The conventional methodology for designing QC-LDPC decoders is applied for fixed configurations used in wireless communication standards, and the supported largest expansion factor Z (the parallelism of the layered de...The conventional methodology for designing QC-LDPC decoders is applied for fixed configurations used in wireless communication standards, and the supported largest expansion factor Z (the parallelism of the layered decoding) is a fixed number. In this paper, we study the circular-shifting network for decoding LDPC codes with arbitrary Z factor, especially for decoding large Z (Z P) codes, where P is the decoder parallelism. By buffering the P-length slices from the memory, and assembling the shifted slices in a fixed routine, the P-parallelism shift network can process Z-parallelism circular-shifting tasks. The implementation results show that the proposed network for arbitrary sized data shifting consumes only one times of additional resource cost compared to the traditional solution for only maximum P sized data shifting, and achieves significant saving on area and routing complexity.展开更多
Quantum key distribution is increasingly transitioning toward network applications,necessitating advancements in system performance,including photonic integration for compact designs,enhanced stability against environ...Quantum key distribution is increasingly transitioning toward network applications,necessitating advancements in system performance,including photonic integration for compact designs,enhanced stability against environmental disturbances,higher key rates,and improved efficiency.In this letter,we propose an orthogonal polarization exchange reflector Michelson interferometer model to address quantum channel disturbances caused by environmental factors.Based on this model,we designed a Sagnac reflector-Michelson interferometer decoder and verified its performance through an interference system.The interference fringe visibility exceeded 98%across all four coding phases at 625 MHz.These results indicate that the decoder effectively mitigates environmental interference while supporting high-speed modulation frequencies.In addition,the proposed anti-interference decoder,which does not rely on magneto-optical devices,is well-suited for photonic integration,aligning with the development trajectory for next-generation quantum communication devices.展开更多
In this paper,an improved error-rate sliding window decoder is proposed for spatially coupled low-density parity-check(SC-LDPC)codes.For the conventional sliding window decoder,the message retention mechanism causes u...In this paper,an improved error-rate sliding window decoder is proposed for spatially coupled low-density parity-check(SC-LDPC)codes.For the conventional sliding window decoder,the message retention mechanism causes unreliable messages along the edges of belief propagation(BP)decoding in the current window to be kept for subsequent window decoding.To improve the reliability of the retained messages during the window transition,a reliable termination method is embedded,where the retained messages undergo more reliable parity checks.Additionally,decoding failure is unavoidable and even causes error propagation when the number of errors exceeds the error-correcting capability of the window.To mitigate this problem,a channel value reuse mechanism is designed,where the received channel values are utilized to reinitialize the window.Furthermore,considering the complexity and performance of decoding,a feasible sliding optimized window decoding(SOWD)scheme is introduced.Finally,simulation results confirm the superior performance of the proposed SOWD scheme in both the waterfall and error floor regions.This work has great potential in the applications of wireless optical communication and fiber optic communication.展开更多
Quantum computing has the potential to solve complex problems that are inefficiently handled by classical computation.However,the high sensitivity of qubits to environmental interference and the high error rates in cu...Quantum computing has the potential to solve complex problems that are inefficiently handled by classical computation.However,the high sensitivity of qubits to environmental interference and the high error rates in current quantum devices exceed the error correction thresholds required for effective algorithm execution.Therefore,quantum error correction technology is crucial to achieving reliable quantum computing.In this work,we study a topological surface code with a two-dimensional lattice structure that protects quantum information by introducing redundancy across multiple qubits and using syndrome qubits to detect and correct errors.However,errors can occur not only in data qubits but also in syndrome qubits,and different types of errors may generate the same syndromes,complicating the decoding task and creating a need for more efficient decoding methods.To address this challenge,we used a transformer decoder based on an attention mechanism.By mapping the surface code lattice,the decoder performs a self-attention process on all input syndromes,thereby obtaining a global receptive field.The performance of the decoder was evaluated under a phenomenological error model.Numerical results demonstrate that the decoder achieved a decoding accuracy of 93.8%.Additionally,we obtained decoding thresholds of 5%and 6.05%at maximum code distances of 7 and 9,respectively.These results indicate that the decoder used demonstrates a certain capability in correcting noise errors in surface codes.展开更多
