Deterministic Bit Research Articles

Deterministic Shuffling Networks to Implement Stochastic Circuits in Parallel

Stochastic computing (SC) in recent years has been defined as a digital computation approach that operates on streams of random bits that represent probability values. SC can perform complex tasks with much smaller hardware footprints compared with conventional binary methods, but previous methods on SC circuits operated on serial bit streams, which leads to high-latency implementations. This article presents a significant improvement over previous work; it provides a deterministic parallel bit shuffling network that can use a simple deterministic thermometer encoding of data, resulting in zero random fluctuation and high accuracy, yet keeping the output bit-stream length constant. We use core “stochastic” logic circuits that do not employ constant coefficients, making them significantly smaller than traditional stochastic logic that use a significant amount of resources to generate such coefficients. Our experiments show that compared with previous SC methods, our method has up to $3\times $ smaller mean absolute error, and better area $\times $ delay and power efficiency. Compared with conventional binary methods, our method is better in terms of area $\times $ delay at 8-bit resolution. It shows better power efficiency ( $40\times $ , $18\times $ , and $8\times $ Gops/W at 8-, 10-, and 12-bit resolutions) compared with conventional binary.

Corrected runs distribution test for pseudorandom number generators

Runs distribution test is an important statistical test based on Golomb's randomness postulates. It is an important randomness test included in some famous evaluation standards and statistical test suites such as Application Notes and Interpretation of the Scheme (AIS20), the Federal Information Processing Standards (FIPS140), CryptX, and so on. However, when it is applied on some well-known good Deterministic Random Bit Generators (DRBGs), the test results show apparent bias from randomness. This shows that there exist some inaccuracies in the runs distribution test. The problem is solved and the runs distribution test through two ways is corrected. First, the degree of freedom of statistical value V is adjusted. Secondly, the expected number of different lengths runs is modified. The experiment results show that the corrected runs distribution test is more accurate under two-level evaluation approach proposed by National Institute of Standards and Technology (NIST) Special Publication (SP) 800–22.

Multi-Layer Iterative LDPC Decoding for Broadband Wireless Access Networks: A Recursive Shortening Algorithm

This paper presents a novel multi-layer iterative decoding scheme using deterministic bits to lower the decoding threshold of LDPC codes in wireless multimedia communication systems. These deterministic bits serve as known information in the LDPC decoding process to reduce the redundancy during data transmission. Unlike the existing work, our proposed scheme addresses the controllable deterministic bits, such as MPEG null packets, rather than the widely investigated protocol headers. In particular, our multi-layer scheme integrates deterministic bit identification into LDPC decoding process. This integration is able to make significant performance improvement if adequate deterministic bits are discovered. We find that the length of deterministic bits may vary from time to time, and the exploring of deterministic bits is associated to a series of shortened parity matrices during the decoding iterations. Accordingly, we develop a recursive LDPC shortening algorithm to match the dynamic length and facilitate the iterative identification of deterministic bits. Simulation results show that our proposed scheme can achieve considerable gain in today's most popular broadband wireless access networks such as WiMAX.

Chirp Managed Laser and Applications

The chirp managed laser (CML) is an alternative transmitter technology that allows a directly modulated laser (DML) to be used in high-performance applications with a smaller size, lower power consumption, less device complexity, and lower cost. The CML comprises a DML and an isolated, passive optical filter. A key feature of the CML is its large tolerance to fiber dispersion, reaching over 360 km transmission at 10 Gb/s in standard single-mode fiber without dispersion compensation. The combination of adiabatic chirp from the laser and filter edge response produces high extinction ratio pulses with nearly uniform phase, abrupt phase shifts at bit transitions, and a correlation between the 1 bits; 1 bits separated by odd number of 0 bits are out of phase. This results in destructive intersymbol interference after fiber transmission. Other applications of CML are generation of advanced modulation formats, such as return-to-zero (RZ) alternate mark inversion, and RZ differential phase-shift keying. Tunable CML using DFB array, and sampled grating deterministic bit rate (DBR) technologies for 10 Gb/s metro applications have now been demonstrated. A four-element DBR array CML has also been demonstrated for 100 Gb/s metro applications. This paper reviews the principle of operation of CML and its various applications, as well as recent advances in CML-based devices.

Memory efficient error diffusion

Because of its good image quality and moderate computational requirements, error diffusion has become a popular halftoning solution for desktop printers, especially inkjet printers. By making the weights and thresholds tone-dependent and using a predesigned halftone bitmap for tone-dependent threshold modulation, it is possible to achieve image quality very close to that obtained with far more computationally complex iterative methods. However, the ability to implement error diffusion in very low cost or large format products is hampered by the requirement to store the tone-dependent parameters and halftone bitmap, and also the need to store error information for an entire row of the image at any given point during the halftoning process. For the first problem, we replace the halftone bitmap by deterministic bit flipping, which has been previously applied to halftoning, and we linearly interpolate the tone-dependent weights and thresholds from a small set of knot points. We call this implementation a reduced lookup table. For the second problem, we introduce a new serial block-based approach to error diffusion. This approach depends on a novel intrablock scan path and the use of different parameter sets at different points along that path. We show that serial block-based error diffusion reduces off-chip memory access by a factor equal to the block height. With both these solutions, satisfactory image quality can only be obtained with new cost functions that we have developed for the training process. With these new cost functions and moderate block size, we can obtain image quality that is very close to that of the original tone-dependent error diffusion algorithm.

