Due to some datasets may acquire a large amount of information to train the annotation model, but some datasets may lack sufficient information, multi-label transfer learning has attracted a lot of attention recently. The correlation between inherent class label among multiple labels is integrated into data structure and can cover the relationship of samples. However, it is usually not accurate enough. In order to use their advantages and eliminate their weaknesses, this paper presents a new framework which integrates the techniques of Green's function inspired label correlation and label covariance analysis. Meanwhile, the framework adopts low-rank based method to learn the groups clearly. At the same time, due to the edge of annotation system be very difficult to determine, our framework is designed to carry out the appropriate hyperplane margin for multi-label datasets. In this framework, label covariance is first utilized by perceptron criterion analyzer, then linear discriminant function can dispose the unlabeled part of multi-label data easily. In addition, a low-rank based margin mechanism is presented to optimize the performance of our framework. We evaluate our framework in the performance of label correlation and covariance builders for transfer annotation. The derived results are encouraging and we achieve about 5% improvement on average accuracy.

Decoding sparse quantum codes can be accomplished by syndrome-based decoding using a belief propagation (BP) algorithm. We significantly improve this decoding scheme by developing a new feedback adjustment strategy for the standard BP algorithm. In our feedback procedure, we exploit much of the information from stabilizers, not just the syndrome but also the values of the frustrated checks on individual qubits of the code and the channel model. Furthermore we show that our decoding algorithm is superior to belief propagation algorithms using only the syndrome in the feedback procedure for all cases of the depolarizing channel. Our algorithm does not increase the measurement overhead compared to the previous method, as the extra information comes for free from the requisite stabilizer measurements.

In this paper, we first propose a general framework for constructing quasi-cyclic low-density parity-check (QC-LDPC) codes based on a two-dimensional (2-D) maximum distance separable (MDS) code. Two classes of QC-LDPC codes are defined, whose parity-check matrices are transposes of each other. We then use a 2-D generalized Reed-Solomon (GRS) code to give a concrete construction. The decoding parity-check matrices have a large number of redundant parity-check equations while their Tanner graphs have a girth of at least 6. The minimum distances of the codes are very respectable as far as LDPC codes are concerned. We further show that many existing constructions of QC-LDPC codes based on finite fields in the literature can be unified under this construction. Experimental studies show that the constructed QC-LDPC codes perform well with the sum-product algorithm (SPA).

This paper presents an algebraic method for constructing quasi-cyclic (QC) low-density parity-check (LDPC) codes based on the structural properties of Euclidean geometries. The construction method results in a class of QC-LDPC codes with girth of at least 6. Codes in this class perform very close to the Shannon limit with iterative decoding. Simulations show that the designed QC-LDPC codes have almost the same performance with the existing QC Euclidean geometry LDPC codes.

We explicitly construct an infinite family of asymptotically good concatenated quantum stabilizer codes where the outer codes use quantum Reed-Solomon codes and the inner codes use a set of special quantum codes. In the field of quantum error-correcting codes, this is the first time that a family of asymptotically good quantum codes is derived from bad codes. Its occurrence supplies a gap in quantum coding theory.

To reduce the detection complexity of the space-time bit-interleaved coded modulation(ST-BICM) scheme, a turbo type iterative receiver structure consisting of a new group Gaussian approximation(GGA) algorithm and an a posteriori probability(APP) decoder is presented. When calculating the log-likelihood ratio(LLR) of the i-th transmitted symbol, the GGA algorithm first divide the N transmitted signals into two groups, a detecting group consisting of G signals with the i-th transmitted signal included, and an interfering group consisting of the other N-G signals. Taking the superposition of the N-G signals as a Gaussian random variable, the GGA algorithm implements the maximum a posteriori(MAP) detection to obtain the LLR of the signal transmitted from antenna i. At high SNR values or if the receiver has more than two replicas of each transmitted signal, the GGA detection exhibits a good performance as well as low decoding complexity.

Concatenated twist Hadamard code (CTwHQ is a new class of low-rate error-correcting codes designed based on two single parity check constraints in bi-orthogonal Hadamard codes, which can present good error floor performance with low decoding complexity. To analyze CTwHC from a theoretical perspective, a simple recursive technique based on trellis structure is proposed to achieve union bound of CTwHC in this paper.

