Neuromorphic computing is attracting significant attention due to its bio-mimetic characteristics. Consequently, neuromorphic hardware platforms have emerged as innovative computing architectures for acceleration. However, the fixed nature of data flow and resources leads to considerable inefficiencies in storage and computation, thereby limiting both utilization efficiency and overall performance. This severely hinders the deployment of edge artificial intelligence (AI) models. To address these issues, we present a multi-core neuromorphic accelerator named CROSSCUT. This crossbar-based system supports both spiking neural network (SNN) and artificial neural network (ANN) paradigms and has a capacity of 256K neurons and 288M synapses. By leveraging the Neuron Package Mechanism (NPM) and Synapse Compress Mechanism (SCM), CROSSCUT can increase input data scale by 64 times and reduce wasted resources and computations by 46.7%, ensuring high compatibility with diverse network structures in machine learning models. Additionally, a Tree-Mesh hybrid network on chip (NoC) is constructed for inter-core communication. Implemented on Xilinx XCVU9P FPGA, CROSSCUT can achieve a peak performance of 431.9 GSOPS/s and 121.13 GSOPS/W energy efficiency. The inference accuracy on MNIST is 98.2%.

The application of artificial neural network (ANN) in video processing encounters significant challenges, including large data volumes, numerous linear operations, and high power consumption. The fusion of convolutional neural network (CNN) and spiking neural network (SNN) provides a dual benefit of achieving high accuracy while maintaining low power consumption. However, ongoing challenges remain in minimizing multiply-accumulate (MAC) operations and optimizing data movement. In this work, we propose a reconfigurable digital compute-in-memory (RDCIM) heterogeneous macro without the sense amplifier, tailored for the diverse computational demands of CNN and SNN. To improve energy efficiency, differential frame convolution (DFC) is adopted to mitigate the computational overhead. In addition, computational resources are functionally reused to accommodate four data flow types, supporting both DFC and SNN operations. Implemented by TSMC 28nm technology, the proposed RDCIM heterogeneous macro achieves the peak energy efficiency of 29.13 TOPS/W for DFC and 0.56 pJ/SOP for SNN, operating at a frequency of 284 MHz.

To achieve optimal performance of municipal solid waste incineration (MSWI) process with nonstationary time-varying dynamics, a dynamic multi-objective operation optimization method (DSE-TrMOPSO), based on transfer learning, is proposed in this paper. First, the operation optimization models are established using data stream ensemble learning, where incremental updating and selective ensemble strategies are adopted to cope with changing working conditions. Second, a dynamic multi-objective particle swarm optimization algorithm based on transfer learning (CTrDMOPSO) is designed for optimization calculation. In this algorithm, the hierarchical clustering-based transfer learning strategy is proposed to construct the initial population with high quality. Afterwards, the knee point-based decision making is performed to determine the final setpoints of manipulated variables for index optimization. Then, the feasibility of the designed algorithm CTrDMOPSO is verified on the benchmark problems. Finally, the proposed DSE-TrMOPSO is applied to the MSWI process. The results demonstrate that the proposed method can achieve satisfactory operation performance in terms of combustion efficiency and nitrogen oxides emissions. Note to Practitioners—This study aims to develop an optimal operation method for the MSWI process with nonstationary time-varying dynamics. To achieve this goal, an operation optimization method based on transfer learning is proposed, which includes two key points, the operation optimization modeling and the dynamic multi-objective optimization. In practice, practitioners can implement the proposed method utilizing real-time data streams. The optimization objective models are constructed by data stream ensemble learning, and the time-varying dynamics can be captured with incremental updating and selective ensemble strategies. After that, the optimal setpoints of manipulated variables are derived by the designed dynamic multi-objective particle swarm optimization algorithm as well as intelligent decision making. The hierarchical clustering-based transfer learning strategy can deal with stochasticity and uncertainty with high optimization efficiency. The applicability and superiority of the proposed method are verified by the process data of a real MSWI plant. The proposed method provides valuable insights to realize the optimal operation of MSWI process.

Inspired by the brain, Spiking Neural Network (SNN) applies temporally sparse spiking communication to gain more bio-mimetic and highly energy efficient computing. The current mainstream platforms for SNN applications are typically the combination of Host+FPGA+Chip Array, which requires an efficient host to preprocess and encode data. It’s not suitable for end-to-end tasks in edge due to its high system power consumption of host and non-negligible high latency of protocol conversion on FPGA. In addition, Convolutional Neural Network (CNN), exhibits strong feature extraction capabilities. Like the brain's visual system, a hierarchical CNN-SNN hybrid network, in which SNN can make use of CNN’s feature extraction capabilities during encoding, can achieve better performance. In this study, we design a 64Neural-Core Array and integrate it with a CNN encoder and a low-power RISC-V CPU within a System-on-Chip (SoC) to enable comprehensive end-to-end hybrid network application support. The proposed heterogeneous SoC is implemented on a Virtex UltraScale+ XCVU9P FPGA, featuring 32.8K neurons, 37.7M synapses and 578GOPS/s peak performance. It processes MNIST classification with a peak throughput of 2022 images per second at frequency of 250MHz. This design gains a balance between high throughput and recognition accuracy simultaneously.