In recent years, the complexity of neural network (NN) models has increase continuously, resulting in a substantial increase in multiply-accumulate (MAC) operations. Therefore, the need for area- and power-efficient DRAM compute-in-memory (CIM) designs has become a pressing issue. The proposed CIM design uses the reverse rotating charge sharing (RRCS) technique to save 79% of silicon area compared to conventional designs. In addition, the use of the proposed low-threshold prediction technique can reduce analog-to-digital converter (ADC) calculations by 25% and ADC power consumption by 17%. Energy efficiency and area reduction are 10x and 1900x, respectively, higher than previous CIM designs.

This paper proposes a flying-capacitor sharing dynamic four-path (FCSD-4P) hybrid converter, designed to achieve a duty cycle (D) ranging from 0% to 100%, addressing the conventional limitation of D2. In a 6V to 1.8V conversion, the converter achieves a peak efficiency of 94.2%.

This paper proposes a two-step down converter based on Pseudo 4-Phase Mutual Operation (P4MO) that uses inductors and coupled inductors to regulate output voltage (VOUT). By doubling the inductor current slew rate, the design reduces the intermediate capacitor (CINT) at VINTby nearly 50%. Additionally, conduction losses in the dual-output triple step-down (DOTS) input stage are reduced by approximately 30%. Moreover, a delayed feedback PWM technique further reduces the VINTripple by about 20%. An integrated soft turn-off circuit minimizes voltage variations at VOUTto less than 2 mV during phase shedding. Experimental results validate the design’s performance, showing a peak efficiency of 93.1%, undershoot of 106 mV with a recovery time of 1.9 μs for a load step transition from 1 A to 20 A, and a transient performance of 5.58 mV/A.

The fifth-generation mobile network (5G) integrates various application services in a heterogeneous network environment. Compared to the traditional networks, 5G is not just an extension of the 4th generation, which contains three important properties, enhanced mobile broadband (eMBB), massive machine type communications (mMTC), and ultra-reliable and low latency communications (URLLC). 5G applies the functionalities of Network Function Virtualization and Software-Defined Networking to support multiple services and proposes a new concept called Network Slicing. Users can access different services quickly in the 5G network supported by network slicing. In a traditional network like 4G, if a user wants to access different services, it will be necessary to perform different authentication procedures that cause additional burden and operation cost in the user's device. However, the 5G network inherits the previous network architecture. Hence, the user's device still needs to be authenticated by the core network. Besides, providing a guarantee of connecting to a correct network slice is one of the prime concerns. The paper presents an authentication scheme tailored for the 5G network. In the proposed scheme, the authentication is decentralized to the edge clouds to achieve low latency. Moreover, the authentication flow is no longer attached to the operator all the time to reduce time latency. The proposed scheme is secure against the attackers who aim to impersonate users, network operators, or even network slices, and it also provides secure session key exchange. Empirical performance assessment in terms of its functionalities gains better acceptability of the proposed scheme than other existing ones.