Neuromorphic computing for content-based image retrieval.

Neuromorphic computing mimics the neural activity of the brain through emulating spiking neural networks. In numerous machine learning tasks, neuromorphic chips are expected to provide superior solutions in terms of cost and power efficiency. Here, we explore the application of Loihi, a neuromorphic computing chip developed by Intel, for the computer vision task of image retrieval. We evaluated the functionalities and the performance metrics that are critical in content-based visual search and recommender systems using deep-learning embeddings. Our results show that the neuromorphic solution is about 2.5 times more energy-efficient compared with an ARM Cortex-A72 CPU and 12.5 times more energy-efficient compared with NVIDIA T4 GPU for inference by a lightweight convolutional neural network when batch size is 1 while maintaining the same level of matching accuracy. The study validates the potential of neuromorphic computing in low-power image retrieval, as a complementary paradigm to the existing von Neumann architectures.

Analysis of the gel properties, microstructural characteristics, and intermolecular forces of soybean protein isolate gel induced by transglutaminase.

Soybean protein isolate (SPI) is a high-quality plant protein that is primarily used to process various soybean products coagulated by transglutaminase (TGase). In this study, the degree of hydrolysis (DH), sulfhydryl content (SH), surface hydrophobicity (H0 ), secondary structural constitution, and microstructure of TGase-treated soybean protein (SPI, 7S, and 11S) were determined, as well as the effects of NaCl, urea, and SDS on the properties and intermolecular forces of SPI gel were analyzed. The results show that the H0 and SH content of SPI, 7S, and 11S decreased significantly with TGase treatment time (p < .05), while the DH gradually increased and reached its highest value (3.72%, 7.41%, and 1.27%, respectively) at 30 min. As the concentration of these two secondary structures exhibited an inverse relationship, the degradation of ß-turns resulted in the increase in ß-sheets. The microstructures of SPI and 11S gels were similar, being denser and more ordered than 7S gel. The low concentration of NaCl solution (0.2 mol/L) enhanced gel properties and intermolecular forces, promoting the formation of SPI gel, whereas a high concentration (0.4-0.8 mol/L) had a significant inhibitory effect. Urea and SDS solutions substantially inhibited the formation of SPI gel, leading to significant decreases in the water holding capacity and hardness as well as a considerable increase in the coagulation time (p < .05). The results revealed that hydrogen bonds and hydrophobic interactions were the main intermolecular forces responsible for the gel formation. This study provides adequate technical support and a theoretical basis for soybean protein gel products.