Deterministic bit-stream digital neurons

In this paper, we present the design of a deterministic bit-stream neuron, which makes use of the memory rich architecture of fine-grained field-programmable gate arrays (FPGAs). It is shown that deterministic bit streams provide the same accuracy as much longer stochastic bit streams. As these bit streams are processed serially, this allows neurons to be implemented that are much faster than those that utilize stochastic logic. Furthermore, due to the memory rich architecture of fine-grained FPGAs, these neurons still require only a small amount of logic to implement. The design presented here has been implemented on a Virtex FPGA, which allows a very regular layout facilitating efficient usage of space. This allows for the construction of neural networks large enough to solve complex tasks at a speed comparable to that provided by commercially available neural-network hardware.

On-Line Approximate String Searching Algorithms: Survey and Experimental Results

The problem of approximate string searching comprises two classes of problems: string searching with k mismatches and string searching with k differences. In this paper we present a short survey and experimental results for well known sequential approximate string searching algorithms. We consider algorithms based on different approaches including dynamic programming, deterministic finite automata, filtering, counting and bit parallelism. We compare these algorithms in terms of running time against pattern length and for several values of k for four different kinds of text: binary alphabet, alphabet of size 8, English alphabet and DNA alphabet. Finally, we compare the experimental results of the algorithms with their theoretical complexities.

Adaptive threshold modulation for error diffusion halftoning

Grayscale digital image halftoning quantizes each pixel to one bit. In error diffusion halftoning, the quantization error at each pixel is filtered and fed back to the input in order to diffuse the quantization error among the neighboring grayscale pixels. Error diffusion introduces nonlinear distortion (directional artifacts), linear distortion (sharpening), and additive noise. Threshold modulation, which alters the quantizer input, has been previously used to reduce either directional artifacts or linear distortion. This paper presents an adaptive threshold modulation framework to improve halftone quality by optimizing error diffusion parameters in the least squares sense. The framework models the quantizer implicitly, so a wide variety of quantizers may be used. Based on the framework, we derive adaptive algorithms to optimize 1) edge enhancement halftoning and 2) green noise halftoning. In edge enhancement halftoning, we minimize linear distortion by controlling the sharpening control parameter. We may also break up directional artifacts by replacing the thresholding quantizer with a deterministic bit flipping (DBF) quantizer. For green noise halftoning, we optimize the hysteresis coefficients.

Accelerated Signalling for the Internet over ATM (ASIA)

AbstractTo face up the rapid development of Internet services with even more growing needs in terms of quality of service and bandwidth, we introduce in this paper an accelerated signalling capability for the ATM network. As a preliminary, we draw attention to the fact that cell spacing can regulate the dynamics of TCP over ATM if the spacing period is adequately chosen. Then, we propose to associate with each Internet information flow (generated by TCP or UDP) a Deterministic Bit Rate (DBR) virtual channel connection within the ATM network. The cell spacing mechanism actually makes the cell stream of an Internet information flow compliant with a DBR connection, for which quality of service objectives in terms of negligible cell loss and bounded cell transfer delays can be guaranteed. To achieve this, we study a synchronization method between the two connection establishment procedures, namely that at the Internet protocol level and the signalling procedure for the ATM network. The ATM connection is established throughout the network via a light‐weight signalling procedure, which relies on the exchange of single cell messages consisting of Resource Management (RM) cells. The addressing system proposed for this accelerated signalling capability is that of classical IP, which is and will certainly be the predominant one in the near future.

MPEG over ATM DBR service class for high quality transmission and effective bandwidth usage

The highly bursty nature of the moving picture experts group (MPEG) determines the bandwidth when it is applied to the asynchronous transfer mode (ATM) deterministic bit rate (DBR) service class. Furthermore, the MPEG source must declare peak cell rate (PCR), one of the traffic descriptors used in the call admission control (CAC) procedure, before actual transmission in ATM networks. However, the real-time encoding of the MPEG video introduces a difficulty in predicting a maximum picture size of the video, which is necessary to determine PCR. When an MPEG source can be equipped with the capability to adjust the encoding rate to the allocated bandwidth, the burstiness of the MPEG video can be lowered and the uncertainty in predicting PCR is avoided to some extent. In this case, a problem is apparently quality degradation caused by introducing such a rate adaptation mechanism, which is the main subject of this study. For the rate adaptation method, we consider a priority control method, a space reduction method, a time reduction method and a bandwidth renegotiation method. Quality of the video is evaluated and compared in terms of signal-to-noise ratio (SNR) as well as mean opinion score (MOS) because users' perception becomes an important measure of the quality of video transmissions. Through experimental results, it is shown that the space reduction method is useful for real-time transmission, whereas the time reduction method can be applicable if the users' application can tolerate the relatively large delay between source and destination. Furthermore, the applicability of the bandwidth renegotiation method is also validated. © 1997 by John Wiley & Sons, Ltd.

The controlled cell transfer capability

A Controlled Cell Transfer (CCT) capability has been proposed to ITU for definition in Recommendation I.371. The CCT capability, if accepted, will bring the number of ATM transfer layer capabilities to five. The other four capabilities being Deterministic Bit Rate (DBR), Statistical Bit Rate (SBR), ATM Block Transfer (ABT) and Available Bit Rate (ABR). It is believed that this capability will endow the ATM network with LAN like service over the wide area. In this article, we discuss the features and mechanisms of the proposed CCT capability. We also discuss the applications that CCT is targeted at supporting and present results of simulation studies, highlighting the advantages of CCT.