Advances in neural architecture search.

Automated machine learning (AutoML) has achieved remarkable success in automating the non-trivial process of designing machine learning models. Among the focal areas of AutoML, neural architecture search (NAS) stands out, aiming to systematically explore the complex architecture space to discover the optimal neural architecture configurations without intensive manual interventions. NAS has demonstrated its capability of dramatic performance improvement across a large number of real-world tasks. The core components in NAS methodologies normally include (i) defining the appropriate search space, (ii) designing the right search strategy and (iii) developing the effective evaluation mechanism. Although early NAS endeavors are characterized via groundbreaking architecture designs, the imposed exorbitant computational demands prompt a shift towards more efficient paradigms such as weight sharing and evaluation estimation, etc. Concurrently, the introduction of specialized benchmarks has paved the way for standardized comparisons of NAS techniques. Notably, the adaptability of NAS is evidenced by its capability of extending to diverse datasets, including graphs, tabular data and videos, etc., each of which requires a tailored configuration. This paper delves into the multifaceted aspects of NAS, elaborating on its recent advances, applications, tools, benchmarks and prospective research directions.

Artificial intelligence algorithms permits rapid acute kidney injury risk classification of patients with acute myocardial infarction.

Objective: This study aimed to develop and validate several artificial intelligence (AI) models to identify acute myocardial infarction (AMI) patients at an increased risk of acute kidney injury (AKI) during hospitalization. Methods: Included were patients diagnosed with AMI from the Medical Information Mart for Intensive Care (MIMIC) III and IV databases. Two cohorts of AMI patients from Changzhou Second People's Hospital and Xuzhou Center Hospital were used for external validation of the models. Patients' demographics, vital signs, clinical characteristics, laboratory results, and therapeutic measures were extracted. Totally, 12 AI models were developed. The area under the receiver operating characteristic curve (AUC) were calculated and compared. Results: AKI occurred during hospitalization in 1098 (28.3 %) of the 3882 final enrolled patients, split into training (3105) and test (777) sets randomly. Among them, Random Forest (RF), C5.0 and Bagged CART models outperformed the other models in both the training and test sets. The AUCs for the test set were 0.754, 0.734 and 0.730, respectively. The incidence of AKI was 9.8 % and 9.5 % in 2202 patients in the Changzhou cohort and 807 patients in the Xuzhou cohort with AMI, respectively. The AUCs for patients in the Changzhou cohort were RF, 0.761; C5.0, 0.733; and bagged CART, 0.725, respectively, and Xuzhou cohort were RF, 0.799; C5.0, 0.808; and bagged CART, 0.784, respectively. Conclusion: Several machines learning-based prediction models for AKI after AMI were developed and validated. The RF, C5.0 and Bagged CART model performed robustly in identifying high-risk patients earlier. Clinical trial approval statement: This Trial was registered in the Chinese clinical trials registry: ChiCTR1800014583. Registered January 22, 2018 (http://www.chictr.org.cn/searchproj.aspx).

Inhibitory effect and mechanism of violacein on planktonic growth, spore germination, biofilm formation and toxin production of Bacillus cereus and its application in grass carp preservation.

Bacillus cereus is a ubiquitous foodborne pathogen commonly found in various foods. Its ability to form spores, biofilms and diarrhoeal and/or emetic toxins further exacerbates the risk of food poisoning. Violacein is a tryptophan derivative with excellent antibacterial activity. However, the knowledge on the antibacterial action of violacein against B. cereus was lacking, and thus this study aimed to investigate the antibacterial activity and mechanism. The antibacterial results demonstrated that minimum inhibitory concentration and minimum bactericidal concentration of violacein were 3.125 mg/L and 12.50 mg/L, respectively. Violacein could effectively inhibit planktonic growth, spore germination and biofilm formation of B. cereus (P < 0.001). Meanwhile, violacein significantly downregulated the expression of toxin genes, including nheA (P < 0.05), nheB (P < 0.001), bceT (P < 0.01), cytK (P < 0.001), hblC (P < 0.001) and hblD (P < 0.001). Results of extracellular alkaline phosphatase, nucleotide and protein leakage assays and scanning and transmission electron microscopy observation tests showed violacein destroyed cell walls and membranes of B. cereus. In addition, 6.25 mg/kg of violacein could significantly inhibit B. cereus in grass carp fillets (P < 0.05). These results demonstrate that violacein has great potential as an effective natural antimicrobial preservative to control food contamination and poisoning events caused by B. cereus.

