Hong Kong Information Theory PostGraduate Day
Saturday, November 26, 2005

Sponsored by the
IEEE Hong Kong Information Theory Chapter


Session 1 HKUST 7th Floor Council Chamber    
10:30AM -- 11:00AM
Complex Lattice Reduction Algorithm for Low-Complexity MIMO Detection
  GAN Ying Hung
11:00AM -- 11:30AM
Identity Based Group Signatures
  Yuen Tsz Hon
11:30AM -- 12:00PM
Superposition Coding with Peak-Power Limitation
  Tong Jun
Lunch HKUST G/F Chinese restaurant  
12:00 PM - 1:30 PM    
Session 2 HKUST 7th Floor Council Chamber    
1:30PM -- 2:00PM
Turbo Equalization Based on Vector Factor Graphs
  Qinghua Guo
2:00PM -- 2:30PM
The Quadrangle-Inequality Dynamic-Programming Speedup is a Consequence of Totally Monotonicity
2:30PM -- 3:00PM

On the Discontinuity of the Shannon Information Measures

  Ho Siu Wai


All talks will be given in the 7th floor council chamber at the Hong Kong University of Science and Technology. For information on how to get to HKUST please see this.  Once at HKUST take lifts 13-15 to the 7th floor.  Upon exiting the lifts make  a right and you will be at the  council chamber.

Confirmed Attendees


Name University Status Email
GAN Ying Hung HKUST (EEE) PhD eegyh@ust.hk
CHAN Ho Yin, Herbert HKUST(EE) PhD eechan@ust.hk
AU Kwok Shum, Edward HKUST(EE) PhD eeedward@ust.hk
CHENG Sin Ying HKUST(CS) MPhil vivying@cs.ust.hk
ZHANG Yan HKUST(CS) PhD cszy@cs.ust.hk
ZHEN Zhou HKUST(CS) PhD cszz@cs.ust.hk
WANG Yajun HKUST(CS) PhD yalding@cs.ust.hk
ZHU Wenqi HKUST(CS) PhD wqzhu@cs.ust.hk
HUANG  Qiong HKUST(CS) PhD bestivy@cs.ust.hk
XIA Jian HKUST(CS) PhD piper@cs.ust.hk
Gerhard Trippen HKUST(CS) PhD trippen@cs.ust.hk
Mordecai GOLIN HKUST(CS) Staff golin@cs.ust.hk
MOW Wai Ho HKUST(EE) Staff eewhmow@ust.hk
DING Cunsheng HKUST(CS) Staff csding@cs.ust.hk
TONG Jun CityU(EE) PhD jun.tong@student.cityu.edu.hk
Qinghua Guo CityU(EE) PhD qh.guo@student.cityu.edu.hk
Peng Wang CityU(EE) PhD pengwang@cityu.edu.hk
Yuan Xiaojun CityU(EE) PhD xjyuan@cityu.edu.hk
LI Yueqian CityU(EE) MPhil yueqian.li@student.cityu.edu.hk
Shuling Che CityU(EE) Staff shlche@cityu.edu.hk
Wu Hao CityU(EE) MPhil Jason.Wu@student.cityu.edu.hk
Ho Siu Wai CUHK(IE) PhD swho4@ie.cuhk.edu.hk
YANG Shenghao CUHK(IE) PhD shyang5@ie.cuhk.edu.hk
Tsz Hon YUEN CUHK(IE) MPhil thyuen4@ie.cuhk.edu.hk
KWOK, Pui Wing CUHK(IE) MPhil pwkwok4@ie.cuhk.edu.hk
Ngai Chi Kin CUHK(IE) PhD ckngai2@ie.cuhk.edu.hk
Fong Lik Hang Silas CUHK(IE) MPhil lhfong5@ie.cuhk.edu.hk
Ho Siu Ting CUHK(IE) PhD stho3@ie.cuhk.edu.hk
Shen Yuxiu CUHK(IE) MPhil yxshen5@ie.cuhk.edu.hk
Sun Qifu CUHK(IE) MPhil qfsun5@ie.cuhk.edu.hk
Yuen Pak Ho CUHK(IE) MPhil phyuen@cuhk.edu.hk
Raymond Yeung CUHK(IE) Staff whyeung@ie.cuhk.edu.hk



Registration will be $70HKper student and $100HK per non-student to be paid on site.   The registration fee includes lunch



Title: Complex Lattice Reduction Algorithm for Low-Complexity MIMO Detection
Speaker: GAN Ying Hung

 Recently, lattice-reduction-aided detectors have been proposed for multiple-input multiple-output (MIMO) communication systems to give performance with full diversity like maximum likelihood optimal receiver yet with complexity similar to linear receiver. However, these lattice-reduction-aided detectors are based on the traditional LLL reduction algorithm that was originally introduced for reducing real lattice basis, even though the channel matrices are inherently complex-valued.

