Mitigate Error Propagation Research Articles

Enhancing handwritten text recognition accuracy with gated mechanisms

Handwritten Text Recognition (HTR) is a challenging task due to the complex structures and variations present in handwritten text. In recent years, the application of gated mechanisms, such as Long Short-Term Memory (LSTM) networks, has brought significant advancements to HTR systems. This paper presents an overview of HTR using a gated mechanism and highlights its novelty and advantages. The gated mechanism enables the model to capture long-term dependencies, retain relevant context, handle variable length sequences, mitigate error propagation, and adapt to contextual variations. The pipeline involves preprocessing the handwritten text images, extracting features, modeling the sequential dependencies using the gated mechanism, and decoding the output into readable text. The training process utilizes annotated datasets and optimization techniques to minimize transcription discrepancies. HTR using a gated mechanism has found applications in digitizing historical documents, automatic form processing, and real-time transcription. The results show improved accuracy and robustness compared to traditional HTR approaches. The advancements in HTR using a gated mechanism open up new possibilities for effectively recognizing and transcribing handwritten text in various domains. This research does a better job than the most recent iteration of the HTR system when compared to five different handwritten datasets (Washington, Saint Gall, RIMES, Bentham and IAM). Smartphones and robots are examples of low-cost computing devices that can benefit from this research.

BCSLinker: automatic method for constructing a knowledge graph of venous thromboembolism based on joint learning.

Venous thromboembolism (VTE) is characterized by high morbidity, mortality, and complex treatment. A VTE knowledge graph (VTEKG) can effectively integrate VTE-related medical knowledge and offer an intuitive description and analysis of the relations between medical entities. However, current methods for constructing knowledge graphs typically suffer from error propagation and redundant information. In this study, we propose a deep learning-based joint extraction model, Biaffine Common-Sequence Self-Attention Linker (BCSLinker), for Chinese electronic medical records to address the issues mentioned above, which often occur when constructing a VTEKG. First, the Biaffine Common-Sequence Self-Attention (BCsSa) module is employed to create global matrices and extract entities and relations simultaneously, mitigating error propagation. Second, the multi-label cross-entropy loss is utilized to diminish the impact of redundant information and enhance information extraction. We used the electronic medical record data of VTE patients from a tertiary hospital, achieving an F1 score of 86.9% on BCSLinker. It outperforms the other joint entity and relation extraction models discussed in this study. In addition, we developed a question-answering system based on the VTEKG as a structured data source. This study has constructed a more accurate and comprehensive VTEKG that can provide reference for diagnosing, evaluating, and treating VTE as well as supporting patient self-care, which is of considerable clinical value.

An inspired chaos-based estimation-theory optimization for low-density parity-check (LDPC) code decoding

Optimization techniques have emerged as powerful tools for solving complex engineering problems in real-time applications, surpassing traditional heuristic methods. This paper introduces a novel chaos optimization algorithm that integrates the bisection search method to enhance the convergence ability, solution accuracy, and search speed of conventional optimization problems. The proposed algorithm is applied to the design and analysis of a reduced complexity decoding scheme using Low-Density Parity-Check (LDPC) codes. By leveraging the Linear Minimum Mean Square Error (LMMSE) estimation criterion, the proposed method approximates the estimated parameters acquired during the message computation process. This combined approach effectively mitigates error propagation and premature convergence issues. The presented research offers a low-complexity decoding algorithm for LDPC codes, leveraging an optimized chaos bisection search method. The adoption of the LMMSE criterion overcomes data dependency and error propagation concerns during the estimation process. The modified chaos optimization technique, integrated with the bisection search, efficiently determines optimal values from perturbed observations, leading to improved convergence rates and data processing precision by reducing the observation dimension (search space). Simulation results validate the proposed approach, showcasing its ability to mitigate error propagation and premature convergence while maintaining coding gain accuracy. Experimental validation on an FPGA RF-SoC device (model XCZU27DR) demonstrates that the proposed method achieves a maximum throughput of 8.7 Gbps at a clock frequency of 386.72 MHz, with reduced hardware complexity. The experimental verification on an FPGA RF-SoC device highlights the algorithm's minimal hardware complexity, superior decoding throughput, and Bit Error Rate (BER) performance.