Rail surface damage is a critical component of high-speed railway infrastructure,directly affecting train operational stability and safety.Existing methods face limitations in accuracy and speed for small-sample,multi...Rail surface damage is a critical component of high-speed railway infrastructure,directly affecting train operational stability and safety.Existing methods face limitations in accuracy and speed for small-sample,multi-category,and multi-scale target segmentation tasks.To address these challenges,this paper proposes Pyramid-MixNet,an intelligent segmentation model for high-speed rail surface damage,leveraging dataset construction and expansion alongside a feature pyramid-based encoder-decoder network with multi-attention mechanisms.The encoding net-work integrates Spatial Reduction Masked Multi-Head Attention(SRMMHA)to enhance global feature extraction while reducing trainable parameters.The decoding network incorporates Mix-Attention(MA),enabling multi-scale structural understanding and cross-scale token group correlation learning.Experimental results demonstrate that the proposed method achieves 62.17%average segmentation accuracy,80.28%Damage Dice Coefficient,and 56.83 FPS,meeting real-time detection requirements.The model’s high accuracy and scene adaptability significantly improve the detection of small-scale and complex multi-scale rail damage,offering practical value for real-time monitoring in high-speed railway maintenance systems.展开更多
With the rapid development of low altitude economic industry,low altitude adhoc network technology has been getting more and more intensive attention.In the adhoc network protocol designed in this paper,the convolutio...With the rapid development of low altitude economic industry,low altitude adhoc network technology has been getting more and more intensive attention.In the adhoc network protocol designed in this paper,the convolutional code used is(3,1,7),and the design of a low power Viterbi decoder adapted to multi-rate variations is proposed.In the traditional Viterbi decoding method,the high complexity of path metric(PM)accumulation and Euclidean distance computation leads to the problems of low efficiency and large storage resources in the decoder.In this paper,an improved add compare select(ACS)algorithm,a generalized formula for branch metric(BM)based on Manhattan distance,and a method to reduce the accumulated PM for different Viterbi decoders are put forward.A simulation environment based on Vivado and Matlab to verify the accuracy and effectiveness of the proposed Viterbi decoder is also established.The experimental results show that the total power consumption is reduced by 15.58%while the decoding accuracy of the Viterbi decoder is guaranteed,which meets the design requirements of a low power Viterbi decoder.展开更多
Medical image segmentation has witnessed rapid advancements with the emergence of encoder-decoder based methods.In the encoder-decoder structure,the primary goal of the decoding phase is not only to restore feature ma...Medical image segmentation has witnessed rapid advancements with the emergence of encoder-decoder based methods.In the encoder-decoder structure,the primary goal of the decoding phase is not only to restore feature map resolution,but also to mitigate the loss of feature information incurred during the encoding phase.However,this approach gives rise to a challenge:multiple up-sampling operations in the decoder segment result in the loss of feature information.To address this challenge,we propose a novel network that removes the decoding structure to reduce feature information loss(CBL-Net).In particular,we introduce a Parallel Pooling Module(PPM)to counteract the feature information loss stemming from conventional and pooling operations during the encoding stage.Furthermore,we incorporate a Multiplexed Dilation Convolution(MDC)module to expand the network's receptive field.Also,although we have removed the decoding stage,we still need to recover the feature map resolution.Therefore,we introduced the Global Feature Recovery(GFR)module.It uses attention mechanism for the image feature map resolution recovery,which can effectively reduce the loss of feature information.We conduct extensive experimental evaluations on three publicly available medical image segmentation datasets:DRIVE,CHASEDB and MoNuSeg datasets.Experimental results show that our proposed network outperforms state-of-the-art methods in medical image segmentation.In addition,it achieves higher efficiency than the current network of coding and decoding structures by eliminating the decoding component.展开更多
This paper presents a software turbo decoder on graphics processing units(GPU).Unlike previous works,the proposed decoding architecture for turbo codes mainly focuses on the Consultative Committee for Space Data Syste...This paper presents a software turbo decoder on graphics processing units(GPU).Unlike previous works,the proposed decoding architecture for turbo codes mainly focuses on the Consultative Committee for Space Data Systems(CCSDS)standard.However,the information frame lengths of the CCSDS turbo codes are not suitable for flexible sub-frame parallelism design.To mitigate this issue,we propose a padding method that inserts several bits before the information frame header.To obtain low-latency performance and high resource utilization,two-level intra-frame parallelisms and an efficient data structure are considered.The presented Max-Log-Map decoder can be adopted to decode the Long Term Evolution(LTE)turbo codes with only small modifications.The proposed CCSDS turbo decoder at 10 iterations on NVIDIA RTX3070 achieves about 150 Mbps and 50Mbps throughputs for the code rates 1/6 and 1/2,respectively.展开更多