In this talk, we will introduce the complex LLL algorithm for direct application to the channel matrix which naturally defines the basis of a complex lattice. Simulation results reveal that the new complex LLL algorithm can achieve a saving in complexity of nearly 50% over the traditional LLL algorithm, when applied to MIMO detection. In addition, we shall present a novel technique called the "joint basis labeling and reduction" which, by exploiting the degree of freedom of labeling the basis vectors, can further accelerate the LLL reduction algorithm. It is noteworthy that the complex LLL algorithms aforementioned incur negligible bit-error-rate performance loss relative to the traditional LLL algorithm

Title: Identity Based Group Signatures
Speaker: Yuen Tsz Hon

We present the first group signature scheme with provable security and short signature size where the group manager, the group members, and the Open Authority are all identity-based. We use the security model of Bellare, Shi, and Zhang, except to add three identity managers for manager, members, and OA respectively, and we discard the Open Oracle. Our construction uses identity-based signatures summarized in Bellare, Namprempre, and Neven for manager, Boneh and Franklin's IBE for OA, and we extend Bellare et al.'s group signature construction by verifiably encrypt an image of the member public key, instead of the public key itself.

Title: Superposition Coding with Peak-Power Limitation
Speaker: Tong Jun

In this work, we apply clipping to superposition coding systems to reduce the peak-to-average power ratio (PAPR) of the transmitted signal. The performance limit is investigated through evaluating the mutual information driven by the induced peak-power-limited input signals. It is shown that the channel capacity can be approached by clipped superposition coding systems. To alleviate the performance degradation due to clipping noises, we develop a soft compensation algorithm that is combined with soft-input-soft-output (SISO) decoding algorithms in an iterative manner. Simulation results show that with the proposed algorithm, most performance loss can be recovered.


Title: Turbo Equalization Based on Vector Factor Graphs
Speaker: Qinghua Guo

A factor graph approach to turbo equalization is proposed. Unlike the existing linear MMSE turbo equalization methods, which operate with truncated windows (sliding or extending window), the proposed is a full-window approach with low complexity. This approach supports a high-speed parallel implementation technique, which makes it an attractive option in practice.


Title: The Quadrangle-Inequality Dynamic-Programming Speedup is a Consequence of Totally Monotonicity
Speaker: ZHANG Yan

There exist several general techniques in the literature for speeding up naive implementations of dynamic programming. Two of the best known are the Knuth-Yao quadrangle inequality speedup and the SMAWK algorithm for finding the row-minima of totally monotone matrices. Although both of these techniques use a quadrangle inequality and seem similar they are actually quite different and have been used differently in the literature.

In this talk we show that the Knuth-Yao technique is actually a direct consequence of total monotonicity. As well as providing new derivations of the Knuth-Yao result, this also permits showing how to solve the Knuth-Yao problem directly using the SMAWK algorithm. Another consequence of this approach is a method for solving online versions of problems with the Knuth-Yao property. The online algorithms given here are symptotically as fast as the best previously known static ones.

This is joint work with Wolf Bein, Mordecai Golin and Larry Larmore

Title: On the Discontinuity of the Shannon Information Measures
Speaker:  Ho Siu Wai

It is well known that the Shannon information measures are continuous functions of the probability distribution when the support is finite. This, however, does not hold when the support is countably infinite. In this work, we investigate the continuity of the Shannon information measures for countably infinite support. With respect to some commonly used divergence measures including the Kullback-Liebler divergence and the variational distance, we use two different approaches to show that all the Shannon information measures are in fact discontinuous at all probability distributions with countably infinite support. For probability distributions with finite alphabet, some bounds are given to relate their support size and the difference of their entropy. These bounds show the limitation of certain algorithms of entropy estimation.


Page maintained by Mordecai Golin.
Last updated
11/25/2005 02:41 PM +0800