Swarm intelligence-based packet scheduling for future intelligent networks.

Network operations involve several decision-making tasks. Some of these tasks are related to operators, such as extending the footprint or upgrading the network capacity. Other decision tasks are related to network functions, such as traffic classifications, scheduling, capacity, coverage trade-offs, and policy enforcement. These decisions are often decentralized, and each network node makes its own decisions based on the preconfigured rules or policies. To ensure effectiveness, it is essential that planning and functional decisions are in harmony. However, human intervention-based decisions are subject to high costs, delays, and mistakes. On the other hand, machine learning has been used in different fields of life to automate decision processes intelligently. Similarly, future intelligent networks are also expected to see an intense use of machine learning and artificial intelligence techniques for functional and operational automation. This article investigates the current state-of-the-art methods for packet scheduling and related decision processes. Furthermore, it proposes a machine learning-based approach for packet scheduling for agile and cost-effective networks to address various issues and challenges. The analysis of the experimental results shows that the proposed deep learning-based approach can successfully address the challenges without compromising the network performance. For example, it has been seen that with mean absolute error from 6.38 to 8.41 using the proposed deep learning model, the packet scheduling can maintain 99.95% throughput, 99.97% delay, and 99.94% jitter, which are much better as compared to the statically configured traffic profiles.

Adenosine Triphosphate Protects from Elevated Extracellular Calcium-Induced Damage in Human Proximal Kidney Cells: Using Deep Learning to Predict Cytotoxicity.

BACKGROUND/AIMS: In kidney, extracellular [Ca2+] can modulate intracellular [Ca2+] to control key cellular processes. Hence, extracellular [Ca2+] is normally maintained within narrow range. We tested effect of extracellular ATP on viability of human proximal (HK-2) cells at high calcium. Modulation of intracellular calcium was assessed by imaging cytosolic [Ca2+], and expression of calcium-binding proteins (CaBPs). We present an artificial intelligence enabled deep learning model for prediction of injury and protection against extracellular [Ca2+] in HK-2 cells. METHODS: HK-2 cells were cultured in calcium-free DMEM supplemented with CaCl2. Morphological changes were detected using light microscopy. Cell viability was determined using MTT Assay. Intracellular [Ca2+] was detected using fluorescence microscopy. For easy detection of HK-2 cells injury, we performed light microscopy image classification based on Convolutional Neural Network. Expression of CaBPs, p21, and Mcl-1 was measured using real-time PCR. RESULTS: We show decreased viability of HK-2 cells cultured in elevated calcium levels, which was prevented by adenosine triphosphate (ATP). Exposure of cells to elevated extracellular [Ca2+] correlated with increasing fluorescence of intracellular calcium indicator, which was attenuated in presence of ATP. Since features cannot be detected easily by human eyes, we propose a customized deep learning-based CNN model for classification of HK-2 cells injury by extracellular calcium with high accuracy of 98%. Our data demonstrated significant increase in mRNA levels of calmodulin, S100A8, S100A14 and CaBP28k, with elevated extracellular [Ca2+]. Expression of these genes was enhanced with ATP. CONCLUSION: The results suggest that ATP protects human proximal (HK-2) cells against elevated extracellular calcium levels. We present a CNN model as user friendly tool to study calcium dependent injury in (HK-2) cells. Finally, we show that ATP-mediated protection is correlated with enhanced expression of calcium-binding proteins.

A Low Power Sigma-Delta Modulator with Hybrid Architecture.

Analogue-to-digital converters (ADC) using oversampling technology and the Σ-∆ modulation mechanism are widely applied in digital audio systems. This paper presents an audio modulator with high accuracy and low power consumption by using a discrete second-order feedforward structure. A 5-bit successive approximation register (SAR) quantizer is integrated into the chip, which reduces the number of comparators and the power consumption of the quantizer compared with flash ADC-type quantizers. An analogue passive adder is used to sum the input signals and it is embedded in a SAR ADC composed of a capacitor array and a dynamic comparator which has no static power consumption. To validate the design concept, the designed modulator is developed in a 180 nm CMOS process. The peak signal to noise distortion ratio (SNDR) is calculated as 106 dB and the total power consumption of the chip is recorded as 3.654 mW at the chip supply voltage of 1.8 V. The input sine wave of 0 to 25 kHz is sampled at a sampling frequency of 3.2 Ms/s. Moreover, the results achieve a 16-bit effective number of bits (ENOB) when the amplitude of the input signal is varied between 0.15 and 1.65 V. By comparing with other modulators which were realized by a 180 nm CMOS process, the proposed architecture outperforms with lower power consumption.