Bo Yuan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Bo Yuan has a total of 180 co-authors. Here are the top co-authors by number of shared publications:
- Qingli Wu
Department of Plant Biology, Rutgers, The State University of New Jersey, New Brunswick, NJ 08901, USA
9 shared publications Danyue Zhao
New Use Agriculture & Natural Plant Products Program, Department of Plant Biology, School of Environmental and Biological Sciences, Rutgers University, New Brunswick, NJ, 08901, USA
6 shared publications- James Edward Simon
New Use Agriculture and Natural Plant Products, Department of Plant Biology, Rutgers University
6 shared publications - Yang Sui
Department of Electrical and Computer Engineering, Rutgers, the State University of New Jersey, Piscataway, USA
6 shared publications - Huy Phan
Rutgers University
5 shared publications - Miao Yin
Rutgers University
5 shared publications
Latest Publications
- Top Cited Papers
- 12 February 2026
- Published by American Association for the Advancement of Science (AAAS) in Science
- Vol. 391 (6786) , eady5532
Mitochondrial control of fuel switching via carnitine biosynthesis
Environmental adaptation often involves a shift in energy utilization toward mitochondrial fatty acid oxidation, which requires carnitine. Besides dietary sources of animal origin, carnitine biosynthesis from trimethyllysine (TML) is essential, particularly for those who consume plant-based diets; however, its molecular regulation and physiological role remain elusive. Here, we identify SLC25A45 as a mitochondrial TML carrier that controls carnitine biosynthesis and fuel switching. SLC25A45 deficiency decreased the carnitine pool and impaired mitochondrial fatty acid oxidation, shifting reliance to carbohydrate metabolism. Slc25a45 -deficient mice were cold-intolerant and resistant to lipid mobilization by glucagon-like peptide-1 receptor agonist (GLP-1RA), rendering them resistant to adipose tissue loss. Our study suggests that mitochondria serve as a regulatory checkpoint in fuel switching, with implications for metabolic adaptation and the efficacy of GLP-1RA–based anti-obesity therapy.
- 12 February 2026
- Published by American Association for the Advancement of Science (AAAS) in Science
- Vol. 391 (6786) , eady5532
Top Cited PapersMitochondrial control of fuel switching via carnitine biosynthesis
Environmental adaptation often involves a shift in energy utilization toward mitochondrial fatty acid oxidation, which requires carnitine. Besides dietary sources of animal origin, carnitine biosynthesis from trimethyllysine (TML) is essential, particularly for those who consume plant-based diets; however, its molecular regulation and physiological role remain elusive. Here, we identify SLC25A45 as a mitochondrial TML carrier that controls carnitine biosynthesis and fuel switching. SLC25A45 deficiency decreased the carnitine pool and impaired mitochondrial fatty acid oxidation, shifting reliance to carbohydrate metabolism. Slc25a45 -deficient mice were cold-intolerant and resistant to lipid mobilization by glucagon-like peptide-1 receptor agonist (GLP-1RA), rendering them resistant to adipose tissue loss. Our study suggests that mitochondria serve as a regulatory checkpoint in fuel switching, with implications for metabolic adaptation and the efficacy of GLP-1RA–based anti-obesity therapy.
- Open Access
Mitochondrial control of fuel switching via carnitine biosynthesis
Metabolic adaptation to environmental changes, such as fasting and cold exposure, involves a dynamic shift in fuel utilization from glucose to fatty acid oxidation, a process that relies on carnitine-mediated fatty acid oxidation in mitochondria. While dietary sources of animal origin (e.g., red meat) contribute to the carnitine pool, de novo carnitine synthesis from trimethyllysine (TML) is essential, particularly for those whose dietary sources are vegetables and fruits that contain negligible amounts of carnitine. However, the molecular pathway of de novo carnitine synthesis and its physiological significance remain poorly understood. Here, we showed that SLC25A45 is a mitochondrial TML carrier that controls de novo carnitine biosynthesis in vivo. Genetic loss of SLC25A45 results in systemic carnitine and acylcarnitine deficiency, leading to impaired fatty acid oxidation and thermogenesis during cold adaptation, while promoting glucose catabolism. Notably, Slc25a45-deficient mice maintained a high respiratory exchange ratio and impaired lipid mobilization following treatment with a GLP1 receptor agonist (GLP-1RA), rendering them resistant to GLP-1RA-induced adipose tissue loss. Together, the present study identifies SLC25A45 as a regulatory checkpoint in fuel switching during adaptation, with implications for systemic energy balance and response to GLP-1RA-mediated anti-obesity therapy.