Modulated Memory Network for Video Object Segmentation

Existing video object segmentation (VOS) methods based on matching techniques commonly employ a reference set comprising historical segmented frames, referred to as ‘memory frames’, to facilitate the segmentation process. However, these methods suffer from the following limitations: (i) Inherent segmentation errors in memory frames can propagate and accumulate errors when utilized as templates for subsequent segmentation. (ii) The non-local matching technique employed in top-leading solutions often fails to incorporate positional information, potentially leading to incorrect matching. In this paper, we introduce the Modulated Memory Network (MMN) for VOS. Our MMN enhances matching-based VOS methods in the following ways: (i) Introducing an Importance Modulator, which adjusts memory frames using adaptive weight maps generated based on the segmentation confidence associated with each frame. (ii) Incorporating a Position Modulator that encodes spatial and temporal positional information for both memory frames and the current frame. The proposed modulator improves matching accuracy by embedding positional information. Meanwhile, the Importance Modulator mitigates error propagation and accumulation by incorporating confidence-based modulation. Through extensive experimentation, we demonstrate the effectiveness of our proposed MMN, which also achieves promising performance on VOS benchmarks.

Enhancing cryptographic robustness with dual key chaining

In this paper, we introduce an advanced mode of operation for block ciphers, named Dual Key Chaining Mode (DKC), aimed at bolstering cryptographic security for safeguarding sensitive information. Building upon the foundations laid by established modes while adhering to guidelines set by the National Institute of Standards and Technology (NIST), DKC innovates through a dual-key mechanism and the generation of highly unpredictable values. This novel approach markedly enhances security, particularly against chosen plaintext attacks, a common vulnerability in traditional modes. Through rigorous mathematical analysis, we demonstrate DKC’s superiority, proving its indistinguishability under chosen plaintext attacks (IND-CPA) and showing that an adversary cannot practically distinguish DKC-encrypted ciphertexts from those produced by a random permutation. Our security proof employs a structured approach, contrasting DKC with conventional modes to highlight its robust defense mechanisms and its capacity to mitigate error propagation, reduce chain dependency, and resist pattern recognition attacks. The DKC mode not only surpasses existing standards in cryptographic security but also offers significant improvements in efficiency and security complexity, making it particularly suited for environments demanding stringent data protection. This study’s findings underscore DKC’s potential as a leading candidate for securing communication channels, financial transactions, and cloud storage services against an array of cryptographic attacks.

Deep Learning-Based Detection Algorithm for the Multi-User MIMO-NOMA System

Recently, non-orthogonal multiple access (NOMA) has become prevalent in 5G communication. However, the traditional successive interference cancellation (SIC) receivers for NOMA still encounter challenges. The near-far effect between the users and the base stations (BS) results in a higher bit error rate (BER) for the SIC receiver. Additionally, the linear detection algorithm used in each SIC stage fails to eliminate the interference and is susceptible to error propagation. Consequently, designing a high-performance NOMA system receiver is a crucial challenge in NOMA research and particularly in signal detection. Focusing on the signal detection of the receiver in the NOMA system, the main work is as follows. (1) This thesis leverages the strengths of deep neural networks (DNNs) for nonlinear detection and incorporates the low computational complexity of the successive interference cancellation (SIC) structure. The proposed solution introduces a feedback deep neural network (FDNN) receiver to replace the SIC in signal detection. By employing a deep neural network for nonlinear detection at each stage, the receiver mitigates error propagation, lowers the BER in NOMA systems, and enhances resistance against inter-user interference (IUI). (2) We describe its algorithm flow and provide simulation results comparing FDNN and SIC receivers under MIMO-NOMA scenarios. The simulations clearly demonstrate that FDNN receivers outperform SIC receivers in terms of BER for MIMO-NOMA systems.

