Intelligent control of self-driving vehicles based on adaptive sampling supervised actor-critic and human driving experience.

Due to the complexity of the driving environment and the dynamics of the behavior of traffic participants, self-driving in dense traffic flow is very challenging. Traditional methods usually rely on predefined rules, which are difficult to adapt to various driving scenarios. Deep reinforcement learning (DRL) shows advantages over rule-based methods in complex self-driving environments, demonstrating the great potential of intelligent decision-making. However, one of the problems of DRL is the inefficiency of exploration; typically, it requires a lot of trial and error to learn the optimal policy, which leads to its slow learning rate and makes it difficult for the agent to learn well-performing decision-making policies in self-driving scenarios. Inspired by the outstanding performance of supervised learning in classification tasks, we propose a self-driving intelligent control method that combines human driving experience and adaptive sampling supervised actor-critic algorithm. Unlike traditional DRL, we modified the learning process of the policy network by combining supervised learning and DRL and adding human driving experience to the learning samples to better guide the self-driving vehicle to learn the optimal policy through human driving experience and real-time human guidance. In addition, in order to make the agent learn more efficiently, we introduced real-time human guidance in its learning process, and an adaptive balanced sampling method was designed for improving the sampling performance. We also designed the reward function in detail for different evaluation indexes such as traffic efficiency, which further guides the agent to learn the self-driving intelligent control policy in a better way. The experimental results show that the method is able to control vehicles in complex traffic environments for self-driving tasks and exhibits better performance than other DRL methods.

Patient-derived organoid elucidates the identical clonal origin of bilateral breast cancer with diverse molecular subtypes.

Bilateral breast cancer (BBC), an infrequent breast cancer subtype, has primarily been studied in terms of incidence, prognosis, and through comparative analysis of synchronous (SBBC) and metachronous (MBBC) manifestations. The advent and application of organoid technology hold profound implications for tumor research and clinical management. This study represents the pioneering use of organoid models in BBC research. We established organoid lines from two surgical tumor specimens of a BBC patient, with one line undergoing detailed pathological and genomic analysis. The BBC organoid from the right breast demonstrated a marker expression profile of ER (-), PR (-), HER-2 (0), and Ki67 index 10%, indicating that it may derived from the TNBC tissue. Whole Exome Sequencing (WES) displayed consistent set of Top10 cancer driver genes affected by missense mutations, frameshift mutation, or splice site mutations in three tumor tissues and the organoid samples. The organoids' single nucleotide polymorphisms (SNPs) were more closely aligned with the TNBC tissue than other tumor tissues. Evolutionary analysis suggested that different tumor regions might evolve from a common ancestral layer. In this case, the development of BBC organoids indicated that simultaneous lesions with diverse molecular profiles shared a high degree of consistency in key tumor-driving mutations. These findings suggest the feasibility of generating BBC organoids representing various molecular types, accurately replicating significant markers and driver mutations of the originating tumor. Consequently, organoids serve as a valuable in vitro model for exploring treatment strategies and elucidating the underlying mechanisms of BBC.

Rectify representation bias in vision-language models for long-tailed recognition.

Natural data typically exhibits a long-tailed distribution, presenting great challenges for recognition tasks. Due to the extreme scarcity of training instances, tail classes often show inferior performance. In this paper, we investigate the problem within the trendy visual-language (VL) framework and find that the performance bottleneck mainly arises from the recognition confusion between tail classes and their highly correlated head classes. Building upon this observation, unlike previous research primarily emphasizing class frequency in addressing long-tailed issues, we take a novel perspective by incorporating a crucial additional factor namely class correlation. Specifically, we model the representation learning procedure for each sample as two parts, i.e., a special part that learns the unique properties of its own class and a common part that learns shared characteristics among classes. By analysis, we discover that the learning process of common representation is easily biased toward head classes. Because of the bias, the network may lean towards the biased common representation as classification criteria, rather than prioritizing the crucial information encapsulated within the specific representation, ultimately leading to recognition confusion. To solve the problem, based on the VL framework, we introduce the rectification contrastive term (ReCT) to rectify the representation bias, according to semantic hints and training status. Extensive experiments on three widely-used long-tailed datasets demonstrate the effectiveness of ReCT. On iNaturalist2018, it achieves an overall accuracy of 75.4%, surpassing the baseline by 3.6 points in a ResNet-50 visual backbone.