Open AccessMitochondrial control of fuel switching via carnitine biosynthesis
Metabolic adaptation to environmental changes, such as fasting and cold exposure, involves a dynamic shift in fuel utilization from glucose to fatty acid oxidation, a process that relies on carnitine-mediated fatty acid oxidation in mitochondria. While dietary sources of animal origin (e.g., red meat) contribute to the carnitine pool, de novo carnitine synthesis from trimethyllysine (TML) is essential, particularly for those whose dietary sources are vegetables and fruits that contain negligible amounts of carnitine. However, the molecular pathway of de novo carnitine synthesis and its physiological significance remain poorly understood. Here, we showed that SLC25A45 is a mitochondrial TML carrier that controls de novo carnitine biosynthesis in vivo. Genetic loss of SLC25A45 results in systemic carnitine and acylcarnitine deficiency, leading to impaired fatty acid oxidation and thermogenesis during cold adaptation, while promoting glucose catabolism. Notably, Slc25a45-deficient mice maintained a high respiratory exchange ratio and impaired lipid mobilization following treatment with a GLP1 receptor agonist (GLP-1RA), rendering them resistant to GLP-1RA-induced adipose tissue loss. Together, the present study identifies SLC25A45 as a regulatory checkpoint in fuel switching during adaptation, with implications for systemic energy balance and response to GLP-1RA-mediated anti-obesity therapy.
- Open Access
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1250-1263
Crane: Inter-Layer Scheduling Framework for DNN Inference and Training Co-Support on Tiled Architecture
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1250-1263
Open AccessCrane: Inter-Layer Scheduling Framework for DNN Inference and Training Co-Support on Tiled Architecture
- 1 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (4) , 1649-1662
TT-RRAM: Joint Improvement of Sparsity and Repetition in Tensor-Train Inference With Fused Processing on RRAM Accelerators
Deep neural networks compressed using tensor-train decomposition, known as TT-DNNs, significantly reduce model size to decrease storage requirements but suffer from more frequent data movement during inference. To address this issue, resistive random access memory (RRAM)-based deep neural network (DNN) accelerators, by leveraging computing-in-memory, can mitigate the data movement and efficiently execute the vector–matrix multiplication (VMM) operations required during inference, making them an ideal candidate for accelerating the inference for TT-DNNs. However, RRAM-based accelerators face three challenges in TT-format DNN inference: 1) weights distribution on the RRAM crossbar with low sparsity and repetition; 2) low input sparsity; and 3) high data dependencies and storage overhead during inference. To address these challenges, we propose three corresponding methods: 1) base-offset splitting (BOS) to improve the weights distribution on the crossbar, thereby enhancing performance potential; 2) inverse-input reusing (IIR) to enhance the sparsity of inputs during inference; and 3) fused scheduling (FS) to reorganize the computation order and leveraging the parallel processing capabilities of the RRAM-based accelerator to improve overall inference efficiency while also reducing data dependencies and storage overhead. Experimental results demonstrate that our proposed methods achieve a$3.44\times $ –$5.53\times $ performance improvement and 55.6%–67.1% energy savings compared with the state-of-the-art RRAM-based accelerator, while also reducing storage overhead by 66.7%–99.1%.- 1 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (4) , 1649-1662
TT-RRAM: Joint Improvement of Sparsity and Repetition in Tensor-Train Inference With Fused Processing on RRAM Accelerators
Deep neural networks compressed using tensor-train decomposition, known as TT-DNNs, significantly reduce model size to decrease storage requirements but suffer from more frequent data movement during inference. To address this issue, resistive random access memory (RRAM)-based deep neural network (DNN) accelerators, by leveraging computing-in-memory, can mitigate the data movement and efficiently execute the vector–matrix multiplication (VMM) operations required during inference, making them an ideal candidate for accelerating the inference for TT-DNNs. However, RRAM-based accelerators face three challenges in TT-format DNN inference: 1) weights distribution on the RRAM crossbar with low sparsity and repetition; 2) low input sparsity; and 3) high data dependencies and storage overhead during inference. To address these challenges, we propose three corresponding methods: 1) base-offset splitting (BOS) to improve the weights distribution on the crossbar, thereby enhancing performance potential; 2) inverse-input reusing (IIR) to enhance the sparsity of inputs during inference; and 3) fused scheduling (FS) to reorganize the computation order and leveraging the parallel processing capabilities of the RRAM-based accelerator to improve overall inference efficiency while also reducing data dependencies and storage overhead. Experimental results demonstrate that our proposed methods achieve a$3.44\times $ –$5.53\times $ performance improvement and 55.6%–67.1% energy savings compared with the state-of-the-art RRAM-based accelerator, while also reducing storage overhead by 66.7%–99.1%.- Open Access