SRoUDA: Meta Self-Training for Robust Unsupervised Domain Adaptation

As acquiring manual labels on data could be costly, unsupervised domain adaptation (UDA), which transfers knowledge learned from a rich-label dataset to the unlabeled target dataset, is gaining increasingly more popularity. While extensive studies have been devoted to improving the model accuracy on target domain, an important issue of model robustness is neglected. To make things worse, conventional adversarial training (AT) methods for improving model robustness are inapplicable under UDA scenario since they train models on adversarial examples that are generated by supervised loss function. In this paper, we present a new meta self-training pipeline, named SRoUDA, for improving adversarial robustness of UDA models. Based on self-training paradigm, SRoUDA starts with pre-training a source model by applying UDA baseline on source labeled data and taraget unlabeled data with a developed random masked augmentation (RMA), and then alternates between adversarial target model training on pseudo-labeled target data and fine-tuning source model by a meta step. While self-training allows the direct incorporation of AT in UDA, the meta step in SRoUDA further helps in mitigating error propagation from noisy pseudo labels. Extensive experiments on various benchmark datasets demonstrate the state-of-the-art performance of SRoUDA where it achieves significant model robustness improvement without harming clean accuracy.

A bilateral context and filtering strategy-based approach to Chinese entity synonym set expansion

AbstractEntity synonyms play a significant role in entity-based tasks. Previous approaches use linguistic syntax, distributional, and semantic features to expand entity synonym sets from text corpora. Due to the flexibility and complexity of the Chinese language expression, the aforementioned approaches are still difficult to expand entity synonym sets robustly from Chinese text, because these approaches fail to track holistic semantics among entities and suffer from error propagation. This paper introduces an approach for expanding Chinese entity synonym sets based on bilateral context and filtering strategy. Specifically, the approach consists of two novel components. First, a bilateral-context-based Siamese network classifier is proposed to determine whether a new entity should be inserted into the existing entity synonym set. The classifier tracks the holistic semantics of bilateral contexts and is capable of imposing soft holistic semantic constraints to improve synonym prediction. Second, a filtering-strategy-based set expansion algorithm is presented to generate Chinese entity synonym sets. The filtering strategy enhances semantic and domain consistencies to filter out wrong synonym entities, thereby mitigating error propagation. Experimental results on two Chinese real-world datasets demonstrate that the proposed approach is effective and outperforms the selected existing state-of-the-art approaches to the Chinese entity synonym set expansion task.

DEERE: Document-Level Event Extraction as Relation Extraction

The descriptions of complex events usually span sentences, so we need to extract complete event information from the whole document. To address the challenges of document-level event extraction, we propose a novel framework named Document-level Event Extraction as Relation Extraction (DEERE), which is suitable for document-level event extraction tasks without trigger-word labelling. By well-designed task transformation, DEERE remodels event extraction as single-stage relation extraction, which can mitigate error propagation. A long text supported encoder is adopted in the relation extraction model to aware the global context effectively. A fault-tolerant event integration algorithm is designed to improve the prediction accuracy. Experimental results show that our approach advances the SOTA for the ChFinAnn dataset by an average F1-score of 3.7. The code and data are available at https://github.com/maomaotfntfn/DEERE.

Timing Recovery and Adaptive Equalization for Discrete Multi-Tone Signalling in Wireline Applications

This paper proposes a discrete multi-tone timing-recovery system with adaptive equalization for ultra-high-speed wireline applications. It combines frequency-domain clock recovery with decision-directed equalization to improve receiver performance while eliminating the need for pilot carriers, thereby increasing spectral efficiency. Compared to a conventional pilot-carrier-based technique employing four pilot carriers and a 32-point FFT, this approach improves phase-error sensitivity by 3.6 times, tracking bandwidth by 1.7 times, increases the jitter tolerance slope by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$20dB$ </tex-math></inline-formula> per decade at low frequency, and removes residual equalization error, resulting in an overall data-rate increase of 27%. The concept is validated at the system-level and gate-level through synthesis in an FPGA. A convergence analysis of both the adaptive equalizer and clock synchronization shows the system’s ability to mitigate error propagation and remain synchronized in the presence of impairments. Finally, we highlight the system’s ability to trade-off clock convergence versus phase error sensitivity. Either parameter can be adjusted by 15 times, optimizing the receiver over a broad range of signal conditions.