The Equalization Losses: Gradient-Driven Training for Long-tailed Object Recognition.

Long-tail distribution is widely spread in real-world applications. Due to the extremely small ratio of instances, tail categories often show inferior accuracy. In this paper, we find such performance bottleneck is mainly caused by the imbalanced gradients, which can be categorized into two parts: (1) positive part, deriving from the samples of the same category, and (2) negative part, contributed by other categories. Based on comprehensive experiments, it is also observed that the gradient ratio of accumulated positives to negatives is a good indicator to measure how balanced a category is trained. Inspired by this, we come up with a gradient-driven training mechanism to tackle the long-tail problem: re-balancing the positive/negative gradients dynamically according to current accumulative gradients, with a unified goal of achieving balance gradient ratios. Taking advantage of the simple and flexible gradient mechanism, we introduce a new family of gradient-driven loss functions, namely equalization losses. We conduct extensive experiments on a wide spectrum of visual tasks, including two-stage/single-stage long-tailed object detection (LVIS), long-tailed image classification (ImageNet-LT, Places-LT, iNaturalist), and long-tailed semantic segmentation (ADE20 K). Our method consistently outperforms the baseline models, demonstrating the effectiveness and generalization ability of the proposed equalization losses.

Research and implementation of variable-domain fuzzy PID intelligent control method based on Q-Learning for self-driving in complex scenarios.

In the control of the self-driving vehicles, PID controllers are widely used due to their simple structure and good stability. However, in complex self-driving scenarios such as curvature curves, car following, overtaking, etc., it is necessary to ensure the stable control accuracy of the vehicles. Some researchers used fuzzy PID to dynamically change the parameters of PID to ensure that the vehicle control remains in a stable state. It is difficult to ensure the control effect of the fuzzy controller when the size of the domain is not selected properly. This paper designs a variable-domain fuzzy PID intelligent control method based on Q-Learning to make the system robust and adaptable, which is dynamically changed the size of the domain to further ensure the control effect of the vehicle. The variable-domain fuzzy PID algorithm based on Q-Learning takes the error and the error rate of change as input and uses the Q-Learning method to learn the scaling factor online so as to achieve online PID parameters adjustment. The proposed method is verified on the Panosim simulation platform.The experiment shows that the accuracy is improved by 15% compared with the traditional fuzzy PID, which reflects the effectiveness of the algorithm.

PicassoNet: Searching Adaptive Architecture for Efficient Facial Landmark Localization.

Since recent facial landmark localization methods achieve satisfying accuracy, few of them enable fast inference speed, which, however, is critical in many real-world facial applications. Existing methods typically employ complicated network structure and predict all the key points through uniform computation, which is inefficient since individual facial part might take different computation to obtain the best performance. Taking both accuracy and efficiency into consideration, we propose the PicassoNet, a lightweight cascaded facial landmark detector with adaptive computation for individual facial part. Different from the conventional cascaded methods, PicassoNet integrates refinement submodules into a single network with group convolution, where each convolution group predicts landmarks from an individual facial part. Note that the groups' structures are flexible in the training process. Then, a novel grouping search algorithm is proposed to optimize the group division. With formulating the optimization as a network architecture search (NAS) problem, the grouping search adaptively allocates computation to each group and obtains an efficient structure. In addition, we propose a boundary-aware loss to optimize along tangent and normal of facial boundaries, instead of optimizing along horizontal and vertical as the conventional loss (L2, SmoothL1, WingLoss, and so on) do. The novel loss improves the joint locations of predicted keypoints. Experiments on three benchmark datasets AFLW, 300W, and WFLW show that the proposed method runs over 6× times faster than the state of the arts and meanwhile achieves comparable accuracy.

