Actor-Critic With Synthesis Loss for Solving Approximation Biases.

Approximation biases of value functions are considered a key problem in reinforcement learning (RL). In particular, existing RL algorithms are hindered by overestimation and underestimation biases, i.e., value mismatching between RL's actual returns and action-value approximations limits the performance of RL algorithms. In this article, we first develop a new synthesis loss function for RL's action-value estimation integrating a regularization term and a modified "clipped double Q -learning" structure for solving overestimation and underestimation biases. To minimize the differences between action-value estimations and actual returns in RL, we develop a new discrepancy function to determine the type and magnitude of approximation biases. Then, two coefficients embedded in the synthesis loss are automatically tuned by minimizing the discrepancy function during training to minimize approximation biases. We further design a new actor-critic (AC) algorithm, named AC with synthesis loss (ACSL), by integrating the synthesis loss function and an error-controlled mechanism. Experimental results on continuous control tasks illustrate that the proposed ACSL algorithm outperforms other cutting-edge RL methods in many tasks and that the proposed synthesis loss function is easily implemented into other algorithms and significantly reduces approximation biases while improving performance. The proposed method can successfully handle many complex continuous control tasks and can greatly outperform other state-of-the-art algorithms on most tasks.

13-Oxyingenol-dodecanoate inhibits the growth of non-small cell lung cancer cells by targeting ULK1.

Lung cancer is the leading cause of cancer deaths worldwide. Non-small cell lung cancer (NSCLC) accounts for 80-85% of all lung cancers. Euphorbia kansui yielded 13-oxyingenol-dodecanoate (13OD), an ingenane-type diterpenoid, which had a strong cytotoxic effect on NSCLC cells. The underlying mechanism and potential target, however, remained unknown. The study found that 13OD effectively inhibited the cell proliferation and colony formation of NSCLC cells (A549 and H460 cells), with less toxicity in normal human lung epithelial BEAS-2B cells. Moreover, 13OD can cause mitochondrial dysfunction, and apoptosis in NSCLC cells. Mechanistically, the transcriptomics results showed that differential genes were mainly enriched in the mTOR and AMPK signaling pathways, which are closely related to cellular autophagy, the related indicators were subsequently validated. Additionally, bafilomycin A1 (Baf A1), an autophagy inhibitor, reversed the mitochondrial damage caused by 13OD. Furthermore, the Omics and Text-based Target Enrichment and Ranking (OTTER) method predicted ULK1 as a potential target of 13OD against NSCLC cells. This hypothesis was further confirmed using molecular docking, the cellular thermal shift assay (CETSA), and Western blot analysis. Remarkably, ULK1 siRNA inhibited 13OD's toxic activity in NSCLC cells. In line with these findings, 13OD was potent and non-toxic in the tumor xenograft model. Our findings suggested a possible mechanism for 13OD's role as a tumor suppressor and laid the groundwork for identifying targets for ingenane-type diterpenoids.

Ingenane-type diterpenoids inhibit non-small cell lung cancer cells by regulating SRC/PI3K/Akt pathway.

Ingenane-type diterpenoids (ITDs) are distinct components of plants belonging to the genus Euphorbia. These compounds have significant cytotoxic effects on non-small cell lung cancer (NSCLC) cells. However, the underlying molecular mechanism has yet to be reported. To explore the mechanism of the anticancer effect of ITDs, we carried out a network pharmacology prediction study. PPI network suggested that SRC and PI3K had high levels of interaction. In addition, KEGG analysis revealed that these common targets were significantly enriched in the PI3K/Akt signalling pathway. 13-oxyingenol-dodecanoate (13OD) was used for validation after the biological evaluation of some ITDs against NSCLC cells. It demonstrated that 13OD could significantly inhibit the growth of NSCLC cells by inducing apoptosis. The results from molecular docking and Western blotting showed that 13OD interacted with SRC and PI3K and down-regulated the SRC/PI3K/Akt signalling pathway in NSCLC cells. This study provided the underlying mechanism of ITDs against NSCLC.

Using multiple indicators to predict the risk of surgical site infection after ORIF of tibia fractures: a machine learning based study.

Objective: Surgical site infection (SSI) are a serious complication that can occur after open reduction and internal fixation (ORIF) of tibial fractures, leading to severe consequences. This study aimed to develop a machine learning (ML)-based predictive model to screen high-risk patients of SSI following ORIF of tibial fractures, thereby aiding in personalized prevention and treatment. Methods: Patients who underwent ORIF of tibial fractures between January 2018 and October 2022 at the Department of Emergency Trauma Surgery at Ganzhou People's Hospital were retrospectively included. The demographic characteristics, surgery-related variables and laboratory indicators of patients were collected in the inpatient electronic medical records. Ten different machine learning algorithms were employed to develop the prediction model, and the performance of the models was evaluated to select the best predictive model. Ten-fold cross validation for the training set and ROC curves for the test set were used to evaluate model performance. The decision curve and calibration curve analysis were used to verify the clinical value of the model, and the relative importance of features in the model was analyzed. Results: A total of 351 patients who underwent ORIF of tibia fractures were included in this study, among whom 51 (14.53%) had SSI and 300 (85.47%) did not. Of the patients with SSI, 15 cases were of deep infection, and 36 cases were of superficial infection. Given the initial parameters, the ET, LR and RF are the top three algorithms with excellent performance. Ten-fold cross-validation on the training set and ROC curves on the test set revealed that the ET model had the best performance, with AUC values of 0.853 and 0.866, respectively. The decision curve analysis and calibration curves also showed that the ET model had the best clinical utility. Finally, the performance of the ET model was further tested, and the relative importance of features in the model was analyzed. Conclusion: In this study, we constructed a multivariate prediction model for SSI after ORIF of tibial fracture through ML, and the strength of this study was the use of multiple indicators to establish an infection prediction model, which can better reflect the real situation of patients, and the model show great clinical prediction performance.