Disentangled Representation Learning.

Disentangled Representation Learning (DRL) aims to learn a model capable of identifying and disentangling the underlying factors hidden in the observable data in representation form. The process of separating underlying factors of variation into variables with semantic meaning benefits in learning explainable representations of data, which imitates the meaningful understanding process of humans when observing an object or relation. As a general learning strategy, DRL has demonstrated its power in improving the model explainability, controlability, robustness, as well as generalization capacity in a wide range of scenarios such as computer vision, natural language processing, and data mining. In this article, we comprehensively investigate DRL from various aspects including motivations, definitions, methodologies, evaluations, applications, and model designs. We first present two well-recognized definitions, i.e., Intuitive Definition and Group Theory Definition for disentangled representation learning. We further categorize the methodologies for DRL into four groups from the following perspectives, the model type, representation structure, supervision signal, and independence assumption. We also analyze principles to design different DRL models that may benefit different tasks in practical applications. Finally, we point out challenges in DRL as well as potential research directions deserving future investigations. We believe this work may provide insights for promoting the DRL research in the community.

Dynamic Spatio-Temporal Graph Reasoning for VideoQA With Self-Supervised Event Recognition.

Video question answering (VideoQA) requires the ability of comprehensively understanding visual contents in videos. Existing VideoQA models mainly focus on scenarios involving a single event with simple object interactions and leave event-centric scenarios involving multiple events with dynamically complex object interactions largely unexplored. These conventional VideoQA models are usually based on features extracted from the global visual signals, making it difficult to capture the object-level and event-level semantics. Although there exists a recent work utilizing a static spatio-temporal graph to explicitly model object interactions in videos, it ignores the dynamic impact of questions for graph construction and fails to exploit the implicit event-level semantic clues in questions. To overcome these limitations, we propose a Self-supervised Dynamic Graph Reasoning (SDGraphR) model for video question answering (VideoQA). Our SDGraphR model learns a question-guided spatio-temporal graph that dynamically encodes intra-frame spatial correlations and inter-frame correspondences between objects in the videos. Furthermore, the proposed SDGraphR model discovers event-level cues from questions to conduct self-supervised learning with an auxiliary event recognition task, which in turn helps to improve its VideoQA performances without using any extra annotations. We carry out extensive experiments to validate the substantial improvements of our proposed SDGraphR model over existing baselines.

Performance optimization and comparison of vertical motion-based triboelectric nanogenerators.

The vertical motion configuration is a common design in triboelectric nanogenerators (TENGs) for energy harvesting; however, the performance optimization and comparison are still vague between various vertical motion-based structures. In this paper, time-averaged power density is defined as a metric to compare the power output performances of vertically structured TENGs, including contact mode and freestanding mode. To ensure comparisons under the same circumstances, a novel sandwich-structured dielectric layer was designed to maintain a stable and consistent surface charge density, with an extra rotating triboelectric nanogenerator working as a charge pump. We also investigated the impact of parasitic capacitance, which is a primary source of error in theoretical optimization. The freestanding TENG (FTENG) with a single dielectric layer demonstrates superior power performance, even when accounting for the influence of parasitic capacitance. This work provides valuable insights and guidelines for the design of high-performance mechanical energy harvesting devices.

From mechanism to application: Decrypting light-regulated denitrifying microbiome through geometric deep learning.