A review of the current state of the computer-aided diagnosis (CAD) systems for breast cancer diagnosis.

Breast cancer is one of the most common cancers affecting females worldwide. Early detection and diagnosis of breast cancer may aid in timely treatment, reducing the mortality rate to a great extent. To diagnose breast cancer, computer-aided diagnosis (CAD) systems employ a variety of imaging modalities such as mammography, computerized tomography, magnetic resonance imaging, ultrasound, and histological imaging. CAD and breast-imaging specialists are in high demand for early detection and diagnosis. This system has the potential to enhance the partiality of traditional histopathological image analysis. This review aims to highlight the recent advancements and the current state of CAD systems for breast cancer detection using different modalities.

Overexpression of PBK/TOPK relates to poor prognosis of patients with breast cancer: a retrospective analysis.

BACKGROUND: PDZ-binding kinase/T-lymphokine-activated killer cell-derived protein kinase (PBK/TOPK) is a potential prognostic indicator for patients with breast cancer. The objective of the present study was to explore the relationship between PBK/TOPK expression and clinicopathological indicators as well as the survival of patients with breast cancer. METHODS: Immunohistochemical staining was used to detect the expression of PBK/TOPK in 202 cases of breast cancer tissues. The relationship between PBK/TOPK and clinicopathological parameters was evaluated using Spearman's rank-order correlation. The difference in PBK/TOPK expression among different molecular types was analyzed with the chi-square test. Kaplan-Meier analysis was used to create a survival curve and the log rank test was used to analyze the overall survival (OS) and disease-free survival (DFS). Prognostic correlation was assessed using univariate and multivariate Cox regression analyses. RESULTS: Among 202 breast cancer samples, PBK/TOPK was expressed ("+" and "++") in 182 samples (90.1%). In addition, the histological grade, TNM stages, lymph node metastasis, estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER-2), and Ki-67 were positively associated with PBK/TOPK expression. With regard to the molecular type, the expression of PBK/TOPK is different. The expression level of PBK/TOPK was negatively correlated with both the OS and DFS of breast cancer patients. The difference in the above results is meaningful (P < 0.05). CONCLUSIONS: PBK/TOPK is overexpressed in breast cancer, and the expression is closely related to the clinicopathological characteristics of the disease. Breast cancer patients with high expression of PBK/TOPK have a poor prognosis. Therefore, healthcare providers can optimize breast cancer management using this indicator.

Circ_0061825 Acts as a miR-593-3p Sponge to Promote Breast Cancer Progression by Regulating FGFR3 Expression.

BACKGROUND: Breast cancer (BC) remains the most common malignancy among women. Circular RNAs (circRNAs) have been demonstrated to play important roles in human cancers, including BC. In this study, we sought to identify the precise parts of circ_0061825 (circRNA trefoil factor 1, circ_TFF1) in BC pathogenesis. METHODS: The expression levels of circ_0061825, miR-593-3p and fibroblast growth factor receptor 3 (FGFR3) were detected by quantitative real-time polymerase chain reaction (qRT-PCR) or Western blot. Circ_0061825 was characterized using ribonuclease (RNase) R digestion, actinomycin D and subcellular fractionation assays. Cell viability, colony formation, migration, invasion, cell cycle progression and apoptosis were evaluated using Cell Counting Kit-8 (CCK-8), colony formation, wound-healing, transwell and flow cytometry assays, respectively. Targeted relationships among circ_0061825, miR-593-3p and FGFR3 were determined by a dual-luciferase reporter assay. Animal studies were used to assess the impact of circ_0061825 in tumor growth in vivo. RESULTS: Our data indicated that circ_0061825 was overexpressed in BC tissues and cells, and it was mainly localized in the cytoplasm of BC cells. Circ_0061825 knockdown hampered BC cell viability, colony formation, migration, invasion, cell cycle progression and enhanced cell apoptosis in vitro and weakened tumor growth in vivo. Mechanistically, circ_0061825 functioned as a molecular sponge of miR-593-3p, and circ_0061825 knockdown repressed BC cell malignant progression in vitro by miR-593-3p. FGFR3 was a direct target of miR-593-3p, and circ_0061825 modulated FGFR3 expression through sponging miR-593-3p. Moreover, miR-593-3p overexpression hindered BC cell malignant progression in vitro by down-regulating FGFR3. CONCLUSION: Our current work provided evidence that circ_0061825, an up-regulated circRNA in BC, regulated BC malignant progression at least in part through targeting the miR-593-3p/FGFR3 axis, illuminating a novel therapeutic target for BC management.