In this paper,we innovatively associate the mutual information with the frame error rate(FER)performance and propose novel quantized decoders for polar codes.Based on the optimal quantizer of binary-input discrete mem...In this paper,we innovatively associate the mutual information with the frame error rate(FER)performance and propose novel quantized decoders for polar codes.Based on the optimal quantizer of binary-input discrete memoryless channels(BDMCs),the proposed decoders quantize the virtual subchannels of polar codes to maximize mutual information(MMI)between source bits and quantized symbols.The nested structure of polar codes ensures that the MMI quantization can be implemented stage by stage.Simulation results show that the proposed MMI decoders with 4 quantization bits outperform the existing nonuniform quantized decoders that minimize mean-squared error(MMSE)with 4 quantization bits,and yield even better performance than uniform MMI quantized decoders with 5 quantization bits.Furthermore,the proposed 5-bit quantized MMI decoders approach the floating-point decoders with negligible performance loss.展开更多
Aiming at the problem that the bit error rate(BER)of asymmetrically clipped optical orthogonal frequency division multiplexing(ACO-OFDM)space optical communication system is significantly affected by different turbule...Aiming at the problem that the bit error rate(BER)of asymmetrically clipped optical orthogonal frequency division multiplexing(ACO-OFDM)space optical communication system is significantly affected by different turbulence intensities,the deep learning technique is proposed to the polarization code decoding in ACO-OFDM space optical communication system.Moreover,this system realizes the polarization code decoding and signal demodulation without frequency conduction with superior performance and robustness compared with the performance of traditional decoder.Simulations under different turbulence intensities as well as different mapping orders show that the convolutional neural network(CNN)decoder trained under weak-medium-strong turbulence atmospheric channels achieves a performance improvement of about 10^(2)compared to the conventional decoder at 4-quadrature amplitude modulation(4QAM),and the BERs for both 16QAM and 64QAM are in between those of the conventional decoder.展开更多
Constituted by BCH component codes and its ordered statistics decoding(OSD),the successive cancellation list(SCL)decoding of U-UV structural codes can provide competent error-correction performance in the short-to-med...Constituted by BCH component codes and its ordered statistics decoding(OSD),the successive cancellation list(SCL)decoding of U-UV structural codes can provide competent error-correction performance in the short-to-medium length regime.However,this list decoding complexity becomes formidable as the decoding output list size increases.This is primarily incurred by the OSD.Addressing this challenge,this paper proposes the low complexity SCL decoding through reducing the complexity of component code decoding,and pruning the redundant SCL decoding paths.For the former,an efficient skipping rule is introduced for the OSD so that the higher order decoding can be skipped when they are not possible to provide a more likely codeword candidate.It is further extended to the OSD variant,the box-andmatch algorithm(BMA),in facilitating the component code decoding.Moreover,through estimating the correlation distance lower bounds(CDLBs)of the component code decoding outputs,a path pruning(PP)-SCL decoding is proposed to further facilitate the decoding of U-UV codes.In particular,its integration with the improved OSD and BMA is discussed.Simulation results show that significant complexity reduction can be achieved.Consequently,the U-UV codes can outperform the cyclic redundancy check(CRC)-polar codes with a similar decoding complexity.展开更多
Neural machine translation(NMT)has advanced with deep learning and large-scale multilingual models,yet translating lowresource languages often lacks sufficient training data and leads to hallucinations.This often resu...Neural machine translation(NMT)has advanced with deep learning and large-scale multilingual models,yet translating lowresource languages often lacks sufficient training data and leads to hallucinations.This often results in translated content that diverges significantly from the source text.This research proposes a refined Contrastive Decoding(CD)algorithm that dynamically adjusts weights of log probabilities from strong expert and weak amateur models to mitigate hallucinations in lowresource NMT and improve translation quality.Advanced large language NMT models,including ChatGLM and LLaMA,are fine-tuned and implemented for their superior contextual understanding and cross-lingual capabilities.The refined CD algorithm evaluates multiple candidate translations using BLEU score,semantic similarity,and Named Entity Recognition accuracy.Extensive experimental results show substantial improvements in translation quality and a significant reduction in hallucination rates.Fine-tuned models achieve higher evaluation metrics compared to baseline models and state-of-the-art models.An ablation study confirms the contributions of each methodological component and highlights the effectiveness of the refined CD algorithm and advanced models in mitigating hallucinations.Notably,the refined methodology increased the BLEU score by approximately 30%compared to baseline models.展开更多
