Xinxiang Huang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Xinxiang Huang has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Bo Liu
Southeast University, Nanjing, Jiangsu, China
2 shared publications - Hao Cai
Southeast University, Nanjing, Jiangsu, China
2 shared publications Shungao Xu
Department of Biochemistry and Molecular Biology, School of Medicine, Jiangsu University, Zhenjiang, China
2 shared publications- Xiumei Sheng
Department of Biochemistry and Molecular Biology, Jiangsu University School of Medicine, Zhenjiang, China
2 shared publications - Yang Zhang
Southeast University
2 shared publications - Chen Zhang
School of Integrated Circuits, Southeast University, Nanjing, China
1 shared publication
Latest Publications
- 1 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (10) , 4652-4664
Layer-Wise Mixed-Modes CNN Processing Architecture With Double-Stationary Dataflow and Dimension-Reshape Strategy
With the development of convolutional neural networks (CNN) across various domains, the growth in network structure complexity and computational load has increasingly become a research focus in the deployment of neural networks. The key to current research on neural network accelerators lies in striking a balance between computational accuracy and energy efficiency. This paper proposes a software-hardware co-design to strike the balance for CNN edge applications. On the hardware side, a 3-dimensional tensor engine (3D-TE), achieved with reconfigurable Tensor Processing Units (TPUs), is introduced for efficient convolution computation. We optimize the CNN dataflow on 3D-TE using a dimension reshaping method for feature maps rearrangement, and a double stationary dataflow scheduling to reduce memory access. This paper adopts a configurable approximate multiplier design based on Boolean Matrix Factorization (BMF) based logic synthesis applied in the architecture of TPU. The proposed 3D-TE, characterized by its configurable precision, enables the TPUs to dynamically adapt the bitwidth of features and weights in response to varying precision requirements. On the software side, a hessian-guided layer precision mapping is adopted to reduce unnecessary computational overhead, and a progressive re-training approach is proposed to enable a better approximation configuration and higher power reduction. Fabricated on 28-nm CMOS, this work achieves an optimized energy efficiency of 14.9 TOPS/W and 12.1 TOPS/W for ResNet56 and MobileNetV2 respectively, with 0.6V supply voltage and 150MHz clock frequency, representing an improvement of $1.33\times\sim8.28\times$ over the state-of-the-art works.
- 1 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (10) , 4652-4664
Layer-Wise Mixed-Modes CNN Processing Architecture With Double-Stationary Dataflow and Dimension-Reshape Strategy
With the development of convolutional neural networks (CNN) across various domains, the growth in network structure complexity and computational load has increasingly become a research focus in the deployment of neural networks. The key to current research on neural network accelerators lies in striking a balance between computational accuracy and energy efficiency. This paper proposes a software-hardware co-design to strike the balance for CNN edge applications. On the hardware side, a 3-dimensional tensor engine (3D-TE), achieved with reconfigurable Tensor Processing Units (TPUs), is introduced for efficient convolution computation. We optimize the CNN dataflow on 3D-TE using a dimension reshaping method for feature maps rearrangement, and a double stationary dataflow scheduling to reduce memory access. This paper adopts a configurable approximate multiplier design based on Boolean Matrix Factorization (BMF) based logic synthesis applied in the architecture of TPU. The proposed 3D-TE, characterized by its configurable precision, enables the TPUs to dynamically adapt the bitwidth of features and weights in response to varying precision requirements. On the software side, a hessian-guided layer precision mapping is adopted to reduce unnecessary computational overhead, and a progressive re-training approach is proposed to enable a better approximation configuration and higher power reduction. Fabricated on 28-nm CMOS, this work achieves an optimized energy efficiency of 14.9 TOPS/W and 12.1 TOPS/W for ResNet56 and MobileNetV2 respectively, with 0.6V supply voltage and 150MHz clock frequency, representing an improvement of $1.33\times\sim8.28\times$ over the state-of-the-art works.
