Serdar Boztas

Prof. Serdar Boztas

orcid

Research Focus & Metrics

Academic Career Length

35Years (1992-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 May 1992 in IEEE Transactions on Information Theory
Verified
Journal Article
Closed access
SBSerdar Boztas
Serdar Boztas
RHR. Hammons
R. Hammons
PKP.Y. Kumar
P.Y. Kumar
Two families of four-phase sequences are constructed using irreducible polynomials over Z/sub 4/. Family A has period L=2/sup r/-1. size L+2. and maximum nontrivial correlation magnitude C/sub max/<or=1+ square root (L+1), where r is a positive integer. Family B has period L=2(2/sup r/-1). size (L+2)/4. and C/sub max/ for complex-valued sequences. Of particular interest, family A has the same size and period as the family of binary Gold sequences. but its maximum nontrivial correlation is smaller by a factor of square root 2. Since the Gold family for r odd is optimal with respect to the Welch bound restricted to binary sequences, family A is thus superior to the best possible binary design of the same family size. Unlike the Gold design, families A and B are asymptotically optimal whether r is odd or even. Both families are suitable for achieving code-division multiple-access and are easily, implemented using shift registers. The exact distribution of correlation values is given for both families.<>

26 October 2022 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
KDKemal Davaslioglu
Kemal Davaslioglu
SBSerdar Boztas
Serdar Boztas
MEMehmet Can Ertem
Mehmet Can Ertem
YSYalin E. Sagduyu
Yalin E. Sagduyu
EAEnder Ayanoglu
Ender Ayanoglu
Deep learning (DL) finds rich applications in the wireless domain to improve spectrum awareness. Typically, DL models are either randomly initialized following a statistical distribution or pretrained on tasks from other domains in the form of transfer learning without accounting for the unique characteristics of wireless signals. Self-supervised learning (SSL) enables the learning of useful representations from Radio Frequency (RF) signals themselves even when only limited training data samples with labels are available. We present a self-supervised RF signal representation learning method and apply it to the automatic modulation recognition (AMR) task by specifically formulating a set of transformations to capture the wireless signal characteristics. We show that the sample efficiency (the number of labeled samples needed to achieve a certain performance) of AMR can be significantly increased (almost an order of magnitude) by learning signal representations with SSL. This translates to substantial time and cost savings. Furthermore, SSL increases the model accuracy compared to the state-of-the-art DL methods and maintains high accuracy when limited training data is available.

19 June 2013 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
ZLZilong Liu
Zilong Liu
Udaya ParampalliUdaya Parampalli
Udaya Parampalli
YGYong Liang Guan
Yong Liang Guan
SBSerdar Boztas
Serdar Boztas
Compared with the perfect complementary sequence sets, quasi-complementary sequence sets (QCSSs) have the advantage of supporting more users in multicarrier CDMA communications. Constructions for optimal and near-optimal periodic QCSSs are proposed in this paper by using the Singer difference sets and the existing optimal quaternary sequence sets. The maximum periodic correlation magnitude of the proposed optimal QCSS achieves the derived periodic correlation lower bound asymptotically. To the authors' best knowledge, such optimal QCSSs haven't been reported before.

1 November 1997 in IEEE Transactions on Information Theory
Verified
Journal Article
Closed access
SBSerdar Boztas
Serdar Boztas
In the above paper by E. Arikan (see ibid., vol.42, no.1, p.99-105, 1996) an asymptotically tight upper bound on the /spl rho/th moment (/spl rho//spl ges/0) of the minimal number of guesses required to determine the value of a random variable was derived. We show that we can tighten this bound for the case of positive integer moments (when /spl rho/=1, the bound is improved by a factor of 2) and that the new bound also applies to a class of nonminimal guessing sequences.

1 January 2010
Verified
Book Chapter
Closed access
PDF file
LSLi Sun
Li Sun
SVSteven Versteeg
Steven Versteeg
SBSerdar Boztas
Serdar Boztas
TYTrevor Yann
Trevor Yann

1 August 2017
Verified
Conference Paper
Closed access
AGAnagi Gamachchi
Anagi Gamachchi
SBSerdar Boztas
Serdar Boztas
While most organizations continue to invest in traditional network defences, a formidable security challenge has been brewing within their own boundaries. Malicious insiders with privileged access in the guise of a trusted source have carried out many attacks causing far reaching damage to financial stability, national security and brand reputation for both public and private sector organizations. Growing exposure and impact of the whistleblower community and concerns about job security with changing organizational dynamics has further aggravated this situation. The unpredictability of malicious attackers, as well as the complexity of malicious actions, necessitates the careful analysis of network, system and user parameters correlated with insider threat problem. Thus it creates a high dimensional, heterogeneous data analysis problem in isolating suspicious users. This research work proposes an insider threat detection framework, which utilizes the attributed graph clustering techniques and outlier ranking mechanism for enterprise users. Empirical results also confirm the effectiveness of the method by achieving the best area under curve value of 0.7648 for the receiver operating characteristic curve.