Regulation on denitrifying microbiomes is crucial for sustainable industrial biotechnology and ecological nitrogen cycling. The holistic genetic profiles of microbiomes can be provided by meta-omics. However, precise decryption and further applications of highly complex microbiomes and corresponding meta-omics data sets remain great challenges. Here, we combined optogenetics and geometric deep learning to form a discover-model-learn-advance (DMLA) cycle for denitrification microbiome encryption and regulation. Graph neural networks (GNNs) exhibited superior performance in integrating biological knowledge and identifying coexpression gene panels, which could be utilized to predict unknown phenotypes, elucidate molecular biology mechanisms, and advance biotechnologies. Through the DMLA cycle, we discovered the wavelength-divergent secretion system and nitrate-superoxide coregulation, realizing increasing extracellular protein production by 83.8% and facilitating nitrate removal with 99.9% enhancement. Our study showcased the potential of GNNs-empowered optogenetic approaches for regulating denitrification and accelerating the mechanistic discovery of microbiomes for in-depth research and versatile applications.

miR-873-5p Suppression Reinvigorates Aging Mesenchymal Stem Cells and Improves Cardiac Repair after Myocardial Infarction.

Aging poses obstacles to the functionality of human mesenchymal stem cells (MSCs), resulting in a notable decline in their valuable contribution to myocardial infarction (MI). MicroRNAs (miRNAs) play a pivotal role in governing MSC aging; nonetheless, the specific mechanisms remain puzzling. This research delved into the value of miR-873-5p in the management of MSC aging and investigated whether the restraint of miR-873-5p could regenerate aged MSCs (AMSCs), thereby enhancing their healing success for MI. In this study, MSCs were isolated from both young donors (referred to as YMSCs) and aged donors (referred to as AMSCs). The senescence status of these MSCs was evaluated through the application of age-related ß-galactosidase (SA-ß-gal) staining. Following this assessment, the MSCs, including those treated with anti-miR-873-5p-AMSCs, were then transplanted into the hearts of Sprague-Dawley rats experiencing acute myocardial infarction. Increasing miR-873-5p levels in YMSCs resulted in elevated cellular aging, whereas reducing miR-873-5p expression decreased aging in AMSCs. Mechanistically, miR-873-5p inhibited autophagy in MSCs through the AMPK signaling pathway, leading to cellular aging by suppressing the Cab39 expression. Partial alleviation of these effects was achieved by the administration of the autophagy inhibitor 3-methyladenine. Grafting of anti-miR-873-5p-AMSCs, by enhancing angiogenesis and bolstering cell survival, led to an improvement in cardiac function in the rat model, unlike the transplantation of AMSCs. miR-873-5p which serves as a pivotal element in mediating MSC aging through its regulation of the Cab39/AMPK signaling pathway. It represents an innovative target for revitalizing AMSCs and enhancing their heart-protective abilities.

Erratum: Circulating Exosomal miR-1-3p from Rats with Myocardial Infarction Plays a Protective Effect on Contrast-Induced Nephropathy via Targeting ATG13 and activating the AKT Signaling Pathway: Erratum.

[This corrects the article DOI: 10.7150/ijbs.55887.].

Exosomes derived from mir-214-3p overexpressing mesenchymal stem cells promote myocardial repair.

AIMS: Exosomes are known as nanovesicles that are naturally secreted, playing an essential role in stem-mediated cardioprotection. This study mainly focused on investigating if exosomes derived from miR-214 overexpressing mesenchymal stem cells (MSCs) show more valid cardioprotective ability in a rat model of acute myocardial infarction (AMI) and its potential mechanisms. METHODS: Exosomes were isolated from control MSCs (Ctrl-Exo) and miR-214 overexpressing MSCs (miR-214OE-Exo) and then they were delivered to cardiomyocytes and endothelial cells in vitro under hypoxia and serum deprivation (H/SD) condition or in vivo in an acutely infarcted Sprague-Dawley rat heart. Regulated genes and signal pathways by miR-214OE-Exo treatment were explored using western blot analysis and luciferase assay. RESULTS IN VITRO: , miR-214OE-Exo enhanced migration, tube-like formation in endothelial cells. In addition, miR-214OE-Exo ameliorated the survival of cardiomyocytes under H/SD. In the rat AMI model, compared to Ctrl-Exo, miR-214OE-Exo reduced myocardial apoptosis, and therefore reduced infarct size and improved cardiac function. Besides, miR-214OE-Exo accelerated angiogenesis in peri-infarct region. Mechanistically, we identified that exosomal miR-214-3p promoted cardiac repair via targeting PTEN and activating p-AKT signal pathway. CONCLUSION: Exosomes derived from miR-214 overexpressing MSCs have greatly strengthened the therapeutic efficacy for treatment of AMI by promoting cardiomyocyte survival and endothelial cell function.