- Open Access
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
FDCA: Fine-grained Digital-CIM based CNN Accelerator with Hybrid Quantization and Weight-Stationary Dataflow
Digital-Compute-in-memory (DCIM) has demonstrated significant energy and area efficiency in convolutional neural network (CNN) accelerators, particularly for high precision applications. However, to mitigate parasitic effects on word and bit lines, most DCIMs employ fine-grained multiply-accumulate operations, which introduces new challenges and opportunities but has not been widely explored. This paper proposes FDCA: a fine-grained digital-CIM based CNN accelerator with hybrid quantization and weight-stationary dataflow, in which the key contributions are :1) a hybrid quantization approach for CNNs leveraging hessian trace and approximation is utilized. This method incorporates the ratio of computation time and storage time into quantization, achieving high efficiency while maintaining accuracy; 2) a Cartesian Genetic Programming based approximate shift and accumulate with error compensation is proposed, where an approximate adder tree is generated to compensate for errors introduced by DCIM; 3) an optimized weight-stationary dataflow is used to improve the utilization of CIM and eliminate dataflow stalls. The experimental results demonstrate that under 28-nm process, when running VGG16 and ResNet50 on CIFAR100, the proposed FDCA achieves 17.1TOPS/W and 18.79TOPS/W with only a slight decrease in accuracy by 0.71% and 0.98%, respectively. Compared to previous works, this work achieves 1.76× and 1.28× better in energy efficiency with less accuracy loss.
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessFDCA: Fine-grained Digital-CIM based CNN Accelerator with Hybrid Quantization and Weight-Stationary Dataflow
Digital-Compute-in-memory (DCIM) has demonstrated significant energy and area efficiency in convolutional neural network (CNN) accelerators, particularly for high precision applications. However, to mitigate parasitic effects on word and bit lines, most DCIMs employ fine-grained multiply-accumulate operations, which introduces new challenges and opportunities but has not been widely explored. This paper proposes FDCA: a fine-grained digital-CIM based CNN accelerator with hybrid quantization and weight-stationary dataflow, in which the key contributions are :1) a hybrid quantization approach for CNNs leveraging hessian trace and approximation is utilized. This method incorporates the ratio of computation time and storage time into quantization, achieving high efficiency while maintaining accuracy; 2) a Cartesian Genetic Programming based approximate shift and accumulate with error compensation is proposed, where an approximate adder tree is generated to compensate for errors introduced by DCIM; 3) an optimized weight-stationary dataflow is used to improve the utilization of CIM and eliminate dataflow stalls. The experimental results demonstrate that under 28-nm process, when running VGG16 and ResNet50 on CIFAR100, the proposed FDCA achieves 17.1TOPS/W and 18.79TOPS/W with only a slight decrease in accuracy by 0.71% and 0.98%, respectively. Compared to previous works, this work achieves 1.76× and 1.28× better in energy efficiency with less accuracy loss.
- Open Access
- 20 September 2022
- Published by Frontiers Media SA in Frontiers in Microbiology
- Vol. 13, 976406
Potential of gut-derived short-chain fatty acids to control enteric pathogens
Short-chain fatty acids (SCFAs) are a very important group of metabolites located in the gut that play a crucial role in the regulation of gut function and pathogen resistance. Since many enteric pathogens respond differently to various SCFAs, substantial efforts have been made to understand the regulatory effects of SCFA types on enteric pathogens. The application of protein post-translational modifications (PTMs) in bacterial research provides a new perspective for studying the regulation of enteric pathogens by different SCFAs. Existing evidence suggests that the SCFAs acetate, propionate, and butyrate influence bacterial processes by extensively promoting the acylation of key bacterial proteins. SCFAs can also prevent the invasion of pathogenic bacteria by regulating the barrier function and immune status of the host gut. In this review, we describe the mechanisms by which different SCFAs modulate the pathogenicity of enteric pathogens from multiple perspectives. We also explore some recent findings on how enteric pathogens counteract SCFA inhibition. Lastly, we discuss the prospects and limitations of applying SCFAs to control enteric pathogens.