R2D2: A Region-Based Recursive Doppler Dealiasing Algorithm for Operational Weather Radar

AbstractRegion-based Recursive Doppler Dealiasing (R2D2) is a novel dealiasing algorithm to unfold Doppler velocity fields obtained by operational radar measurements. It specializes in resolving issues when the magnitude of the gate-to-gate velocity shear approaches or exceeds the Nyquist velocity. This occurs either in highly sheared situations, or when the Nyquist velocity is low. Highly sheared situations, such as convergence lines or mesocyclones, are of particular interest for nowcasting and warnings. R2D2 masks high-shear areas and adds a spatial buffer around them. The areas between the buffers are then identified as continuous regions that lie within the same Nyquist interval. Each region subsequently is assigned its most likely Nyquist interval by applying vertical and temporal continuity constraints, as well as supplemental wind information from an operational mesoscale model. The shear zones are then resolved using 2D continuity in azimuth and range. This 4D procedure is repeated until no further improvement can be achieved. Each iteration with fewer folds identifies fewer but larger continuous regions and less shear zones until an optimum is reached. Residual errors, often related to shear greater than the Nyquist velocity, are contained to small areas within the buffer zones. This approach maximizes operational performance in high-shear situations and restricts errors to minimal areas to mitigate error propagation.

Bidirectional Soft-Decision Feedback Equalization for OFDM Systems

A bidirectional soft-decision feedback equalizer (BiSDFE) for orthogonal frequency-division multiplexing (OFDM) systems over doubly selective channels is proposed. Soft, rather than hard, decision feedback equalizer (SDFE) is adopted to suppress intercarrier interference (ICI) and the band structure of the frequency-domain channel matrix is exploited to trade off between performance and computational complexity. By combining a conventional SDFE with a frequency-reversed SDFE, the proposed scheme can eliminate ICI induced by neighboring subcarriers of both sides, thereby obtaining frequency-reversal diversity and mitigating error propagation. Simulation results have demonstrated the advantage of this approach in bit error rate (BER), especially in the channels with large Doppler spreads.

Asymmetric Adaptive Modulation With Symbol-Based Network Coding for Heterogeneous Two-Way Relay Network

In this paper, we propose an asymmetric adaptive modulation scheme for a heterogeneous two-way relay network (TWRN). In this scheme, due to the heterogeneity of the TWRN, different modulation schemes with different bit rates are assigned to the source nodes. Due to the mismatch between the bit rates of the two opposite traffic flows, the conventional XOR-based network coding is not applicable to this scheme. To cope with this heterogeneity at the relay node, we propose a network coding scheme in which the two data streams are combined at the level of symbol. We optimize the pair of modulation schemes and bit rates such that the instantaneous transmission rate is maximized under the constraint that a quality-of-service requirement in terms of the end-to-end BER is satisfied. To find the solution of this multi-dimensional problem, we propose a simple procedure with very low computational complexity by reducing the problem dimension to one. We analyze the system performance in terms of the average end-to-end BER, the average transmission rate, and the outage probability over Nakagami fading channel. Our analysis shows that this scheme not only achieves high spectral efficiency but also mitigates error propagation from the relay to the transceivers.

Ordered Sequence Detection and Barrier Signal Design for Digital Pulse Interval Modulation in Optical Wireless Communications

