Turbo codes are the most recent breakthrough in coding theory. However, the decoder’s implementation cost limits their incorporation in commercial systems. Although the decoding algorithm is highly data dominated, no true memory optimization study has been performed yet. We have extensively and systematically investigated different memory optimizations for the MAP (Maximum A Posteriori) class o...

Analog soft iterative error control decoders use analog computation to decode digital information. Analog soft-information processors offer the power of iterative decoding with a very small transistor count, enabling fully parallel decoding at low cost. They can be implemented in several technologies: BiCMOS and SiGe analog decoders are suitable for high-speed applications beyond 10 Gbit per se...

The performance of algebraic soft-decision decoding (ASD) of Reed-Solomon (RS) codes using bit-level soft information is investigated. Optimal multiplicity assignment strategies (MAS) of ASD with infinite cost are first studied over erasure channels and binary symmetric channels (BSC). The corresponding decoding radii are calculated in closed forms and tight bounds on the error probability are ...

In this paper, we study various technics to improve the performance, to reduce the computation cost and the required memory of a recognition system based on HMM. For the efficiency of the system, we first study the optimization of the number of HMM parameters according to training data. We experiment a temporal control of the phonetic transitions on lexical decoding task with a significant 5% i...

Background sprite decoding is an essential part of object-based video coding. The composition and rendering of a final scene involves the placing of individual video objects in a predefined way superimposed on the decoded background image. The MPEG-4 standard includes the decoding algorithm for background image decoding, but this algorithm is not suitable for implementation on a memory-constrai...

IRA (Irregular Repeat Accumulate) codes can be well utilized in wireless sensor systems, computer storage systems due to their approaching to Shannon limit by means of prominent decoding algorithm and etc. Belief propagation algorithm is the best decoding algorithm for IRA codes, but there is great hardware complexity with implementation. Minimum-sum algorithm modifies belief propagation algori...

Recently, minimum perfect hashing (MPH)-based language model (LM) lookup methods have been proposed for fast access of N-gram LM scores in lexical-tree based LVCSR (large vocabulary continuous speech recognition) decoding. Methods of node-based LM cache and LM context pre-computing (LMCP) have also been proposed to combine with MPH for further reduction of LM lookup time. Although these methods...

—Nowadays polar codes are becoming one of the most favorable capacity achieving error correction codes for their low encoding and decoding complexity. However, due to the large code length required by practical applications, the few existing successive cancellation (SC) decoder implementations still suffer from not only the high hardware cost but also the long decoding latency. This paper prese...

Attentional sequence-to-sequence models have become the new standard for machine translation, but one challenge of such models is a significant increase in training and decoding cost compared to phrase-based systems. Here, we focus on efficient decoding, with a goal of achieving accuracy close the state-of-the-art in neural machine translation (NMT), while achieving CPU decoding speed/throughpu...

A novel priority-based heading one detector for Exp-Golomb/CAVLC decoding of H.264/AVC is presented. It exploits the statistical distribution of input encoded codewords and adopts a nonuniform partition decoding scheme for the detector. Compared with a conventional design without power optimization, the power consumption can be reduced by more than 3 times while the performance is maintained an...