In this paper,Wuzhou City of Guangxi was taken as the research object.Through the design of a climatic data warehousing system,the decoding methods of surface meteorological data and their application in the managemen...In this paper,Wuzhou City of Guangxi was taken as the research object.Through the design of a climatic data warehousing system,the decoding methods of surface meteorological data and their application in the management of climatic data were explored.Based on the parsing technology of the monthly report of surface meteorological records(A-file),the design of Wuzhou climatic data warehousing system was realized,completing the precise extraction and database construction of observational elements such as regional temperature,wind direction,and weather phenomena.Based on this system,the meteorological data in 2024 were analyzed,and the probabilistic characteristics of dominant wind direction in Wuzhou(northeast wind accounting for the largest proportion),the spatiotemporal distribution patterns of extreme temperatures(annual extreme high temperature of 37.1℃in August and extreme low temperature of 1.9℃in January),and the general climatic overview of Wuzhou City(annual precipitation 3.2%higher than the standard value)were revealed.The research shows that climate change has a significant impact on agricultural production and economic development in Wuzhou City,and the construction of a reasonable climatic data database is of great significance for enhancing professional meteorological service capabilities in the context of climate change.This study not only provides a scientific basis for the economic development of Wuzhou region,but also offers reference ideas for other regions to cope with regional climate adaptation planning.展开更多
Transfer RNAs(tRNAs)adopt a stable L-shaped tertiary structure crucial for their involvement in protein translation.Among various divalent metal ions,magnesium ions play a pivotal role in preserving the tertiary struc...Transfer RNAs(tRNAs)adopt a stable L-shaped tertiary structure crucial for their involvement in protein translation.Among various divalent metal ions,magnesium ions play a pivotal role in preserving the tertiary structure of tRNA.However,the precise location of the Mg^(2+)binding pocket in human tRNA remains elusive.In this investigation,we identified the Mg^(2+)binding site within human tRNAGln using suppressor tRNA^(Gln).This variant of tRNA recognizes premature stop codons(specificlly UAG)and facilitates the expression of fll-length proteis.By mutating sites 8 and C72 in supprssr tRNAcl,we assessed the decoding efficiency of the resulting mutant suppressor tRNAs,which serves as a measure of tRNA's ability to decode genetic information.Our analysis revealed that the U8C mutant suppressor tRNA exhibited a significantly lower Mg^(2+)content compared to the C72U mutant.Furthermore,we observed a notable reduction in decoding efficiency in the U8-mutated suppressor tRNA,as evidenced by GFP fluorescence and Western blotting analysis.Conversely,mutations at the C72 site had a comparatively minor impact on decoding efficiency.These findings underscored the tight binding of Mg^(2+)to the U8 site of human tRNAGln,crucial for maintaining the stability of tRNA tertiary structure and translation efficacy.Additionally,our investigation delved into the influence of glutamine availability on tRNA decoding efficiency at the cellular level.The results indicated that both the concentration of amino acids and the codon context of TAG could modulate tRNA decoding efficiency.This study provided valuable insights into the structure and function of tRNA,laying the groundwork for further exploration in this field.展开更多
Action recognition,a fundamental task in the field of video understanding,has been extensively researched and applied.In contrast to an image,a video introduces an extra temporal dimension.However,many existing action...Action recognition,a fundamental task in the field of video understanding,has been extensively researched and applied.In contrast to an image,a video introduces an extra temporal dimension.However,many existing action recognition networks either perform simple temporal fusion through averaging or rely on pre-trained models from image recognition,resulting in limited temporal information extraction capabilities.This work proposes a highly efficient temporal decoding module that can be seamlessly integrated into any action recognition backbone network to enhance the focus on temporal relationships between video frames.Firstly,the decoder initializes a set of learnable queries,termed video-level action category prediction queries.Then,they are combined with the video frame features extracted by the backbone network after self-attention learning to extract video context information.Finally,these prediction queries with rich temporal features are used for category prediction.Experimental results on HMDB51,MSRDailyAct3D,Diving48 and Breakfast datasets show that using TokShift-Transformer and VideoMAE as encoders results in a significant improvement in Top-1 accuracy compared to the original models(TokShift-Transformer and VideoMAE),after introducing the proposed temporal decoder.The introduction of the temporal decoder results in an average performance increase exceeding 11%for TokShift-Transformer and nearly 5%for VideoMAE across the four datasets.Furthermore,the work explores the combination of the decoder with various action recognition networks,including Timesformer,as encoders.This results in an average accuracy improvement of more than 3.5%on the HMDB51 dataset.The code is available at https://github.com/huangturbo/TempDecoder.展开更多
文摘The conventional methodology for designing QC-LDPC decoders is applied for fixed configurations used in wireless communication standards, and the supported largest expansion factor Z (the parallelism of the layered decoding) is a fixed number. In this paper, we study the circular-shifting network for decoding LDPC codes with arbitrary Z factor, especially for decoding large Z (Z P) codes, where P is the decoder parallelism. By buffering the P-length slices from the memory, and assembling the shifted slices in a fixed routine, the P-parallelism shift network can process Z-parallelism circular-shifting tasks. The implementation results show that the proposed network for arbitrary sized data shifting consumes only one times of additional resource cost compared to the traditional solution for only maximum P sized data shifting, and achieves significant saving on area and routing complexity.
