Recursive Encoders Research Articles

A Flexible Hierarchical Framework for Implicit 3D Characterization of Bionic Devices.

In practical applications, integrating three-dimensional models of bionic devices with simulation systems can predict their behavior and performance under various operating conditions, providing a basis for subsequent engineering optimization and improvements. This study proposes a framework for characterizing three-dimensional models of objects, focusing on extracting 3D structures and generating high-quality 3D models. The core concept involves obtaining the density output of the model from multiple images to enable adaptive boundary surface detection. The framework employs a hierarchical octree structure to partition the 3D space based on surface and geometric complexity. This approach includes recursive encoding and decoding of the octree structure and surface geometry, ultimately leading to the reconstruction of the 3D model. The framework has been validated through a series of experiments, yielding positive results.

Gated Recursive and Sequential Deep Hierarchical Encoding for Detecting Incongruent News Articles

With the increase in misinformation across digital platforms, incongruent news detection is becoming an important research problem. Earlier, researchers have exploited various feature engineering approaches and deep learning models with embedding to capture incongruity between news headlines and their respective bodies. Studies have broadly considered different combinations of bag-of-words-based features, sequential encoding, hierarchical encoding, headline-guided attention-based encoding, and so on, of the text in headlines and bodies. In this article, we focus on addressing two important limitations observed with hierarchical encoding and headline-guided attention-based encoding methods. The existing hierarchical encoding-based studies limit the hierarchical structure of the body of a news article to paragraph level only, undermining the importance of incorporating long-term dependence from word level to sentence, paragraph, and body. Furthermore, the existing headline-guided attention-based encoding focuses on contextually similar contents in the body of the headline, undermining the importance of incorporating contextually dissimilar contents. Motivated by the above observations, this article proposes a gated recursive and sequential deep hierarchical encoding (GraSHE) method for detecting incongruent news articles by extending the hierarchical structure of the news body from the body to the word level and incorporating incongruity weight. From various experimental setups over three publicly available benchmarks datasets, the experimental results indicate that the proposed model outperforms baseline models with bag-of-word-based features, sequential, hierarchical, and headline-guided attention-based encoding methods. To further validate the performance of the proposed model, we conduct several ablation studies. The following key observations can be made from the ablation study: 1) models with hierarchical encoding outperform models with nonhierarchical encoding; 2) recursive encoding of sentences boosts the performance of models as compared with sequential encoding of sentences within paragraphs; and 3) incongruent news article detection is domain-dependent. Incorporating explicit features further boosts the performance of proposed model and also decreases the domain dependence of models.

A Review on Higher Order Continuous-Time Sigma-Delta Modulators for ADC Design

Abstract—In this paper, the concept of adaptive filtering has been proposed. It has been shown that signals often undergo variation in the noise environment encountered. Hence it becomes mandatory to design filters based on noise shaping and adaptive noise cancellation. Such filters are called adaptive filters and are used for applications requiring real time noise cancellation such as speech processing, live video broadcasting etc. The paper presents the structure of Sigma Delta or MASH ADCs which employ recursive encoding. SDM ADCs are also used to convert high bit-count, low- frequency digital signals into lower bit-count, higher-frequency digital signals as part of the process to convert digital signals into analog as part of a digital-to-analog converter (DAC). The paper focuses on the various approaches used thus far for the purpose, citing the pros and cons of each approach. Keywords— Adaptive Filter Design, Noise Filtering, Sigma Delta Modulator (SDM), MASH, Signal to Noise Ratio.

RDNet-KD: Recursive Encoder, Bimodal Screening Fusion, and Knowledge Distillation Network for Rail Defect Detection

RDNet-KD: Recursive Encoder, Bimodal Screening Fusion, and Knowledge Distillation Network for Rail Defect Detection

Recursive autoencoder network for prediction of CAD model parameters from STEP files

Databases of 3D CAD (computer aided design) models are often large and lacking in meaningful organisation. Effective tools for automatically searching for, categorising and comparing CAD models, therefore, have many potential applications in improving efficiency within design processes. This paper presents a novel asymmetric autoencoder model, consisting of a recursive encoder network and fully-connected decoder network, for the reproduction of CAD models through prediction of the parameters necessary to generate a 3D part design. Inputs to the autoencoder are STEP (standard for the exchange of product data) files, an ISO standard CAD model format, compatible with all major CAD software. A complete 3D model can be accurately reproduced using a STEP file, meaning that all geometric information can be used to contribute to the final encoded vector, with no loss of small detail.In a CAD model of overall size 10 × 10 × 10 units, for 90% of models, the class of an added feature is estimated with maximum error of 0.6 units, feature size with maximum error of 0.4 units and coordinate values representing position with maximum error of 0.3 units. These results demonstrate the successful encoding of complex geometric information, beyond merely the shape of the 3D object, with potential application in the design of search engine functionality.

