Independent vector analysis incorporating active and inactive states,

    A. Masnadi-Shirazi and B. Rao, “Independent vector analysis incorporating active and inactive states,” Acoustics, Speech and Signal Processing, 2009. ICASSP 2009. IEEE International Conference on , May 2009

  20. [28]

    Glimpsing independent vector analysis: Separating more sources than sensors using active and inactive states,

    A. Masnadi-Shirazi, W. Wang, and B. Rao, “Glimpsing independent vector analysis: Separating more sources than sensors using active and inactive states,” Acoustics, Speech and Signal Processing (ICASSP), 2010 IEEE International Conference on, June 2010

  21. [29]

    Synchronization for lassical blind source separation algorithms in wireless acoustic sensor networks,

    C. Llerena, R. Gil-Pita, D. Ayllon, H. Sanchez-Hevia, I. Mohino-Herranz, and M. Rosa, “Synchronization for lassical blind source separation algorithms in wireless acoustic sensor networks,” 2016 IEEE Statistical Signal Processing Workshop (SSP), June 2016

  22. [30]

    Blind synchronization in wireless acoustic sensor networks,

    D. Cherkassky and S. Gannot, “Blind synchronization in wireless acoustic sensor networks,” IEEE/ACM Trans. on Audio, 15 Speech, and Language Processing , vol. 25, pp. 651–661, 2017

  23. [31]

    Dictionary learning over distributed models,

    J. Chen, Z. J. Towfic, and A. H. Sayed, “Dictionary learning over distributed models,” IEEE Trans. on Signal Processing, vol. 63, pp. 1001–1016, 2015

  24. [32]

    Dictionary identification - sparse matrix-factorization via 𝓁1-minimization,

    R. Gribonval and K. Schnass, “Dictionary identification - sparse matrix-factorization via 𝓁1-minimization,” IEEE Trans. on Information Theory , vol. 56, pp. 3523–3539, 2010

  25. [33]

    Method of optimal directions for frame design,

    K. Engan, S. O. Aase, and J. Hakon Husoy, “Method of optimal directions for frame design,” Proc, IEEE ICASSP, vol. 5, pp. 2443–2446, 1999

  26. [34]

    Improving dictionary learning: Multiple dictionary updates and coefficient reuse,

    L. N. Smith and M. Elad, “Improving dictionary learning: Multiple dictionary updates and coefficient reuse,” IEEE Signal Process. Lett., vol. 20, no. 1, pp. 79–82, 2013

  27. [35]

    Dictionary training for sparse representation as generalization of k-means clustering,

    S. K. Sahoo and A. Makur, “Dictionary training for sparse representation as generalization of k-means clustering,” IEEE Signal Process. Lett. , vol. 20, no. 6, pp. 587–590, 2013

  28. [36]

    Dictionary learning for sparse representation: A novel approach,

    M. Sadeghi, M. Babaie-Zadeh, and C. Jutten, “Dictionary learning for sparse representation: A novel approach,” IEEE Signal Process. Lett. , vol. 20, no. 12, pp. 1195–1198, 2013

  29. [37]

    Regression shrinkage and selection via the lasso,

    R. Tibshirani, “Regression shrinkage and selection via the lasso,” Journal of the Royal Statistical Society , vol. 58, no. 1, pp. 267–288, 1996

  30. [38]

    Convolutional dictionary learning,

    C. Garcia-Cardona and B. Wohlberg, “Convolutional dictionary learning,” CoRR, vol. abs/1709.02893, 2017. [Online]. Available: http://arxiv.org/abs/1709.02893

  31. [39]

    The smooth-lasso and other 𝓁1 +𝓁2-penalized methods,

    M. Hebiri and S. van de Geer, “The smooth-lasso and other 𝓁1 +𝓁2-penalized methods,” Electronic Journal of Statistics , vol. 5, pp. 1184–1226, 2011

  32. [40]

    An interior-point method for large-scale 𝓁1-regularized logistic regression,

    K. Koh, S. J. Kim, and S. Boyd, “An interior-point method for large-scale 𝓁1-regularized logistic regression,” Journal of Machine Learning Research , vol. 8, pp. 1519–1555, 2007

  33. [41]

    Regularization and variable selection via the elastic net,

    H. Zou and T. Hastie, “Regularization and variable selection via the elastic net,” Journal of the Royal Statistical Society , vol. 67, pp. 301–320, 2005

  34. [42]

    Distributed optimization and statistical learning via the alternating direction method of multipliers,

    S. Boyd, N. Parikh, E. Chu, B. Peleato, and J. Eckstein, “Distributed optimization and statistical learning via the alternating direction method of multipliers,” Foundations and Trends in Machine Learning , vol. 3, pp. 1–122, 2011

  35. [43]

    Proximal algorithms,

    N. Parikh and S. Boyd, “Proximal algorithms,” Foundations and Trends in Optimization , vol. 1, pp. 123–231, 2014

  36. [44]

    Multiplier and gradient methods,

    M. R.Hestenes, “Multiplier and gradient methods,” Journal of Optimization Theory and Applications , vol. 4, pp. 302–320, 1969

  37. [45]

    Koller and N

    D. Koller and N. Friedman, Probabilistic Graphical Models: Principles and Techniques . Cambridge, Massachusetts: MIT Press, 2009

  38. [46]

    K. P. Murphy, Machine Learning: A Probabilistic Perspective . Cambridge, Massachusetts: MIT Press, 2012

  39. [47]

    C. M. Bishop, Pattern Recognition and Machine Learning . New York, NY: Springer, 2006

  40. [48]

    An introduction to hidden markov models and bayesian networks,

    Z. Ghahramani, “An introduction to hidden markov models and bayesian networks,” International Journal of Pattern Recognition and Artificial Intelligence , vol. 15, pp. 9–42, 2001

  41. [49]

    Generalized elastic net regression,

    G. Mouret, J.-J. Brault, and V . Partovinia, “Generalized elastic net regression,” the Joint Statistical Meetings , Aug. 2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.