Transmission Line-Planning Method Based on Adaptive Resolution Grid and Improved Dijkstra Algorithm.

An improved Dijkstra algorithm based on adaptive resolution grid (ARG) is proposed to assist manual transmission line planning, shorten the construction period and achieve lower cost and higher efficiency of line selection. Firstly, the semantic segmentation network is used to change the remote sensing image into a ground object-identification image and the grayscale image of the ground object-identification image is rasterized. The ARG map model is introduced to greatly reduce the number of redundant grids, which can effectively reduce the time required to traverse the grids. Then, the Dijkstra algorithm is combined with the ARG and the neighborhood structure of the grid is a multi-center neighborhood. An improved method of bidirectional search mechanism based on ARG and inflection point-correction is adopted to greatly increase the running speed. The inflection point-correction reduces the number of inflection points and reduces the cost. Finally, according to the results of the search, the lowest-cost transmission line is determined. The experimental results show that this method aids manual planning by providing a route for reference, improving planning efficiency while shortening the duration, and reducing the time spent on algorithm debugging. Compared with the comparison algorithm, this method is faster in running speed and better in cost saving and has a broader application prospect.

A New Scoring System for Predicting Ventricular Arrhythmia Risk in Patients with Acute Myocardial Infarction.

Objective: In this study, a risk score for ventricular arrhythmias (VA) were evaluated for predicting the risk of ventricular arrhythmia (VA) of acute myocardial infarction (AMI) patients. Methods: Patients with AMI were divided into two sets according to whether VA occurred during hospitalization. Another cohort was enrolled for external validation. The area under the curve (AUC) of receiver operating characteristic (ROC) was calculated to evaluate the accuracy of the model. Results: A total of 1493 eligible patients with AMI were enrolled as the training set, of whom 70 (4.7%) developed VA during hospitalization. In-hospital mortality was significantly higher in the VA set than in the non-VA set (31.4% vs 2.7%, P=0.001). The independent predictors of VA in patients with AMI including Killip grade ≥3, STEMI patients, LVEF <50%, frequent premature ventricular beats, serum potassium <3.5 mmol/L, type 2 diabetes, and creatinine level. The AUC of the model for predicting VT/VF in the training set was 0.815 (95% CI: 0.763-0.866). A total of 1149 cases were enrolled from Xuzhou Center Hospital as the external validation set. The AUC of the model in the external validation set for predicting VT/VF was 0.755 (95% CI: 0.687-0.823). Calibration curves indicated a good consistency between the predicted and the observed probabilities of VA in both sets. Conclusion: We have established a clinical prediction risk score for predicting the occurrence of VA in AMI patients. The prediction score is easy to use, performs well and can be used to guide clinical practice.

Disentangled Representation Learning for Recommendation.