A feedback belief propagation (BP) decoding algorithm for low-density parity-check convolutional codes is proposed. The proposed algorithm can activate the variable nodes more efficiently by applying feedback decoding at each decoding iteration. Compared with the on-demand BP algorithm, the proposed algorithm has a doubled convergence speed and causes only about half of the decoding delay at similar error performances without any increase of storage requirement. Simulation results show that the proposed algorithm can offer a good trade-off between the error-correcting performance and the decoding complexity.

This paper is concerned with the transmission strategies and relay selection for decode-and-forward parallel relay networks (PRNs) based on superposition coding. Our focus is on the minimization of the aggregate transmission power that supports a given transmission rate. In this paper, two transmission strategies are investigated, i.e., an optimal- cooperation (OC) strategy requiring strict synchronization at the destination/relay nodes and a no-information-overlapping (NIO) strategy with no synchronization requirement for the ease of implementation using current technologies such as interleave- division multiple-access (IDMA). Furthermore, a reduced-relay-selection (RRS) strategy involving only 2-level superposition coding is given to trade off between complexity and performance. Numerical results show that the performance gap between OC and NIO is negligible, and selecting the best two relay nodes is adequate to achieve most performance gain even when the number of relay nodes is large.

By proving the existence of the minimum of the threshold function in (0,1) we show that the threshold of regular low-density erasure codes under erasure recovery algorithms is equal to the minimum of the threshold function in (0, 1). The existence and uniqueness of positive real root of the polynomial equation k(x) = 0 in (0,1) are verified. It follows from the two conclusions obtained above that the threshold of regular low-density erasure codes can be determined by the analytical expression.

We present a new method to forge ElGamal signatures with the cases that the secret key parameters of the system are not known under the chosen signature messages. The attacker can forge the signature substituting the right signature, and also attack the right secret key without depending on the computation of discrete logarithm. With the attacking probability cryptanalysis, it is found that the cryptosystem can be attacked successfully in some conditions.

We propose an identity (ID)-based signature from m-torsion groups of super-singular elliptic curves or hyper-elliptic curves. It is secure against existential forgery under adoptively chosen message and ID attack in the random oracle model. Due to the nice properties of m-torsion groups and Weil pairing, it turns out that the proposed scheme is simple, efficient and has the good property of addition. Based on this ID-based signature scheme, we propose an ID-based threshold signature scheme and show that it has the security properties of unforgeability and robustness.

It is well known that the algebraic expression of the advanced encryption standard (AES) Rijndael S-box is very simple and only 9 terms are involved. In this paper a new algebraic method is proposed to illustrate the reason why AES S-box algebraic expression is very simple, which corrects all the available inaccurate illustrations. An improved AES S-box is also presented to improve the complexity of AES S-box algebraic expression with terms increasing from 9 to 255 and algebraic degree invariable. The improved AES S-box also has good properties of Boolean functions in SAC and balance, and is capable of attacking against differential cryptanalysis with high reliable security.

Using high rate recursive systematic convolutional code (RSCC) as the basic element and the trace criteria as the design principle, a new kind of high rate recursive space-time trellis code (HR-RSTTC) is designed for serial concatenation. By making the number of RSCCs and the coding rate of each RSCC be variable, our designed HR-RSTTC can adjust the data rate according to the transmit antenna number and the modulation scheme. It is verified that, in independent fading channels, the minimum diversity gains of the serially concatenated space-time trellis code (SCSTTC) constructed with HR-RSTTC is 2M(M is the number of receive antennas). In comparison with the traditional SCSTTC, the new SCSTTC can achieve the same data rate with lower complexity and no performance degradation. With the same transmit antenna number and modulation scheme, the new SCSTTC have higher data rate than the traditional SCSTTC

A new class of recursive convolutional space-time codes (ReC-STC) with an arbitrary number of transmit antennas is designed by adopting several parallel two-state recursive systematic convolutional codes (RSCs). An intercross linear mapping rule that distributes the output bits of an RSC to the different positions of different transmitted signals is also described. The proposed ReC-STC can not only increase the data rate with the number of transmit antennas, but also performs well when it is used in a serially concatenated space-time code (SCSTC). The convergence analysis shows that ReC-STC based SCSTC has a lower decoding threshold than the available SCSTC.