基金supported by the National Natural Science Foundation of China under Grant No.62001440。
文摘Quantum key distribution is increasingly transitioning toward network applications,necessitating advancements in system performance,including photonic integration for compact designs,enhanced stability against environmental disturbances,higher key rates,and improved efficiency.In this letter,we propose an orthogonal polarization exchange reflector Michelson interferometer model to address quantum channel disturbances caused by environmental factors.Based on this model,we designed a Sagnac reflector-Michelson interferometer decoder and verified its performance through an interference system.The interference fringe visibility exceeded 98%across all four coding phases at 625 MHz.These results indicate that the decoder effectively mitigates environmental interference while supporting high-speed modulation frequencies.In addition,the proposed anti-interference decoder,which does not rely on magneto-optical devices,is well-suited for photonic integration,aligning with the development trajectory for next-generation quantum communication devices.
基金supported by the National Natural Science Foundation of China (No.62275193)。
文摘In this paper,an improved error-rate sliding window decoder is proposed for spatially coupled low-density parity-check(SC-LDPC)codes.For the conventional sliding window decoder,the message retention mechanism causes unreliable messages along the edges of belief propagation(BP)decoding in the current window to be kept for subsequent window decoding.To improve the reliability of the retained messages during the window transition,a reliable termination method is embedded,where the retained messages undergo more reliable parity checks.Additionally,decoding failure is unavoidable and even causes error propagation when the number of errors exceeds the error-correcting capability of the window.To mitigate this problem,a channel value reuse mechanism is designed,where the received channel values are utilized to reinitialize the window.Furthermore,considering the complexity and performance of decoding,a feasible sliding optimized window decoding(SOWD)scheme is introduced.Finally,simulation results confirm the superior performance of the proposed SOWD scheme in both the waterfall and error floor regions.This work has great potential in the applications of wireless optical communication and fiber optic communication.
基金Project supported by the Natural Science Foundation of Shandong Province,China(Grant No.ZR2021MF049)Joint Fund of Natural Science Foundation of Shandong Province(Grant Nos.ZR2022LLZ012 and ZR2021LLZ001)the Key R&D Program of Shandong Province,China(Grant No.2023CXGC010901)。
文摘Quantum computing has the potential to solve complex problems that are inefficiently handled by classical computation.However,the high sensitivity of qubits to environmental interference and the high error rates in current quantum devices exceed the error correction thresholds required for effective algorithm execution.Therefore,quantum error correction technology is crucial to achieving reliable quantum computing.In this work,we study a topological surface code with a two-dimensional lattice structure that protects quantum information by introducing redundancy across multiple qubits and using syndrome qubits to detect and correct errors.However,errors can occur not only in data qubits but also in syndrome qubits,and different types of errors may generate the same syndromes,complicating the decoding task and creating a need for more efficient decoding methods.To address this challenge,we used a transformer decoder based on an attention mechanism.By mapping the surface code lattice,the decoder performs a self-attention process on all input syndromes,thereby obtaining a global receptive field.The performance of the decoder was evaluated under a phenomenological error model.Numerical results demonstrate that the decoder achieved a decoding accuracy of 93.8%.Additionally,we obtained decoding thresholds of 5%and 6.05%at maximum code distances of 7 and 9,respectively.These results indicate that the decoder used demonstrates a certain capability in correcting noise errors in surface codes.
基金supported in part by the National Natural Science Foundation of China under Grant 6226070954Jiangxi Provincial Key R&D Programme under Grant 20244BBG73002.