Circular RNA-100219 promotes breast cancer progression by binding to microRNA-485-3p.

PURPOSE: To investigate whether circ-100219 could promote the proliferation and migration of breast cancer cells by upregulating NTRK3 after binding to microRNA-485-3p, thus participating in the development of breast cancer. METHODS: Breast cancer cell lines MCF-7, MDA-MB-231, MDA-MB-549 and human mammary epithelial cells were cultured. The expression levels of circ-100219, microRNA-485-3p and NTRK3 in breast cancer and paracancer tissues were determined using real-time quantitative polymerase chain reaction (RT-qPCR). The regulatory effects of circ-100219, microRNA-485-3p and NTRK3 on the proliferative and migratory capacities of breast cancer cells were assessed using cell counting kit-8 (CCK-8) and Transwell assay, respectively. Dual-luciferase reporter gene assay was conducted to determine the binding condition among circ-100219, microRNA-485-3p and NTRK3. Rescue experiments were performed in co-transfected breast cancer cells. RESULTS: RT-qPCR data showed that circ-100219 and NTRK3 were highly expressed, whereas microRNA-485-3p was lowly expressed in breast cancer tissues than those of paracancer tissues. Knockdown of circ-100219 in MCF-7 and MDA-MB-231 cells inhibited their proliferative and migratory capacities. On the contrary, microRNA-485-3p knockdown improved the proliferative and migratory capacities. Dual-luciferase reporter gene assay revealed that circ-100219 could bind to microRNA-485-3p and NTRK3 was the target gene of microRNA-485-3p. Western blot results elucidated that circ-100219 stabilized NTRK3 expression, whereas microRNA-485-3p degraded NTRK3 expression. Rescue experiments demonstrated that overexpression of NTRK3 could partially reverse the inhibited proliferative and migratory capacities induced by circ-100219 knockdown in MCF-7 and MDA-MB-231 cells. CONCLUSIONS: Overexpression of circ-10021 promotes the proliferative and migratory capacities of breast cancer cells by sponging microRNA-485-3p to upregulate the NTRK3 expression.

Probing primordial gravitational waves: Ali CMB Polarization Telescope.

In this paper, we will give a general introduction to the Ali CMB Polarization Telescope (AliCPT) project, which is a Sino-US joint project led by the Institute of High Energy Physics and involves many different institutes in China. It is the first ground-based Cosmic Microwave Background (CMB) polarization experiment in China and an integral part of China's Gravitational-wave Program. The main scientific goal of the AliCPT project is to probe the primordial gravitational waves (PGWs) originating from the very early Universe. The AliCPT project includes two stages. The first stage, referred to as AliCPT-1, is to build a telescope in the Ali region of Tibet at an altitude of 5250 meters. Once completed, it will be the highest ground-based CMB observatory in the world and will open a new window for probing PGWs in the northern hemisphere. The AliCPT-1 telescope is designed to have about 7000 transition-edge sensor detectors at 95 GHz and 150 GHz. The second stage is to have a more sensitive telescope (AliCPT-2) with more than 20 000 detectors. Our simulations show that AliCPT will improve the current constraint on the tensor-to-scalar ratio r by one order of magnitude with three years' observation. Besides the PGWs, AliCPT will also enable a precise measurement of the CMB rotation angle and provide a precise test of the CPT symmetry. We show that three years' observation will improve the current limit by two orders of magnitude.