A novel neural network for improved in-hospital mortality prediction with irregular and incomplete multivariate data

Accurate estimation of in-hospital mortality based on patients’ physiological time series data improves the performance of the clinical decision support systems and assists hospital providers in allocating resources. In practice, the data quality issues of missing values are ubiquitous in electronic health records (EHRs). Since the vital signs are usually observed with irregular temporal intervals and different sampling rates, it is challenging to predict clinical outcomes with sparse and incomplete multivariate time series. We propose an auto-regressive recurrent neural network (RNN) based model, dubbed the bi-directional recursive encoder–decoder network (BiRED), to jointly perform data imputation and mortality prediction. To capture complex patterns of medical time sequences, a 2D cross-regression with an RNN unit (2DCR-RNN) and an imputation block with an RNN unit (IB-RNN) are designed as the recurrent component of the encoder and decoder, respectively. Furthermore, a state initialization method is proposed to alleviate errors accumulated in the generated sequence. The experimental results on two real EHR datasets show that our proposed method can predict hospital mortality with high AUC scores.

Recursive encoder network for the automatic analysis of STEP files

Automated tools which can understand and interface with CAD (computer-aided design) models are of significant research interest due to the potential for improving efficiency in manufacturing processes. At present, most research into the use of artificial intelligence to interpret three-dimensional data takes input in the form of multiple two-dimensional images of the object or in the form of three-dimensional grids of voxels. The transformation of the input data necessary for these approaches inevitably leads to some loss of information and limitations of resolution. Existing research into the direct analysis of model files in STEP (standard for the exchange of product data) format tends to follow a rules-based approach to analyse models of a certain type, resulting in algorithms without the benefits of flexibility and complex understanding which artificial intelligence can provide. In this paper, a novel recursive encoder network for the automatic analysis of STEP files is presented. The encoder network is a flexible model with the potential for adaptation to a wide range of tasks and finetuning for specific CAD model datasets. Performance is evaluated using a machining feature classification task, with results showing accuracy approaching 100% and training time comparable to that of existing multi-view and voxel-based solutions without the need for a GPU.

Finding Patterns in Signals Using Lossy Text Compression

Whether the source is autonomous car, robotic vacuum cleaner, or a quadcopter, signals from sensors tend to have some hidden patterns that repeat themselves. For example, typical GPS traces from a smartphone contain periodic trajectories such as “home, work, home, work, ⋯”. Our goal in this study was to automatically reverse engineer such signals, identify their periodicity, and then use it to compress and de-noise these signals. To do so, we present a novel method of using algorithms from the field of pattern matching and text compression to represent the “language” in such signals. Common text compression algorithms are less tailored to handle such strings. Moreover, they are lossless, and cannot be used to recover noisy signals. To this end, we define the recursive run-length encoding (RRLE) method, which is a generalization of the well known run-length encoding (RLE) method. Then, we suggest lossy and lossless algorithms to compress and de-noise such signals. Unlike previous results, running time and optimality guarantees are proved for each algorithm. Experimental results on synthetic and real data sets are provided. We demonstrate our system by showing how it can be used to turn commercial micro air-vehicles into autonomous robots. This is by reverse engineering their unpublished communication protocols and using a laptop or on-board micro-computer to control them. Our open source code may be useful for both the community of millions of toy robots users, as well as for researchers that may extend it for further protocols.

An Efficient Method to Construct Parity-Check Matrices for Recursively Encoding Spatially Coupled LDPC Codes †

Spatially coupled low-density parity-check (LDPC) codes have attracted considerable attention due to their promising performance. Recursive encoding of the codes with low delay and low complexity has been proposed in the literature but with constraints or restrictions. In this manuscript we propose an efficient method to construct parity-check matrices for recursively encoding spatially coupled LDPC codes with arbitrarily chosen node degrees. A general principle is proposed, which provides feasible and practical guidance for the construction of parity-check matrices. According to the specific structure of the matrix, each parity bit at a coupling position is jointly determined by the information bits at the current position and the encoded bits at former positions. Performance analysis in terms of design rate and density evolution has been presented. It can be observed that, in addition to the feature of recursive encoding, selected code structures constructed by the newly proposed method may lead to better belief-propagation thresholds than the conventional structures. Finite-length simulation results are provided as well, which verify the theoretical analysis.

