5/4 Martin Wainwright (UC Berkeley) (joint with CSPL seminar), Tree-reweighted message-passing algorithms in graphical models: Some theory and applications

5/12 Pierre Moulin (UIUC) (CSPL seminar, note the unusual day), Communication games with side information

Fall 2004

12/15 Prof. Prakash Narayan (UMD), Throughput-Delay Trade-off in Wireless Networks
References:
A. El Gamal, J. Mammen, B. Prabhakar and D. Shah, Throughput-Delay Trade-off in Wireless Networks, in IEEE INFOCOM, Hong Kong, 2004.(Best Paper Award)

12/08 Alkan Soysal (UMD), Network Coding
References:
a. R. Ahlswede, N. Cai, S.-Y. R. Li and R. W. Yeung, "Network information flow," IEEE Trans. on Information Theory, vol. 46, pp. 1204-1216, 2000.
b. S.-Y. R. Li, R. W. Yeung, and N. Cai. "Linear network coding". IEEE Transactions on Information Theory , Februray, 2003.

11/17 Prof. Min Wu and Shan He (UMD), Joint Coding and Embedding Techniques for Multimedia Fingerprinting
Abstract:
Digital fingerprinting is an emerging technology to protect multimedia content from illegal redistribution by uniquely marking copies of the content distributed to each user. Collusion is a powerful attack whereby several differently fingerprinted copies of the same content are combined together to attenuate or even remove the embedded fingerprints. One of the main categories of collusion resistant fingerprinting techniques is coded fingerprinting, which uses compact codes to represent users. Many existing works in this area focus on the code layer and usually do not evaluate the overall performance with the embedding layer.

In this talk, we first review the techniques of robust data embedding, which serves as a basis for putting digital fingerprints (both non-coded and coded) into multimedia signals. From multimedia signal processing perspective, assigning and embedding mutually orthogonal signals as fingerprints to represent each of the users is a natural way to perform fingerprinting. Using this well-studied orthogonal fingerprinting as a baseline, we examine the overall performance of the error correcting code (ECC) based fingerprinting, including both the coding and embedding layers. The results suggest that when compared to the non-coded orthogonal approach for the same number of users and the same host signal length, the ECC fingerprinting has computational advantage and supports efficient distribution, but has in general a lower collusion resistance by one order of magnitude. We also observe very different resistance of ECC fingerprints against averaging and interleaving types of collusion. These observations have motivated us to propose two new techniques to improve the collusion resistance of coded fingerprinting while retaining its advantages. The improvement is benefited from the joint consideration of the coding and embedding layers, and is validated by the experimental results on both synthetic data and real images.

11/03 Onur Kaya and Nan Liu (UMD), Diversity-Multiplexing Tradeoff (Contd...)

10/27 Onur Kaya and Nan Liu (UMD), Diversity-Multiplexing Tradeoff
References:
[1] L. Zheng and D. Tse, "Diversity and Multiplexing: A Fundamental Tradeoff in Multiple Antenna Channels'', IEEE Transactions on Information Theory, vol. 49(5), May 2003.2
[2] D. Tse, P. Viswanath and L. Zheng, "Diversity-Multiplexing Tradeoff in Multiple Access Channels" , to appear in the Transactions on Information Theory. (Revised: March 2004).
[3] H. El Gamal, G. Caire, M. Oussama Damen, "Lattice Coding and Decoding Achieve the Optimal Diversity-Multiplexing Tradeoff of MIMO Channels", IEEE Transactions on Information Theory 50(6): 968-985 (2004)

10/20 J. Justesen (TU Denmark), Decoding more errors in Reed Solomon codes

Abstract: Standard decoding algorithms for RS codes usually lead to decoding failure when more than D/2 errors occur, while decoding errors are very rare. When RS codes are used in various forms of concatenated codes, it is desirable to correct more errors, and often additional information is available from the inner code. In the talk I will consider some recent advances in RS decoding and their potential for improving decoder performance.

10/12 No seminar

10/5 N. Prasanth Anthapadmanabhan (UMD)Optimal Probabilistic Fingerprint Codes

9/28 No seminar due to Allerton conference.

9/21 Chunxuan Ye (UMD) The private key capacity region for three terminals (joint work with P. Narayan)

9/14 Kaushik Chakraborty (UMD), Capacity of the Optical Fading Channel with Channel State Information

Spring 2004

3/24 A. Barg (UMD), Codes from bipartite graphs

3/31 A. Barg (UMD), Bounding the error probability via the distance distribution of codes

4/7 No seminar

4/14 P. Narayan (UMD), Secret key generation and Slepian-Wolf data compression

4/21 Damianos Karakos (JHU), EM algorithm and its connection with alternating minimization technique of Csiszar and Tusnady

4/28 N. Prasanth Anthapadmanabhan and Kaushik Chakraborthy (UMD), Mutual Information and MMSE in Gaussian Channels (Guo-Shamai-Verdu '04)

5/5 Vinay Chande (UMD), Exponential Error bounds for Erasure, List and Decision Feedback Schemes (Forney '68)