scholarly journals Large Deviations Behavior of the Logarithmic Error Probability of Random Codes

2020 ◽  
Vol 66 (11) ◽  
pp. 6635-6659
Author(s):  
Ran Tamir ◽  
Neri Merhav ◽  
Nir Weinberger ◽  
Albert Guillen i Fabregas
Keyword(s):  
Large Deviations ◽  
Error Probability ◽  
Random Codes

scholarly journals Deep Polarized Network for Supervised Learning of Accurate Binary Hashing Codes

2020 ◽  
Author(s):  
Lixin Fan ◽  
Kam Woh Ng ◽  
Ce Ju ◽  
Tianyu Zhang ◽  
Chee Seng Chan
Keyword(s):  
Large Deviations ◽  
Hamming Distance ◽  
State Of The Art ◽  
Metric Learning ◽  
Binary Codes ◽  
Distance Metric Learning ◽  
Learning Framework ◽  
Label Information ◽  
Learning To Hash ◽  
Random Codes

This paper proposes a novel deep polarized network (DPN) for learning to hash, in which each channel in the network outputs is pushed far away from zero by employing a differentiable bit-wise hinge-like loss which is dubbed as polarization loss. Reformulated within a generic Hamming Distance Metric Learning framework [Norouzi et al., 2012], the proposed polarization loss bypasses the requirement to prepare pairwise labels for (dis-)similar items and, yet, the proposed loss strictly bounds from above the pairwise Hamming Distance based losses. The intrinsic connection between pairwise and pointwise label information, as disclosed in this paper, brings about the following methodological improvements: (a) we may directly employ the proposed differentiable polarization loss with no large deviations incurred from the target Hamming distance based loss; and (b) the subtask of assigning binary codes becomes extremely simple --- even random codes assigned to each class suffice to result in state-of-the-art performances, as demonstrated in CIFAR10, NUS-WIDE and ImageNet100 datasets.

Bounds on Data Compression Ratio with a Given Tolerable Error Probability

1998 ◽  
Vol 12 (2) ◽  
pp. 189-210
Author(s):  
Ilan Sadeh
Keyword(s):  
Probability Theory ◽  
Numerical Methods ◽  
Large Deviations ◽  
Data Compression ◽  
Compression Ratio ◽  
Error Probability ◽  
Compression Rate ◽  
Lossy Coding ◽  
Minimal Rate

The paper treats data compression from the viewpoint of probability theory where a certain error probability is tolerable. We obtain bounds for the minimal rate given an error probability for blockcoding of general stationary ergodic sources. An application of the theory of large deviations provides numerical methods to compute for memoryless sources, the minimal compression rate given a tolerable error probability. Interesting connections between Cramer's functions and Shannon's theory for lossy coding are found.

Error Probability of MRC in Frequency Selective Nakagami Fading in the Presence of CCI and ACI

2009 ◽  
Vol E92-A (11) ◽  
pp. 2679-2687
Author(s):  
Mohammad Azizur RAHMAN ◽  
Chin-Sean SUM ◽  
Ryuhei FUNADA ◽  
Shigenobu SASAKI ◽  
Tuncer BAYKAS ◽  
...  
Keyword(s):  
Error Probability ◽  
Nakagami Fading ◽  
Frequency Selective
Sign in / Sign up

Export Citation Format

Share Document