RLT Code Based Handshake-Free Reliable MAC Protocol for Underwater Sensor Networks

The characteristics of underwater acoustic channels such as long propagation delay and low bit rate cause the medium access control (MAC) protocols designed for radio channels to either be inapplicable or have low efficiency for underwater sensor networks (UWSNs). Meanwhile, due to high bit error, conventional end-to-end reliable transfer solutions bring about too many retransmissions and are inefficient in UWSN. In this paper, we present a recursive LT (RLT) code. With small degree distribution and recursive encoding, RLT achieves reliable transmission hop-by-hop while reducing the complexity of encoding and decoding in UWSN. We further propose an RLT code based handshake-free (RCHF) reliable MAC protocol. In RCHF protocol, each node maintains a neighbor table including the field of state, and packages are forwarded according to the state of a receiver, which can avoid collisions of sending-receiving and overhearing. The transmission-avoidance time in RCHF decreases data-ACK collision dramatically. Without RTS/CTS handshaking, the RCHF protocol improves channel utilization while achieving reliable transmission. Simulation results show that, compared with the existing reliable data transport approaches for underwater networks, RCHF can improve network throughput while decreasing end-to-end overhead.

Image Transmission Based on QC-LDPC Codes with Simple Recursive Encoding Form

Low encoding delay and complexity is very important for image transmission. This paper proposes a novel image transmission scheme with low encoding complexity. The proposed scheme is based on quasi-cyclic low density parity check (QC-LDPC) codes with a simple recursive encoding form (SREF QC-LDPC code) which results in low encoding complexity and delay. Constructing the SREF QC-LDPC codes in this scheme composes of two main steps, construction of the base matrix and the exponent matrix. We combine the differential evolution and protograph extrinsic information transfer (PEXIT) method to optimize the base matrix of QC-LDPC code. Consequently, the exponent matrix and the parity check matrix are constructed. Simulation results show that the proposed scheme based on SREF QC-LDPC code can provide a good tradeoff between the performance and complexity.

Recursive encoding and decoding of the noiseless subsystem for qudits

We give a full explanation of the noiseless subsystem that protects a single-qubit against collective errors and the corresponding recursive scheme described by C.-K. Li et. al. [Phys. Rev. A 84, 044301 (2011)] from a representation theory point of view. Furthermore, we extend the construction to qudits under the influence of collective SU($d$) errors. We find that under this recursive scheme, the asymptotic encoding rate is $1/d$.

Recurrent neural collective classification.

With the recent surge in availability of data sets containing not only individual attributes but also relationships, classification techniques that take advantage of predictive relationship information have gained in popularity. The most popular existing collective classification techniques have a number of limitations-some of them generate arbitrary and potentially lossy summaries of the relationship data, whereas others ignore directionality and strength of relationships. Popular existing techniques make use of only direct neighbor relationships when classifying a given entity, ignoring potentially useful information contained in expanded neighborhoods of radius greater than one. We present a new technique that we call recurrent neural collective classification (RNCC), which avoids arbitrary summarization, uses information about relationship directionality and strength, and through recursive encoding, learns to leverage larger relational neighborhoods around each entity. Experiments with synthetic data sets show that RNCC can make effective use of relationship data for both direct and expanded neighborhoods. Further experiments demonstrate that our technique outperforms previously published results of several collective classification methods on a number of real-world data sets.

Polar Codes for Degradable Quantum Channels

Channel polarization is a phenomenon in which a particular recursive encoding induces a set of synthesized channels from many instances of a memoryless channel, such that a fraction of the synthesized channels becomes near perfect for data transmission and the other fraction becomes near useless for this task. Mahdavifar and Vardy have recently exploited this phenomenon to construct codes that achieve the symmetric private capacity for private data transmission over a degraded wiretap channel. In the current paper, we build on their work and demonstrate how to construct quantum wiretap polar codes that achieve the symmetric private capacity of a degraded quantum wiretap channel with a classical eavesdropper. Due to the Schumacher-Westmoreland correspondence between quantum privacy and quantum coherence, we can construct quantum polar codes by operating these quantum wiretap polar codes in superposition, much like Devetak's technique for demonstrating the achievability of the coherent information rate for quantum data transmission. Our scheme achieves the symmetric coherent information rate for quantum channels that are degradable with a classical environment. This condition on the environment may seem restrictive, but we show that many quantum channels satisfy this criterion, including amplitude damping channels, photon-detected jump channels, dephasing channels, erasure channels, and cloning channels. Our quantum polar coding scheme has the desirable properties of being channel-adapted and symmetric capacity-achieving along with having an efficient encoder, but we have not demonstrated that the decoding is efficient. Also, the scheme may require entanglement assistance, but we show that the rate of entanglement consumption vanishes in the limit of large blocklength if the channel is degradable with classical environment.