There exist complex interactions among a large number of latent factors behind the decision making processes of different individuals, which drive the various user behavior patterns in recommender systems. These factors hidden in those diverse behaviors demonstrate highly entangled patterns, covering from high-level user intentions to low-level individual preferences. Uncovering the disentanglement of these latent factors can benefit in enhanced robustness, interpretability, and controllability during representation learning for recommendation. However, the large degree of entanglement within latent factors poses great challenges for learning representations that disentangle them, and remains largely unexplored in literature. In this paper, we present the SEMantic MACRo-mIcro Disentangled Variational Auto-Encoder (SEM-MacridVAE) model for learning disentangled representations from user behaviors, taking item semantic information into account. Our SEM-MacridVAE model achieves macro disentanglement by inferring the high-level concepts associated with user intentions (e.g., to buy a pair of shoes or a laptop) through a prototype routing mechanism, as well as capturing the individual preferences with respect to different concepts separately. The micro disentanglement is guaranteed through a micro-disentanglement regularizer stemming from an information-theoretic interpretation of VAEs, which forces each dimension of the representations to independently reflect an isolated low-level factor (e.g., the size or the color of a shirt). The semantic information including visual and categorical signals extracted from candidate items is utilized to further boost the recommendation performance of the proposed SEM-MacridVAE model. Empirical experiments demonstrate that our proposed approach is able to achieve significant improvement over the state-of-the-art baselines. We also show that the learned representations are interpretable and controllable, capable of potentially leading to a new paradigm for recommendation where users have fine-grained control over some target aspects of the recommendation candidates.

DeepLogic: Joint Learning of Neural Perception and Logical Reasoning.

Neural-symbolic learning, aiming to combine the perceiving power of neural perception and the reasoning power of symbolic logic together, has drawn increasing research attention. However, existing works simply cascade the two components together and optimize them isolatedly, failing to utilize the mutual enhancing information between them. To address this problem, we propose DeepLogic, a framework with joint learning of neural perception and logical reasoning, such that these two components are jointly optimized through mutual supervision signals. In particular, the proposed DeepLogic framework contains a deep-logic module that is capable of representing complex first-order-logic formulas in a tree structure with basic logic operators. We then theoretically quantify the mutual supervision signals and propose the deep&logic optimization algorithm for joint optimization. We further prove the convergence of DeepLogic and conduct extensive experiments on model performance, convergence, and generalization, as well as its extension to the continuous domain. The experimental results show that through jointly learning both perceptual ability and logic formulas in a weakly supervised manner, our proposed DeepLogic framework can significantly outperform DNN-based baselines by a great margin and beat other strong baselines without out-of-box tools.

Association of plasma free triiodothyronine levels with contrast-induced acute kidney injury and short-term survival in patients with acute myocardial infarction.

Objective: Post-treatment contrast-induced acute kidney injury (CI-AKI) is associated with poor outcomes in patients with acute myocardial infarction (AMI). A lower free triiodothyronine (FT3) level predicts a poor prognosis of AMI patients. This study evaluated the effect of plasma FT3 level in predicting CI-AKI and short-term survival among AMI patients. Methods: Coronary arteriography or percutaneous coronary intervention was performed in patients with AMI. A 1:3 propensity score (PS) was used to match patients in the CI-AKI group and the non-CI-AKI group. Results: Of 1480 patients enrolled in the study, 224 (15.1%) patients developed CI-AKI. The FT3 level was lower in CI-AKI patients than in non-CI-AKI patients (3.72 ± 0.88 pmol/L vs 4.01 ± 0.80 pmol/L, P < 0.001). Compared with those at the lowest quartile of FT3, the patients at quartiles 2-4 had a higher risk of CI-AKI respectively (P for trend = 0.005). The risk of CI-AKI increased by 17.7% as FT3 level decreased by one unit after PS-matching analysis (odds ratio: 0.823; 95% CI: 0.685-0.988, P = 0.036). After a median of 31 days of follow-up (interquartile range: 30-35 days), 78 patients died, including 72 cardiogenic deaths and 6 non-cardiogenic deaths, with more deaths in the CI-AKI group than in the non-CI-AKI group (53 vs 25, P < 0.001). Kaplan-Meier survival analysis showed that patients at a lower FT3 quartile achieved a worse survival before and after matching. Conclusion: Lower FT3 may increase the risk of CI-AKI and 1-month mortality in AMI patients.