- 20 September 2022
- Published by Frontiers Media SA in Frontiers in Microbiology
- Vol. 13, 976406
Open AccessPotential of gut-derived short-chain fatty acids to control enteric pathogens
Short-chain fatty acids (SCFAs) are a very important group of metabolites located in the gut that play a crucial role in the regulation of gut function and pathogen resistance. Since many enteric pathogens respond differently to various SCFAs, substantial efforts have been made to understand the regulatory effects of SCFA types on enteric pathogens. The application of protein post-translational modifications (PTMs) in bacterial research provides a new perspective for studying the regulation of enteric pathogens by different SCFAs. Existing evidence suggests that the SCFAs acetate, propionate, and butyrate influence bacterial processes by extensively promoting the acylation of key bacterial proteins. SCFAs can also prevent the invasion of pathogenic bacteria by regulating the barrier function and immune status of the host gut. In this review, we describe the mechanisms by which different SCFAs modulate the pathogenicity of enteric pathogens from multiple perspectives. We also explore some recent findings on how enteric pathogens counteract SCFA inhibition. Lastly, we discuss the prospects and limitations of applying SCFAs to control enteric pathogens.
- Open Access
- 24 January 2019
- Published by Springer Nature in Scientific Reports
- Vol. 9 (1) , 1-6
Nucleotide-level Convolutional Neural Networks for Pre-miRNA Classification
Due to the biogenesis difference, miRNAs can be divided into canonical microRNAs and mirtrons. Compared to canonical microRNAs, mirtrons are less conserved and hard to be identified. Except stringent annotations based on experiments, many in silico computational methods have be developed to classify miRNAs. Although several machine learning classifiers delivered high classification performance, all the predictors depended heavily on the selection of calculated features. Here, we introduced nucleotide-level convolutional neural networks (CNNs) for pre-miRNAs classification. By using “one-hot” encoding and padding, pre-miRNAs were converted into matrixes with the same shape. The convolution and max-pooling operations can automatically extract features from pre-miRNAs sequences. Evaluation on test dataset showed that our models had a satisfactory performance. Our investigation showed that it was feasible to apply CNNs to extract features from biological sequences. Since there are many hyperparameters can be tuned in CNNs, we believe that the performance of nucleotide-level convolutional neural networks can be greatly improved in the future.
- 24 January 2019
- Published by Springer Nature in Scientific Reports
- Vol. 9 (1) , 1-6
Open AccessNucleotide-level Convolutional Neural Networks for Pre-miRNA Classification
Due to the biogenesis difference, miRNAs can be divided into canonical microRNAs and mirtrons. Compared to canonical microRNAs, mirtrons are less conserved and hard to be identified. Except stringent annotations based on experiments, many in silico computational methods have be developed to classify miRNAs. Although several machine learning classifiers delivered high classification performance, all the predictors depended heavily on the selection of calculated features. Here, we introduced nucleotide-level convolutional neural networks (CNNs) for pre-miRNAs classification. By using “one-hot” encoding and padding, pre-miRNAs were converted into matrixes with the same shape. The convolution and max-pooling operations can automatically extract features from pre-miRNAs sequences. Evaluation on test dataset showed that our models had a satisfactory performance. Our investigation showed that it was feasible to apply CNNs to extract features from biological sequences. Since there are many hyperparameters can be tuned in CNNs, we believe that the performance of nucleotide-level convolutional neural networks can be greatly improved in the future.
- 1 April 2017
- Published by Canadian Science Publishing in Canadian Journal of Microbiology
- Vol. 63 (4) , 287-295
The malS-5′UTR regulates hisG, a key gene in the histidine biosynthetic pathway in Salmonella enterica serovar Typhi
Bacterial noncoding RNAs (ncRNA) regulate diverse cellular processes, including virulence and environmental fitness. The malS 5′ untranslated region (named malS-5′UTR) was identified as a regulatory ncRNA that increases the invasive capacity of Salmonella enterica serovar Typhi. An IntaRNA search suggested base pairing between malS-5′UTR and hisG mRNA, a key gene in the histidine biosynthetic pathway. Overexpression of malS-5′UTR markedly reduced bacterial growth in minimal medium without histidine. Overexpression of malS-5′UTR increased mRNA from his operon genes, independently of the bax gene, and decreased HisG protein in Salmonella Typhi. RNA structure analysis showed base pairing of the malS-5′UTR RNA with the hisG mRNA across the ribosome binding site. Thus, we propose that malS-5′UTR inhibited hisG translation, probably by base pairing to the Shine–Dalgarno sequence.