文摘Rail surface damage is a critical component of high-speed railway infrastructure,directly affecting train operational stability and safety.Existing methods face limitations in accuracy and speed for small-sample,multi-category,and multi-scale target segmentation tasks.To address these challenges,this paper proposes Pyramid-MixNet,an intelligent segmentation model for high-speed rail surface damage,leveraging dataset construction and expansion alongside a feature pyramid-based encoder-decoder network with multi-attention mechanisms.The encoding net-work integrates Spatial Reduction Masked Multi-Head Attention(SRMMHA)to enhance global feature extraction while reducing trainable parameters.The decoding network incorporates Mix-Attention(MA),enabling multi-scale structural understanding and cross-scale token group correlation learning.Experimental results demonstrate that the proposed method achieves 62.17%average segmentation accuracy,80.28%Damage Dice Coefficient,and 56.83 FPS,meeting real-time detection requirements.The model’s high accuracy and scene adaptability significantly improve the detection of small-scale and complex multi-scale rail damage,offering practical value for real-time monitoring in high-speed railway maintenance systems.
基金Supported by the National Natural Science Foundation of China(No.62103257).
文摘With the rapid development of low altitude economic industry,low altitude adhoc network technology has been getting more and more intensive attention.In the adhoc network protocol designed in this paper,the convolutional code used is(3,1,7),and the design of a low power Viterbi decoder adapted to multi-rate variations is proposed.In the traditional Viterbi decoding method,the high complexity of path metric(PM)accumulation and Euclidean distance computation leads to the problems of low efficiency and large storage resources in the decoder.In this paper,an improved add compare select(ACS)algorithm,a generalized formula for branch metric(BM)based on Manhattan distance,and a method to reduce the accumulated PM for different Viterbi decoders are put forward.A simulation environment based on Vivado and Matlab to verify the accuracy and effectiveness of the proposed Viterbi decoder is also established.The experimental results show that the total power consumption is reduced by 15.58%while the decoding accuracy of the Viterbi decoder is guaranteed,which meets the design requirements of a low power Viterbi decoder.
基金funded by the National Key Research and Development Program of China(Grant 2020YFB1708900)the Fundamental Research Funds for the Central Universities(Grant No.B220201044).
文摘Medical image segmentation has witnessed rapid advancements with the emergence of encoder-decoder based methods.In the encoder-decoder structure,the primary goal of the decoding phase is not only to restore feature map resolution,but also to mitigate the loss of feature information incurred during the encoding phase.However,this approach gives rise to a challenge:multiple up-sampling operations in the decoder segment result in the loss of feature information.To address this challenge,we propose a novel network that removes the decoding structure to reduce feature information loss(CBL-Net).In particular,we introduce a Parallel Pooling Module(PPM)to counteract the feature information loss stemming from conventional and pooling operations during the encoding stage.Furthermore,we incorporate a Multiplexed Dilation Convolution(MDC)module to expand the network's receptive field.Also,although we have removed the decoding stage,we still need to recover the feature map resolution.Therefore,we introduced the Global Feature Recovery(GFR)module.It uses attention mechanism for the image feature map resolution recovery,which can effectively reduce the loss of feature information.We conduct extensive experimental evaluations on three publicly available medical image segmentation datasets:DRIVE,CHASEDB and MoNuSeg datasets.Experimental results show that our proposed network outperforms state-of-the-art methods in medical image segmentation.In addition,it achieves higher efficiency than the current network of coding and decoding structures by eliminating the decoding component.
基金supported by the Fundamental Research Funds for the Central Universities(FRF-TP20-062A1)Guangdong Basic and Applied Basic Research Foundation(2021A1515110070)。
文摘This paper presents a software turbo decoder on graphics processing units(GPU).Unlike previous works,the proposed decoding architecture for turbo codes mainly focuses on the Consultative Committee for Space Data Systems(CCSDS)standard.However,the information frame lengths of the CCSDS turbo codes are not suitable for flexible sub-frame parallelism design.To mitigate this issue,we propose a padding method that inserts several bits before the information frame header.To obtain low-latency performance and high resource utilization,two-level intra-frame parallelisms and an efficient data structure are considered.The presented Max-Log-Map decoder can be adopted to decode the Long Term Evolution(LTE)turbo codes with only small modifications.The proposed CCSDS turbo decoder at 10 iterations on NVIDIA RTX3070 achieves about 150 Mbps and 50Mbps throughputs for the code rates 1/6 and 1/2,respectively.