Ground-to-satellite quantum teleportation.

An arbitrary unknown quantum state cannot be measured precisely or replicated perfectly. However, quantum teleportation enables unknown quantum states to be transferred reliably from one object to another over long distances, without physical travelling of the object itself. Long-distance teleportation is a fundamental element of protocols such as large-scale quantum networks and distributed quantum computation. But the distances over which transmission was achieved in previous teleportation experiments, which used optical fibres and terrestrial free-space channels, were limited to about 100 kilometres, owing to the photon loss of these channels. To realize a global-scale 'quantum internet' the range of quantum teleportation needs to be greatly extended. A promising way of doing so involves using satellite platforms and space-based links, which can connect two remote points on Earth with greatly reduced channel loss because most of the propagation path of the photons is in empty space. Here we report quantum teleportation of independent single-photon qubits from a ground observatory to a low-Earth-orbit satellite, through an uplink channel, over distances of up to 1,400 kilometres. To optimize the efficiency of the link and to counter the atmospheric turbulence in the uplink, we use a compact ultra-bright source of entangled photons, a narrow beam divergence and high-bandwidth and high-accuracy acquiring, pointing and tracking. We demonstrate successful quantum teleportation of six input states in mutually unbiased bases with an average fidelity of 0.80 ± 0.01, well above the optimal state-estimation fidelity on a single copy of a qubit (the classical limit). Our demonstration of a ground-to-satellite uplink for reliable and ultra-long-distance quantum teleportation is an essential step towards a global-scale quantum internet.

miRNA-21 promotes proliferation and invasion of triple-negative breast cancer cells through targeting PTEN.

MicroRNAs (miRNAs) are small single-stranded RNAs that bind to the 3'UTR of the mRNAs of target genes. They can target multiple genes and regulate translation or degradation of the mRNA. miRNAs target genes in a tissue-specific manner, and the role of a particular miRNA varies according to tumor origin or even subtype within the same cancer. This study evaluated the effect of miR-21 expression in triple-negative breast cancer (TNBC) tissues and MDA-MB-468, a cell line derived from TNBC tissues. miR-21 was consistently upregulated in TNBC and MDA-MB-468 cells compared to normal tissues. Inhibition of miR-21 by miR-21 antisense oligonucleotides decreased the proliferation, viability, and invasiveness of MDA-MB-468 cells and enhanced apoptosis. Furthermore, we confirmed that PTEN was downregulated by miR-21 in MDA-MB-468 cells. The results indicated that PTEN may mediate the oncogenic properties of miR-21 in TNBC. In summary, miR-21 was upregulated in TNBC tissues and cells, and promoted the proliferation and invasion of MDA-MB-468 cells, but negatively regulated the expression of PTEN protein. Inhibition of miR-21 or overexpression of PTEN protein could be promising strategies for the treatment of patients with TNBC.

Breast cancer in 30-year-old or younger patients: clinicopathologic characteristics and prognosis.

BACKGROUND: The number of 30-year-old or younger patients with breast cancer is increasing. The aim was to describe the clinicopathological features and prognosis of 30-year-old or younger patients with breast cancer. METHODS: We reviewed the records of 1,406 consecutive patients aged ≤50 years with first diagnosis of invasive breast cancer referred to surgery from March 2001 to March 2009. A total of 105 patients were aged ≤30 years (group I) and 1,301 were aged 31-50 years (group II). RESULTS: Compared with patients of group II, patients of group I had a higher percentage of tumors classified as estrogen receptors (ER) negative (P < 0.001) and progesterone receptors (PR) negative (P = 0.043), with a Ki-67 labeling index ≥20% of the cells (P = 0.011). There was no difference between the two groups for pT and pN, histology, endocrine therapy, and chemotherapy. The 5-year survival of group I was 67.5% as compared with 75.3% for group II (P = 0.003). CONCLUSIONS: Compared with patients aged between 31 and 50 years, patients aged ≤30 years have a greater chance of having an endocrine-unresponsive tumor and a significantly poor prognosis.