- 1 April 2017
- Published by Canadian Science Publishing in Canadian Journal of Microbiology
- Vol. 63 (4) , 287-295
The malS-5′UTR regulates hisG, a key gene in the histidine biosynthetic pathway in Salmonella enterica serovar Typhi
Bacterial noncoding RNAs (ncRNA) regulate diverse cellular processes, including virulence and environmental fitness. The malS 5′ untranslated region (named malS-5′UTR) was identified as a regulatory ncRNA that increases the invasive capacity of Salmonella enterica serovar Typhi. An IntaRNA search suggested base pairing between malS-5′UTR and hisG mRNA, a key gene in the histidine biosynthetic pathway. Overexpression of malS-5′UTR markedly reduced bacterial growth in minimal medium without histidine. Overexpression of malS-5′UTR increased mRNA from his operon genes, independently of the bax gene, and decreased HisG protein in Salmonella Typhi. RNA structure analysis showed base pairing of the malS-5′UTR RNA with the hisG mRNA across the ribosome binding site. Thus, we propose that malS-5′UTR inhibited hisG translation, probably by base pairing to the Shine–Dalgarno sequence.
- Open Access
- 17 September 2015
- Published by Frontiers Media SA in Frontiers in Microbiology
- Vol. 6, 990
The novel cis-encoded antisense RNA AsrC positively regulates the expression of rpoE-rseABC operon and thus enhances the motility of Salmonella enterica serovar typhi
Bacterial noncoding RNAs are essential in many cellular processes, including response to environmental stress, and virulence. Deep sequencing analysis of the Salmonella enterica serovar Typhi (S. Typhi) transcriptome revealed a novel antisense RNA transcribed in cis on the strand complementary to rseC, an activator gene of sigma factor RpoE. In this study, expression of this antisense RNA was confirmed in S. Typhi by Northern hybridization. Rapid amplification of cDNA ends and sequence analysis identified an 893 bp sequence from the antisense RNA coding region that covered all of the rseC coding region in the reverse direction of transcription. This sequence of RNA was named as AsrC. After overexpression of AsrC with recombinantant plasmid in S. Typhi, the bacterial motility was increased obviously. To explore the mechanism of AsrC function, regulation of rseC and rpoE expression by AsrC was investigated. We found that AsrC increased the levels of rseC mRNA and protein. The expression of rpoE was also increased in S. Typhi after overexpression of AsrC, which was dependent on rseC. Thus, we propose that AsrC increased RseC level and indirectly activating RpoE which can initiate fliA expression and promote the motility of S. Typhi.
- 17 September 2015
- Published by Frontiers Media SA in Frontiers in Microbiology
- Vol. 6, 990
Open AccessThe novel cis-encoded antisense RNA AsrC positively regulates the expression of rpoE-rseABC operon and thus enhances the motility of Salmonella enterica serovar typhi
Bacterial noncoding RNAs are essential in many cellular processes, including response to environmental stress, and virulence. Deep sequencing analysis of the Salmonella enterica serovar Typhi (S. Typhi) transcriptome revealed a novel antisense RNA transcribed in cis on the strand complementary to rseC, an activator gene of sigma factor RpoE. In this study, expression of this antisense RNA was confirmed in S. Typhi by Northern hybridization. Rapid amplification of cDNA ends and sequence analysis identified an 893 bp sequence from the antisense RNA coding region that covered all of the rseC coding region in the reverse direction of transcription. This sequence of RNA was named as AsrC. After overexpression of AsrC with recombinantant plasmid in S. Typhi, the bacterial motility was increased obviously. To explore the mechanism of AsrC function, regulation of rseC and rpoE expression by AsrC was investigated. We found that AsrC increased the levels of rseC mRNA and protein. The expression of rpoE was also increased in S. Typhi after overexpression of AsrC, which was dependent on rseC. Thus, we propose that AsrC increased RseC level and indirectly activating RpoE which can initiate fliA expression and promote the motility of S. Typhi.