- 6 August 2025
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 97 (5-6) , 247-256
Pruning 3D Convolutional Neural Networks via Channel Independence
Network pruning provides a promising approach for deploying costly Deep Neural Network (DNN) models on resource-constrained devices. However, most existing pruning works focus on compressing traditional convolutional DNN models for image-related classification and object-detection tasks in 2D scenarios. Due to the complicated structure of 3D CNN models, pruning such models has not been well studied. In this paper, we analyze the different properties between 2D and 3D tasks, then propose a Filter/Depth-wise Independence Score (FDIS) to evaluate the importance of each filter. In addition, we adopt several granularity schemes to improve the performance of the proposed method. To achieve fine-grained pruning, we prune the networks gradually using an iterative pruning procedure. In addition, we experimentally show that weights with low independence scores contain less important information, enabling the removal of filters without serious accuracy degradation. Our proposed FDIS-based approach maintains high accuracy with certain FLOP reduction and practical acceleration.
- 6 August 2025
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 97 (5-6) , 247-256
Open AccessPruning 3D Convolutional Neural Networks via Channel Independence
Network pruning provides a promising approach for deploying costly Deep Neural Network (DNN) models on resource-constrained devices. However, most existing pruning works focus on compressing traditional convolutional DNN models for image-related classification and object-detection tasks in 2D scenarios. Due to the complicated structure of 3D CNN models, pruning such models has not been well studied. In this paper, we analyze the different properties between 2D and 3D tasks, then propose a Filter/Depth-wise Independence Score (FDIS) to evaluate the importance of each filter. In addition, we adopt several granularity schemes to improve the performance of the proposed method. To achieve fine-grained pruning, we prune the networks gradually using an iterative pruning procedure. In addition, we experimentally show that weights with low independence scores contain less important information, enabling the removal of filters without serious accuracy degradation. Our proposed FDIS-based approach maintains high accuracy with certain FLOP reduction and practical acceleration.
- Open Access
- 10 January 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
- Vol. 30 (2) , 1-21
Global Placement Exploiting Soft 2D Regularity
Cell placement is a step of paramount importance in chip physical design and requests relentless effort for continuous improvement. Recently, designs with two-dimensional (2D) processing element arrays have become popular primarily due to their deep neural network hardware applications. The 2D array regularity is similar to but different from the regularity of conventional datapath designs. To exploit the 2D array regularity, this work develops a new global placement technique, Placement of Arrays with SOft Regularity (PASOR), built upon RePlAce, the state-of-the-art placement framework. Experimental results from various designs show that the proposed approach can reduce global routing wirelength by 11% and 6% compared to RePlAce and a previous work on datapath driven placement, respectively.
- 10 January 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
- Vol. 30 (2) , 1-21
Open AccessGlobal Placement Exploiting Soft 2D Regularity
Cell placement is a step of paramount importance in chip physical design and requests relentless effort for continuous improvement. Recently, designs with two-dimensional (2D) processing element arrays have become popular primarily due to their deep neural network hardware applications. The 2D array regularity is similar to but different from the regularity of conventional datapath designs. To exploit the 2D array regularity, this work develops a new global placement technique, Placement of Arrays with SOft Regularity (PASOR), built upon RePlAce, the state-of-the-art placement framework. Experimental results from various designs show that the proposed approach can reduce global routing wirelength by 11% and 6% compared to RePlAce and a previous work on datapath driven placement, respectively.