[Discussion on Technical Evaluation Requirements of Allergen Detection Reagent Pre-marketing].

In the perspective of technical evaluation, the pre-marketing regulatory requirements of allergen detection reagents in China, America, European Union were compared, and the regulatory risks and performance requirements of this product were analyzed based on the monitoring of post-marketing adverse events, reference standards and domestic and foreign regulatory documents. In view of the "neck-stuck" problems such as the difficulty of clinical trials, the difficulty of finding comparable contrast reagents and the lack of clinical diagnostic gold standards, this paper discusses and gives regulatory suggestions, with a view to providing technical reference for product R&D, production, evaluation, approval and supervision in this field.

[Core Architecture and Clinical Application of Orthopedic Robot Based on Human-like Thinking].

Intelligent and precision medical treatment is the future development trend of surgical operations. We proposed a core architecture of orthopedic robots with human-like thinking based on the growing demand for orthopedic robots and disadvantages of current robots, it consists of brain, eyes and hands three modules according to functions. The architecture design is extremely in line with the doctor's logic so that the work process of the orthopedic robot is similar to the process of traditional surgery which is mainly done by the doctor's brain-eye-hand coordination. It realizes the digitization of the doctor's thinking, the immediacy and visualization of surgical information and the accuracy of surgical operation process. The clinical application proved that the orthopedic robot has the advantages of higher accuracy, less radiation and shorter operation time, which can be further promoted clinically.

Semantic Conditioned Dynamic Modulation for Temporal Sentence Grounding in Videos.

Temporal sentence grounding in videos aims to localize one target video segment, which semantically corresponds to a given sentence. Unlike previous methods mainly focusing on matching semantics between the sentence and different video segments, in this paper, we propose a novel semantic conditioned dynamic modulation (SCDM) mechanism, which leverages the sentence semantics to modulate the temporal convolution operations for better correlating and composing the sentence-relevant video contents over time. The proposed SCDM also performs dynamically with respect to the diverse video contents so as to establish a precise semantic alignment between sentence and video. By coupling the proposed SCDM with a hierarchical temporal convolutional architecture, video segments with various temporal scales are composed and localized. Besides, more fine-grained clip-level actionness scores are also predicted with the SCDM-coupled temporal convolution on the bottom layer of the overall architecture, which are further used to adjust the temporal boundaries of the localized segments and thereby lead to more accurate grounding results. Experimental results on benchmark datasets demonstrate that the proposed model can improve the temporal grounding accuracy consistently, and further investigation experiments also illustrate the advantages of SCDM on stabilizing the model training and associating relevant video contents for temporal sentence grounding. Our code for this paper is available at https://github.com/yytzsy/SCDM-TPAMI.

Syntax Customized Video Captioning by Imitating Exemplar Sentences.

Enhancing the diversity of sentences to describe video contents is an important problem arising in recent video captioning research. In this paper, we explore this problem from a novel perspective of customizing video captions by imitating exemplar sentence syntaxes. Specifically, given a video and any syntax-valid exemplar sentence, we introduce a new task of Syntax Customized Video Captioning (SCVC) aiming to generate one caption which not only semantically describes the video contents but also syntactically imitates the given exemplar sentence. To tackle the SCVC task, we propose a novel video captioning model, where a hierarchical sentence syntax encoder is first designed to extract the syntactic structure of the exemplar sentence, then a syntax conditioned caption decoder is devised to generate the syntactically structured caption expressing video semantics. As there is no available syntax customized groundtruth video captions, we tackle such a challenge by proposing a new training strategy, which leverages the traditional pairwise video captioning data and our collected exemplar sentences to accomplish the model learning. Extensive experiments, in terms of semantic, syntactic, fluency, and diversity evaluations, clearly demonstrate our model capability to generate syntax-varied and semantics-coherent video captions that well imitate different exemplar sentences with enriched diversities. Code is available at https://github.com/yytzsy/Syntax-Customized-Video-Captioning.