基金financially supported in part by National Key R&D Program of China(No.2018YFB1801402)in part by Huawei Technologies Co.,Ltd.
文摘In this paper,we innovatively associate the mutual information with the frame error rate(FER)performance and propose novel quantized decoders for polar codes.Based on the optimal quantizer of binary-input discrete memoryless channels(BDMCs),the proposed decoders quantize the virtual subchannels of polar codes to maximize mutual information(MMI)between source bits and quantized symbols.The nested structure of polar codes ensures that the MMI quantization can be implemented stage by stage.Simulation results show that the proposed MMI decoders with 4 quantization bits outperform the existing nonuniform quantized decoders that minimize mean-squared error(MMSE)with 4 quantization bits,and yield even better performance than uniform MMI quantized decoders with 5 quantization bits.Furthermore,the proposed 5-bit quantized MMI decoders approach the floating-point decoders with negligible performance loss.
基金supported by the National Natural Science Foundation of China(No.12104141).
文摘Aiming at the problem that the bit error rate(BER)of asymmetrically clipped optical orthogonal frequency division multiplexing(ACO-OFDM)space optical communication system is significantly affected by different turbulence intensities,the deep learning technique is proposed to the polarization code decoding in ACO-OFDM space optical communication system.Moreover,this system realizes the polarization code decoding and signal demodulation without frequency conduction with superior performance and robustness compared with the performance of traditional decoder.Simulations under different turbulence intensities as well as different mapping orders show that the convolutional neural network(CNN)decoder trained under weak-medium-strong turbulence atmospheric channels achieves a performance improvement of about 10^(2)compared to the conventional decoder at 4-quadrature amplitude modulation(4QAM),and the BERs for both 16QAM and 64QAM are in between those of the conventional decoder.
基金supported by the National Natural Science Foundation of China(NSFC)with project ID 62071498the Guangdong National Science Foundation(GDNSF)with project ID 2024A1515010213.
文摘Constituted by BCH component codes and its ordered statistics decoding(OSD),the successive cancellation list(SCL)decoding of U-UV structural codes can provide competent error-correction performance in the short-to-medium length regime.However,this list decoding complexity becomes formidable as the decoding output list size increases.This is primarily incurred by the OSD.Addressing this challenge,this paper proposes the low complexity SCL decoding through reducing the complexity of component code decoding,and pruning the redundant SCL decoding paths.For the former,an efficient skipping rule is introduced for the OSD so that the higher order decoding can be skipped when they are not possible to provide a more likely codeword candidate.It is further extended to the OSD variant,the box-andmatch algorithm(BMA),in facilitating the component code decoding.Moreover,through estimating the correlation distance lower bounds(CDLBs)of the component code decoding outputs,a path pruning(PP)-SCL decoding is proposed to further facilitate the decoding of U-UV codes.In particular,its integration with the improved OSD and BMA is discussed.Simulation results show that significant complexity reduction can be achieved.Consequently,the U-UV codes can outperform the cyclic redundancy check(CRC)-polar codes with a similar decoding complexity.
基金M.Faheem is supported by VTT Technical Research Center of Finland.
文摘Neural machine translation(NMT)has advanced with deep learning and large-scale multilingual models,yet translating lowresource languages often lacks sufficient training data and leads to hallucinations.This often results in translated content that diverges significantly from the source text.This research proposes a refined Contrastive Decoding(CD)algorithm that dynamically adjusts weights of log probabilities from strong expert and weak amateur models to mitigate hallucinations in lowresource NMT and improve translation quality.Advanced large language NMT models,including ChatGLM and LLaMA,are fine-tuned and implemented for their superior contextual understanding and cross-lingual capabilities.The refined CD algorithm evaluates multiple candidate translations using BLEU score,semantic similarity,and Named Entity Recognition accuracy.Extensive experimental results show substantial improvements in translation quality and a significant reduction in hallucination rates.Fine-tuned models achieve higher evaluation metrics compared to baseline models and state-of-the-art models.An ablation study confirms the contributions of each methodological component and highlights the effectiveness of the refined CD algorithm and advanced models in mitigating hallucinations.Notably,the refined methodology increased the BLEU score by approximately 30%compared to baseline models.
基金Supported by the Fifth Batch of Innovation Teams of Wuzhou Meteorological Bureau“Wuzhou Innovation Team for Enhancing the Comprehensive Meteorological Observation Ability through Digitization and Intelligence”Project of Wuzhou Science and Technology Bureau(202402122)Wuzhou Science and Technology Planning Project(202402119).
