Forward-forward training of an optical neural network.

Neural networks (NNs) have demonstrated remarkable capabilities in various tasks, but their computation-intensive nature demands faster and more energy-efficient hardware implementations. Optics-based platforms, using technologies such as silicon photonics and spatial light modulators, offer promising avenues for achieving this goal. However, training multiple programmable layers together with these physical systems poses challenges, as they are difficult to fully characterize and describe with differentiable functions, hindering the use of error backpropagation algorithm. The recently introduced forward-forward algorithm (FFA) eliminates the need for perfect characterization of the physical learning system and shows promise for efficient training with large numbers of programmable parameters. The FFA does not require backpropagating an error signal to update the weights, rather the weights are updated by only sending information in one direction. The local loss function for each set of trainable weights enables low-power analog hardware implementations without resorting to metaheuristic algorithms or reinforcement learning. In this paper, we present an experiment utilizing multimode nonlinear wave propagation in an optical fiber demonstrating the feasibility of the FFA approach using an optical system. The results show that incorporating optical transforms in multilayer NN architectures trained with the FFA can lead to performance improvements, even with a relatively small number of trainable weights. The proposed method offers a new path to the challenge of training optical NNs and provides insights into leveraging physical transformations for enhancing the NN performance.

Blood flow restriction training promotes functional recovery of knee joint in patients after arthroscopic partial meniscectomy: A randomized clinical trial.

Purpose: To explore the effect of blood flow restriction training (BFRT) on the recovery of knee function in patients after arthroscopic partial meniscectomy (APM). Methods: Forty patients undergoing APM surgery were included in this parallel group, two-arm, single-assessor blinded, randomized clinical trial. The subjects were randomly divided into two groups: routine rehabilitation group (RR Group, n = 20) and routine rehabilitation + blood flow restriction training group (RR + BFRT Group, n = 20). One subject in each group dropped out during the experiment. All patients received 8 weeks of routine rehabilitation starting from the second day after APM. In addition, patients in the RR + BFRT group required additional BFRT twice a week. Visual analogue scale (VAS) score, range of motion (ROM), one-leg standing test (OLST) score, Lysholm knee score, quadriceps muscle strength, quadriceps thickness, and thigh circumference were evaluated at preoperative, postoperative, 4 and 8 weeks after surgery. SPSS 25.0 software was used for statistical analysis of the data. Repeated measures ANOVA was used if the data were normally distributed and had homogeneity of variance. Generalized estimating equations were chosen if the data were not normally distributed or had homogeneity of variance. Results: There were no significant differences in VAS score, ROM, OLST score, Lysholm knee score, quadriceps muscle strength, quadriceps thickness, and thigh circumference between the two groups before surgery (p > 0.05). Compared with postoperative, VAS score, ROM, OLST score, Lysholm knee score, and thigh circumference were significantly improved in the RR group (p < 0.05), while quadriceps muscle strength and quadriceps thickness were not significantly enhanced at 8 weeks postoperatively (p > 0.05). However, VAS score, ROM, OLST score, Lysholm knee score, quadriceps muscle strength, quadriceps thickness, and thigh circumference were all significantly improved in the RR + BFRT group at 8 weeks postoperatively (p < 0.05). Furthermore, compared with the RR group, VAS score (50% vs. 86%), ROM (7.9% vs. 16.0%), OLST score (57.3% vs. 130.1%), Lysholm knee score (38.4% vs. 55.7%), relative peak torque (11.0% vs. 84.7%), mean power (20.6% vs. 88.1%), rectus femoris thickness (0.40% vs. 13.0%), vastus medialis (0.29% vs. 5.32%), vastus lateralis (0% vs. 6.2%), vastus internus (0% vs. 5.8%), and thigh circumference (2.7% vs. 5.8%) in the RR + BFRT group were significantly improved at 4 and 8 weeks postoperatively (p < 0.05). Conclusion: BFRT combined with routine rehabilitation training can better promote the recovery of knee joint function in patients after APM, especially the improvement of quadriceps muscle strength and thickness.

Synthesis of Benzylidene Analogs of Oleanolic Acid as Potential α-Glucosidase and α-Amylase Inhibitors.

A series of benzylidene analogs of oleanolic acid 4a∼4s were synthesized and assessed for their α-glucosidase and α-amylase inhibitory activities. The results presented that all synthesized analogs exhibited excellent-to-moderate inhibitory effects on α-glucosidase and α-amylase. Analog 4i showed the highest α-glucosidase inhibition (IC50: 0.40 µM), and analog 4o presented the strongest α-amylase inhibition (IC50: 9.59 µM). Inhibition kinetics results showed that analogs 4i and 4o were reversible and mixed-type inhibitors against α-glucosidase and α-amylase, respectively. Simulation docking results demonstrated the interaction between analogs and two enzymes. Moreover, analogs 4i and 4o showed a high level of safety against 3T3-L1 and HepG2 cells.

Synthesis and bioactivities evaluation of oleanolic acid oxime ester derivatives as α-glucosidase and α-amylase inhibitors.

Different oleanolic acid (OA) oxime ester derivatives (3a-3t) were designed and synthesised to develop inhibitors against α-glucosidase and α-amylase. All the synthesised OA derivatives were evaluated against α-glucosidase and α-amylase in vitro. Among them, compound 3a showed the highest α-glucosidase inhibition with an IC50 of 0.35 µM, which was ∼1900 times stronger than that of acarbose, meanwhile compound 3f exhibited the highest α-amylase inhibitory with an IC50 of 3.80 µM that was ∼26 times higher than that of acarbose. The inhibition kinetic studies showed that the inhibitory mechanism of compounds 3a and 3f were reversible and mixed types towards α-glucosidase and α-amylase, respectively. Molecular docking studies analysed the interaction between compound and two enzymes, respectively. Furthermore, cytotoxicity evaluation assay demonstrated a high level of safety profile of compounds 3a and 3f against 3T3-L1 and HepG2 cells.HighlightsOleanolic acid oxime ester derivatives (3a-3t) were synthesised and screened against α-glucosidase and α-amylase.Compound 3a showed the highest α-glucosidase inhibitory with IC50 of 0.35 µM.Compound 3f presented the highest α-amylase inhibitory with IC50 of 3.80 µM.Kinetic studies and in silico studies analysed the binding between compounds and α-glucosidase or α-amylase.