The sensitivities of the bits of low weight codewords in turbo codes to noises are discussed firstly and the result is that the lower weight the higher sensitivity. Theoretical analysis shows that through the protections on such key-bits which have high sensitivities; the minimal weight of the codes can be improved apparently. The simulations also give out the supported results: The BER could be highly improved at high Eb/N0 for code rates both 1/3 and 1/2.

The rate design rules for the multilevel coding (MLC) are studied over the Rayleigh fading channel. By simulations on 16-QAM and Ungerboeck mapping, it is speculated that the coding exponent rule is yet the best for finite length. The MLC scheme using low-density parity-check (LDPC) codes performs within 2dB above the capacity limit at the bit error rate 10-5 for bandwidth efficiency 2-bits/symbol in the fully interleaved channel with channel state information. LDPC schemes suffer less signal-to-noise ratio penalty in correlated channels.

There have not been many successful attacks against the AES algorithm Rijndael so far, of which the best known and effective attack is the Square Attack. It takes advantage of the balance property of the algorithm. A new type of attack which makes use of the activity property is proposed. The consuming time and chosen plaintexts necessary are both reduced to one tenth of the original ones. The analysis principle is described.

This paper presents a family of concatenated two-state trellis-coded modulation (CT-TCM) schemes. Compared with the existing turbo-type bandwidth-efficient coded modulation schemes, the proposed codes have significantly reduced complexity without sacrificing performance. A joint design strategy for all component codes is established. This leads to so-called asymmetrical and time-varying trellis structures, which possess good Hamming and Euclidean distance distributions. The performance of the proposed codes is demonstrated by simulation results.

The general autocorrelation proposed is the extension of the traditional autocorrelation. Besides having the same ability as autocorrelation to reveal the characteristics of the linear structures, the strict avalanche criterion (SAC) and the propagation criterion (PC) of degree k, the general autocorrelation is capable of finding the order of SAC and PC which is a capability that autocorrelation does not possess. The application on AES S-box shows the SAC and PC characteristics of the eight AES Boolean functions.

Another description of the two-dimensional binary (d,k) constrained array is given. Based on this, a lower bound for the capacity of the (0,k) constrained array is presented, and it is reproved that the capacity of the (d,d+1) constrained array is equal to zero.

Some properties of the minimal interpolation problem are investigated, from which a simple proof of the validity of the Welch-Berlekamp (1983) algorithm is presented. A new key equation is derived, which is closely related to the classical key equation of the syndrome-decoding algorithm and can be solved by the Welch-Berlekamp algorithm.

In this correspondence, we derive, with the method of coset decompositions, some new bounds for q-ary linear block codes, where q is the order of some finite field.

The undetected error probability (UEP) of binary (n, 2/spl delta/, m) nonlinear constant weight codes over the binary symmetric channel (BSC) is investigated, where n is the blocklength, m is the weight of codeword and 2/spl delta/ is the minimum distance of the codes. The distance distribution of the (n, 2, m) nonlinear constant weight codes is evaluated. It is proven in this paper that the (5, 2, 2) code, (5, 2, 3) code, (6, 2, 3) code, (7, 2, 4) code, (7, 2, 3) code and (8, 2, 4) code are the only proper error-detecting codes in the (n, 2, m) nonlinear constant weight codes for n/spl ges/5, in the sense that their UEP is increased monotonically with the channel error rate p, of course all these proper codes are m-out-of-n codes. Furthermore, it is conjectured that except for the cases of n/spl les/4/spl delta/, there are no proper error-detecting binary (n, 2/spl delta/, m) nonlinear constant weight codes, for n>8 and /spl deltaspl ges/1.>

It is shown that McEliece's and Niederreiter's public-key cryptosystems are equivalent when set up for corresponding choices of parameters. A security analysis for the two systems based on this equivalence observation, is presented.>