Implementation of a reconfigurable turbo decoder in 3GPP for flat Rayleigh fading

Quality of service is a critical consideration in the design of mobile systems, since it allows the user to receive high quality services. Therefore, in 3GPP systems, in order to realise a particular service, the quality of service requirements in terms of performance and latency, have to be satisfied. Turbo code features include parallel code concatenation, recursive convolutional encoding, nonuniform interleaving and an associated iterative decoding algorithm. Exploiting the quality of service classification according to the priority of latency or performance, possible examples of service scenarios are examined for flat Rayleigh fading channels with emphasis on the turbo decoding algorithm. Particularly, for two operating environments considering SOVA and log-MAP algorithms due to their data-flow similarities, this paper shows that SOVA is clearly optimal for most of the real-time applications, whereas for nonreal time applications with low data rate and small frames log-MAP is preferred. The use of the optimum algorithm in most scenarios results in a more efficient turbo decoder: applications that otherwise would have failed now can be realised.

Near-optimum serial concatenation of single-parity codes with convolutional codes

A class of concatenated codes using convolutional codes and low-density parity check codes as component codes is proposed, extending the author's previous work. A special case is considered serial concatenation of single parity check codes as the outer code, a convolutional code with recursive systematic encoder as the inner code, separated by a random interleaver. A way to design the interleaver ‘or the corresponding LDPC code’ is presented to minimise the correlation between the extrinsic information. The results show that this parity concatenated code can operate within 0.45 dB of the Shannon limit. Similar to the parity concatenated codes presented in earlier work, these codes are superior in approaching the sphere packing bound (SPB), achieve near-SPB decoding can be achieved for block sizes ranging from 200 (about 0.8 dB) to 10 000 (within 0.6 dB) information bits (for a block error rate of 10−4). The improvement over previous coding schemes is 0.3–0.4 dB for block lengths more than 300 bits, and over JPL turbo codes is 0.3 dB for block lengths less than 1000 bits. A reliable error-detection method is proposed that makes this coding scheme very attractive for communications using short packets.

Analytic Expressions for the Bit Error Probabilities of Rate-<tex>$1/2$</tex>Memory<tex>$2$</tex>Convolutional Encoders

Analytic expressions for the exact bit error probabilities of rate R=1/2, memory m=2 convolutional encoders are derived for a maximum-likelihood (ML) decoder and transmission over the binary-symmetric channel (BSC). The resulting expressions are rational functions of the crossover probability of the BSC. In addition to classical nonsystematic encoders without feedback, we consider also recursive systematic encoders, which became especially important as component encoders in concatenated coding schemes. To attest the validity of the results, they are compared to computer simulations. Based on the presented technique also the bit error probability and the probability distribution of the output log-likelihood ratios of the Max-Log-MAP algorithm are derived in analytic form.

Turbo Multiuser Detection for Differentially Modulated CDMA

In this letter, we study differentially modulated, iteratively decoded CDMA. The iterative multiuser receiver proposed consists of an additional soft-input soft-output (SISO) differential decoder, when compared to turbo multiuser detectors for absolutely modulated systems. Algorithms for iterative decoding with and without phase information at the receiver are developed. The resulting turbo receivers with differential modulation outperform coherent receivers with absolute modulation at moderate to high signal to noise ratios due to the interleaver gain associated with recursive inner encoders in serially concatenated encoding structures.

Concatenated codes for fading channels based on recursive space--time trellis codes

We propose a class of codes which combine the principles of turbo coding and space-time trellis codes. It is first shown that several classes of space-time codes have an equivalent recursive realization. This fact is then exploited to design serial concatenated coding schemes with an outer code, interleaver, and an inner recursive space-time encoder. Two solutions are proposed in this paper - the use of convolutional outer codes aimed mainly to improve the power efficiency and the use of very high-rate outer codes to obtain significant improvement in power efficiency with a marginal decrease in spectral efficiency. We show that single parity check based turbo product codes are a good candidate for very high-rate outer codes. Finally, we propose an automatic repeat request scheme based on recursive realizations of space-time codes and show that the proposed scheme provides significant reduction in frame error rate.

Novel algorithm for continuous decoding of turbo codes

A novel type of recursive algorithm for decoding turbo codes with a convolutional interleaver is proposed, based on a symbol-by-symbol maximum a posteriori probability algorithm. It requires only a forward recursion which can be performed in parallel, and the number of variables to be stored does not increase with decoding delay. This algorithm can be used in continuous decoding for both recursive and nonrecursive encoders. Simulation results of demonstrating its performance are presented.