文摘In this paper,Wuzhou City of Guangxi was taken as the research object.Through the design of a climatic data warehousing system,the decoding methods of surface meteorological data and their application in the management of climatic data were explored.Based on the parsing technology of the monthly report of surface meteorological records(A-file),the design of Wuzhou climatic data warehousing system was realized,completing the precise extraction and database construction of observational elements such as regional temperature,wind direction,and weather phenomena.Based on this system,the meteorological data in 2024 were analyzed,and the probabilistic characteristics of dominant wind direction in Wuzhou(northeast wind accounting for the largest proportion),the spatiotemporal distribution patterns of extreme temperatures(annual extreme high temperature of 37.1℃in August and extreme low temperature of 1.9℃in January),and the general climatic overview of Wuzhou City(annual precipitation 3.2%higher than the standard value)were revealed.The research shows that climate change has a significant impact on agricultural production and economic development in Wuzhou City,and the construction of a reasonable climatic data database is of great significance for enhancing professional meteorological service capabilities in the context of climate change.This study not only provides a scientific basis for the economic development of Wuzhou region,but also offers reference ideas for other regions to cope with regional climate adaptation planning.
基金National Natural Science Foundation of China(Grant No.U23A20106)National Key Research and Development Program of China(Grant No.91510100MA6CG8UJ4K)。
文摘Transfer RNAs(tRNAs)adopt a stable L-shaped tertiary structure crucial for their involvement in protein translation.Among various divalent metal ions,magnesium ions play a pivotal role in preserving the tertiary structure of tRNA.However,the precise location of the Mg^(2+)binding pocket in human tRNA remains elusive.In this investigation,we identified the Mg^(2+)binding site within human tRNAGln using suppressor tRNA^(Gln).This variant of tRNA recognizes premature stop codons(specificlly UAG)and facilitates the expression of fll-length proteis.By mutating sites 8 and C72 in supprssr tRNAcl,we assessed the decoding efficiency of the resulting mutant suppressor tRNAs,which serves as a measure of tRNA's ability to decode genetic information.Our analysis revealed that the U8C mutant suppressor tRNA exhibited a significantly lower Mg^(2+)content compared to the C72U mutant.Furthermore,we observed a notable reduction in decoding efficiency in the U8-mutated suppressor tRNA,as evidenced by GFP fluorescence and Western blotting analysis.Conversely,mutations at the C72 site had a comparatively minor impact on decoding efficiency.These findings underscored the tight binding of Mg^(2+)to the U8 site of human tRNAGln,crucial for maintaining the stability of tRNA tertiary structure and translation efficacy.Additionally,our investigation delved into the influence of glutamine availability on tRNA decoding efficiency at the cellular level.The results indicated that both the concentration of amino acids and the codon context of TAG could modulate tRNA decoding efficiency.This study provided valuable insights into the structure and function of tRNA,laying the groundwork for further exploration in this field.
基金Shanghai Municipal Commission of Economy and Information Technology,China (No.202301054)。
文摘Action recognition,a fundamental task in the field of video understanding,has been extensively researched and applied.In contrast to an image,a video introduces an extra temporal dimension.However,many existing action recognition networks either perform simple temporal fusion through averaging or rely on pre-trained models from image recognition,resulting in limited temporal information extraction capabilities.This work proposes a highly efficient temporal decoding module that can be seamlessly integrated into any action recognition backbone network to enhance the focus on temporal relationships between video frames.Firstly,the decoder initializes a set of learnable queries,termed video-level action category prediction queries.Then,they are combined with the video frame features extracted by the backbone network after self-attention learning to extract video context information.Finally,these prediction queries with rich temporal features are used for category prediction.Experimental results on HMDB51,MSRDailyAct3D,Diving48 and Breakfast datasets show that using TokShift-Transformer and VideoMAE as encoders results in a significant improvement in Top-1 accuracy compared to the original models(TokShift-Transformer and VideoMAE),after introducing the proposed temporal decoder.The introduction of the temporal decoder results in an average performance increase exceeding 11%for TokShift-Transformer and nearly 5%for VideoMAE across the four datasets.Furthermore,the work explores the combination of the decoder with various action recognition networks,including Timesformer,as encoders.This results in an average accuracy improvement of more than 3.5%on the HMDB51 dataset.The code is available at https://github.com/huangturbo/TempDecoder.