This paper proposes an ordered sequence detection (OSD) for digital pulse interval modulation (DPIM) in optical wireless communications. Leveraging the sparsity of DPIM sequences, OSD shows a comparable performance to the optimal maximum likelihood sequence detection with much lower complexity. Compared with the widely adopted sample-by-sample optimal threshold detection (OTD), it considerably improves the bit-error-rate (BER) performance by mitigating error propagation. Moreover, this paper proposes a barrier signal-aided digital pulse interval modulation (BDPIM), where the last of every $K$ symbols is allocated with more power as an inserted barrier signal. BDPIM with OSD (BDPIM-OSD) can limit the error propagation between two adjacent barriers. To reduce the storing delay when using OSD to detect extremely large packets, we propose BDPIM with a combination of OTD and OSD (BDPIM-OTD-OSD), within which long sequences are cut into pieces and separately detected. Approximate upper bounds of the average BER performance of DPIM-OTD, DPIM-OSD, BDPIM-OSD, and BDPIM-OTD-OSD are analysed. Simulations are conducted to corroborate our analysis. Optimal parameter settings are also investigated in uncoded and coded systems by simulations. Simulation results show that the proposed OSD and BDPIM bring significant improvement in uncoded and coded systems over various channels.

Symbol-Level Selective Full-Duplex Relaying With Power and Location Optimization

In this paper, a symbol-level selective transmission for full-duplex (FD) relaying networks is proposed to mitigate error propagation effects and improve system spectral efficiency. The idea is to allow the FD relay node to predict the correctly decoded symbols of each frame, based on the generalized square deviation method, and discard the erroneously decoded symbols, resulting in fewer errors being forwarded to the destination node. Using the capability for simultaneous transmission and reception at the FD relay node, our proposed strategy can improve the transmission efficiency without extra cost of signaling overhead. In addition, targeting on the derived expression for outage probability, we compare it with half-duplex relaying case and provide the transmission power and relay location optimization strategy to further enhance the system performances. The results show that our proposed scheme outperforms the classic relaying protocols, such as cyclic redundancy check-based selective decode-and-forward (S-DF) relaying and threshold-based S-DF relaying in terms of outage probability and bit error rate. Moreover, the performances with optimal power allocation are better than those with equal power allocation, especially when the FD relay node encounters strong self-interference and/or it is close to the destination node.

Efficient Low-Latency Implementation of CORDIC-Based Sorted QR Decomposition for Multi-Gbps MIMO Systems

This brief presents the efficient VLSI implementation of coordinate rotation digital computer (CORDIC)-based sorted QR decomposition (SQRD) for multiple-input and multiple-output (MIMO) systems. SQRD is widely adopted in MIMO signal processing to mitigate error propagation. However, its iterative sorting process after every column annihilation requires a large hardware overhead and suffers from a latency issue. The proposed CORDIC-based SQRD modifies a real-value decomposition matrix to utilize symmetry and jointly performs sorting processes and CORDIC rotations with adjacent symmetric columns to reduce the required number of CORDIC rotations and CORDIC stages. Furthermore, the VLSI design of the proposed SQRD has been synthesized and implemented with a Virtex-6 FPGA and a 65-nm CMOS technology, respectively. The 65-nm implementation results show an overall processing latency of 266.5 ns, a throughput of 48.8 MSQRD per second, and can support a 4.7-Gbps MIMO system throughput.

Using volume-weighted average wood specific gravity of trees reduces bias in aboveground biomass predictions from forest volume data

With the improvement of remote sensing techniques for forest inventory application such as terrestrial LiDAR, tree volume can now be measured directly, without resorting to allometric equations. However, wood specific gravity (WSG) remains a crucial factor for converting these precise volume measurements into unbiased biomass estimates. In addition to this WSG values obtained from samples collected at the base of the tree (WSGBase) or from global repositories such as Dryad (WSGDryad) can be substantially biased relative to the overall tree value. Our aim was to assess and mitigate error propagation at tree and stand level using a pragmatic approach that could be generalized to National Forest Inventories or other carbon assessment efforts based on measured volumetric data. In the semi-deciduous forests of Eastern Cameroon, we destructively sampled 130 trees belonging to 15 species mostly represented by large trees (up to 45 Mg). We also used stand-level dendrometric parameters from 21 1-ha plots inventoried in the same area to propagate the tree-level bias at the plot level. A new descriptor, volume average-weighted WSG (WWSG) of the tree was computed by weighting the WSG of tree compartments by their relative volume prior to summing at tree level. As WWSG cannot be assessed non-destructively, linear models were adjusted to predict field WWSG and revealed that a combination of WSGDryad, diameter at breast height (DBH) and species stem morphology (Sm) were significant predictors explaining together 72% of WWSG variation. At tree level, estimating tree aboveground biomass using WSGBase and WSGDryad yielded overestimations of 10% and 7% respectively whereas predicted WWSG only produced an underestimation of less than 1%. At stand-level, WSGBase and WSGDryad gave an average simulated bias of 9% (S.D. = ±7) and 3% (S.D. = ±7) respectively whereas predicted WWSG reduced the bias by up to 0.1% (S.D. = ±8). We also observed that the stand-level bias obtained with WSGBase and WSGDryad decreased with total plot size and plot area. The systematic bias induced by WSGBase and WSGDryad for biomass estimations using measured volumes are clearly not negligible but yet generally overlooked. A simple corrective approach such as the one proposed with our predictive WWSG model is liable to improve the precision of remote sensing-based approaches for broader scale biomass estimations.