- 26 November 2024
- Published by Springer Nature
- p. 273-290
Clean and Compact: Efficient Data-Free Backdoor Defense with Model Compactness
- 26 November 2024
- Published by Springer Nature
- p. 273-290
Clean and Compact: Efficient Data-Free Backdoor Defense with Model Compactness
- Open Access
MoE-I$^2$: Compressing Mixture of Experts Models through Inter-Expert Pruning and Intra-Expert Low-Rank Decomposition
The emergence of Mixture of Experts (MoE) LLMs has significantly advanced the development of language models. Compared to traditional LLMs, MoE LLMs outperform traditional LLMs by achieving higher performance with considerably fewer activated parameters. Despite this efficiency, their enormous parameter size still leads to high deployment costs. In this paper, we introduce a two-stage compression method tailored for MoE to reduce the model size and decrease the computational cost. First, in the inter-expert pruning stage, we analyze the importance of each layer and propose the Layer-wise Genetic Search and Block-wise KT-Reception Field with the non-uniform pruning ratio to prune the individual expert. Second, in the intra-expert decomposition stage, we apply the low-rank decomposition to further compress the parameters within the remaining experts. Extensive experiments on Qwen1.5-MoE-A2.7B, DeepSeek-V2-Lite, and Mixtral-8$\times$7B demonstrate that our proposed methods can both reduce the model size and enhance inference efficiency while maintaining performance in various zero-shot tasks. The code will be available at \url{https://github.com/xiaochengsky/MoEI-2.git}
Open AccessMoE-I$^2$: Compressing Mixture of Experts Models through Inter-Expert Pruning and Intra-Expert Low-Rank Decomposition
The emergence of Mixture of Experts (MoE) LLMs has significantly advanced the development of language models. Compared to traditional LLMs, MoE LLMs outperform traditional LLMs by achieving higher performance with considerably fewer activated parameters. Despite this efficiency, their enormous parameter size still leads to high deployment costs. In this paper, we introduce a two-stage compression method tailored for MoE to reduce the model size and decrease the computational cost. First, in the inter-expert pruning stage, we analyze the importance of each layer and propose the Layer-wise Genetic Search and Block-wise KT-Reception Field with the non-uniform pruning ratio to prune the individual expert. Second, in the intra-expert decomposition stage, we apply the low-rank decomposition to further compress the parameters within the remaining experts. Extensive experiments on Qwen1.5-MoE-A2.7B, DeepSeek-V2-Lite, and Mixtral-8$\times$7B demonstrate that our proposed methods can both reduce the model size and enhance inference efficiency while maintaining performance in various zero-shot tasks. The code will be available at \url{https://github.com/xiaochengsky/MoEI-2.git}
- Open Access
- 27 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-8
SysMix: Mixed-Size Placement for Systolic-Array-Based Hierarchical Designs
- 27 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessSysMix: Mixed-Size Placement for Systolic-Array-Based Hierarchical Designs
- 16 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (1) , 251-265
DiMO-CNN: Deep Learning Toolkit-Accelerated Analytical Modeling and Optimization of CNN Hardware and Dataflow
The growing complexity of CNNs demands both hardware acceleration design and dataflow mapping solutions. The large co-design solution space presents a huge challenge. We introduce an analytical model for assessing CNN hardware design and dataflow solutions, using a matrix-based approach. Our co-optimization method, combining nonlinear programming and parallel local search, excels in addressing the power-performance-area tradeoff. The average relative error of our analytical model compared with Timeloop is as small as 1%. Compared to state-of-the-art methods, our co-optimization achieves solutions with average$3.14\times $ shorter inference latency,$\mathbf {68.2\%}$ less power consumption, and$\mathbf {74\%}$ less area on all testcases. It also provides a$200\times $ speedup of optimization runtime.- 16 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (1) , 251-265
DiMO-CNN: Deep Learning Toolkit-Accelerated Analytical Modeling and Optimization of CNN Hardware and Dataflow
The growing complexity of CNNs demands both hardware acceleration design and dataflow mapping solutions. The large co-design solution space presents a huge challenge. We introduce an analytical model for assessing CNN hardware design and dataflow solutions, using a matrix-based approach. Our co-optimization method, combining nonlinear programming and parallel local search, excels in addressing the power-performance-area tradeoff. The average relative error of our analytical model compared with Timeloop is as small as 1%. Compared to state-of-the-art methods, our co-optimization achieves solutions with average$3.14\times $ shorter inference latency,$\mathbf {68.2\%}$ less power consumption, and$\mathbf {74\%}$ less area on all testcases. It also provides a$200\times $ speedup of optimization runtime.