Soft direct-adaptation based bidirectional turbo equalization for MIMO underwater acoustic communications

This paper proposes a soft direct-adaptation based bidirectional turbo equalizer for multiple-input multiple-output underwater acoustic communication systems. Soft, rather than hard, direct-adaptation based equalizer combined with the fast self-optimized least mean square algorithm is employed to achieve a faster convergence rate, and the second-order phase-locked loop is embedded into the equalizer to track the time-varying channel. Meanwhile, by utilizing a weighted linear combining scheme, the conventional soft direct- adaptation based equalizer is combined with the time-reversed soft direct-adaptation based equalizer to exploit bidirectional diversity and mitigate error propagation. Both the simulation and experimental results demonstrate that the soft direct-adaptation based bidirectional turbo equalizer outperforms the single-direction soft direct-adaptation based turbo equalizer, and achieves a faster convergence rate than the hard direct-adaptation based bidirectional turbo equalizer.

Bidirectional Soft-Decision Feedback Turbo Equalization for MIMO Systems

This paper proposes a bidirectional soft-decision feedback turbo equalizer for severe triply selective fading channels. The proposed bidirectional soft-decision feedback equalizer (Bi-SDFE) uses a time reversed SDFE in conjunction with a normal SDFE to harvest time-reverse diversity and mitigate error propagation. A simple and effective linear combining scheme is derived for combining the extrinsic information on the two SDFEs for multiple-input–multiple-output (MIMO) systems with multilevel modulation. The bit-error-rate (BER) simulation results show that the proposed Bi-SDFE exhibits a lower signal-to-noise-ratio (SNR) threshold than the original SDFE and outperforms the well-known exact turbo minimum mean square error (MMSE) linear equalizers (LEs) at medium-to-high SNRs. Moreover, both extrinsic information transfer (EXIT) chart analysis and BER simulation results show that the Bi-SDFE achieves better performance than the original SDFE, whereas the Bi-SDFE uses only half the number of iterations of the SDFE. Therefore, the total computational complexity of the Bi-SDFE is similar to that of the original SDFE, which is a linear function of the channel length, MIMO size, and modulation constellation size. The performance gain of the proposed Bi-SDFE is also verified through an underwater acoustic (UWA) communication experiment with tough channel conditions.

Network-Aware Reference Frame Control for Error-Resilient H.264/AVC Video Streaming Service

To provide high-quality video streaming services in a mobile communication network, a large bandwidth and reliable channel conditions are required. However, mobile communication services still encounter limited bandwidth and varying channel conditions. The streaming video system compresses video with motion estimation and compensation using multiple reference frames. The multiple reference frame structure can reduce the compressed bit rate of video; however, it can also cause significant error propagation when the video in the channel is damaged. Even though the streaming video system includes error-resilience tools to mitigate quality degradation, error propagation is inevitable because all errors can not be refreshed under the multiple reference frame structure. In this paper, a new network-aware error-resilient streaming video system is introduced. The proposed system can mitigate error propagation by controlling the number of reference frames based on channel status. The performance enhancement is demonstrated by comparing the proposed method to the conventional streaming system using static number of reference frames.