Mutual Correlation Network for few-shot learning.

Most metric-based Few-Shot Learning (FSL) methods focus on learning good embeddings of images. However, these methods either lack the ability to explore the cross-correlation (i.e., correlated information) between image pairs or explore limited consensus among the correlation map constrained by the limited receptive field of CNN. We propose a Mutual Correlation Network (MCNet) to explore global consensus among the correlation map by using the self-attention mechanism which has a global receptive field. Our MCNet contains two core modules: (1) a multi-level embedding module that generates multi-level embeddings for an image pair which capture hierarchical semantics, and (2) a mutual correlation module that refines correlation map of two embeddings and generates more robust relational embeddings. Extensive experiments show that our MCNet achieves competitive results on four widely-used few-shot classification benchmarks miniImageNet, tieredImageNet, CUB-200-2011, and CIFAR-FS. Code is available at https://github.com/DRGreat/MCNet.

Mechanistic causes of sign epistasis and its applications.

Mapping genetic variations to phenotypic variations poses a significant challenge, as mutations often combine unexpectedly, diverging from assumed additive effects even in the same environment. These interactions are known as epistasis or genetic interactions. Sign epistasis, as a specific type of epistasis, involves a complete reversal of mutation effects within altered genetic backgrounds, presenting a substantial hurdle to phenotype prediction. Despite its importance, there is a limited systematic overview of the mechanistic causes of sign epistasis. This review explores the mechanistic causes, highlighting its occurrence in signalling cascades, peaked fitness landscapes, and physical interactions. Moving beyond theoretical discussions, we delve into the practical applications of sign epistasis in agriculture, evolution, and antibiotic resistance. In conclusion, this review aims to enhance the comprehension of sign epistasis and molecular dynamics, anticipating future endeavours in systematic biology engineering that leverage the knowledge of sign epistasis.

DVSRNet: Deep Video Super-Resolution Based on Progressive Deformable Alignment and Temporal-Sparse Enhancement.

Video super-resolution (VSR) is used to compose high-resolution (HR) video from low-resolution video. Recently, the deformable alignment-based VSR methods are becoming increasingly popular. In these methods, the features extracted from video are aligned to eliminate the motion error targeting high super-resolution (SR) quality. However, these methods often suffer from misalignment and the lack of enough temporal information to compose HR frames, which accordingly induce artifacts in the SR result. In this article, we design a deep VSR network (DVSRNet) based on the proposed progressive deformable alignment (PDA) module and temporal-sparse enhancement (TSE) module. Specifically, the PDA module is designed to accurately align features and to eliminate artifacts via the bidirectional information propagation. The TSE module is constructed to further eliminate artifacts and to generate clear details for the HR frame. In addition, we construct a lightweight deep optical flow network (OFNet) to obtain the bidirectional optical flows for the implementation of the PDA module. Moreover, two new loss functions are designed for our proposed method. The first one is adopted in OFNet and the second one is constructed to guarantee the generation of sharp and clear details for the HR frames. The experimental results demonstrate that our method performs better than the state-of-the-art methods.

A Generic Strategy to Stabilize Wide Bandgap Perovskites for Efficient Tandem Solar Cells.

Efficient wide bandgap (WBG) perovskite solar cells (PSCs) are essential for fully maximizing the potential of tandem solar cells. However, these cells currently face challenges such as high photovoltage losses and the presence of phase segregation, which impede the attainment of their expected efficiency and stability. Herein, the root cause of halide segregation is investigated, uncovering a close association with the presence of locally aggregated lead iodide (PbI2 ), particularly at the perovskite/C60 interface. Kelvin-probe atomic force microscopy results indicate that the remaining PbI2 at the interface leads to potential electrical differences between the domain surface and boundaries, which drives the formation of halide segregation. By reacting the surface PbI2 residue with ethanediamine dihydroiodide (EDAI2 ) at proper temperature, it is possible to effectively mitigate the phase segregation. By applying this surface reaction strategy in WBG inverted cells, a notable improvement of ≈100 mV is achieved in photovoltage over a wide range of WBG cells (1.67-1.78 eV), resulting in a champion efficiency of 23.1% (certified 22.95%) for 1.67 eV cells and 19.7% (certified 18.81%) for 1.75 eV cells. Furthermore, efficiency of 26.1% is demonstrated in a monolithic all-perovskite tandem cell.

Silencing long noncoding RNA LINC01133 suppresses pancreatic cancer through regulation of microRNA-1299-dependent IGF2BP3.

The deregulation of long noncoding RNAs (lncRNAs) holds great potential in the treatment of multiple cancers, including pancreatic cancer (PC). However, the specific molecular mechanisms by which LINC01133 contributes to pancreatic cancer remain unknown. Subsequent to bioinformatics analysis, we predicted and analyzed differentially expressed lncRNAs, microRNAs, and genes in pancreatic cancer. We determined the expression patterns of LINC01133, miR-1299, and insulin-like growth factor 2 mRNA binding protein 3 (IGF2BP3) in pancreatic cancer cells, and validated their interactions through luciferase reporter and RNA immunoprecipitation assays. We implemented loss-of-function and gain-of-function experiments for LINC01133, miR-1299, and IGF2BP3 to assay their potential effects on pancreatic cancer cell functions. We observed high expression of LINC01133 and IGF2BP3, but low expression of miR-1299, in pancreatic cancer cells. Furthermore, we found that LINC01133 enhances IGF2BP3 through binding with miR-1299. Silencing LINC01133 or IGF2BP3 and/or overexpressing miR-1299 limited pancreatic cancer cell proliferation, invasion, epithelial-mesenchymal transition, and suppressed tumorigenic abilities in mice lacking T cells (nude mice). Overall, our findings identified that silencing LINC01133 downregulates IGF2BP3 by upregulating miR-1299 expression, ultimately leading to the prevention of pancreatic cancer.

SegT: Separated edge-guidance transformer network for polyp segmentation.

Accurate segmentation of colonoscopic polyps is considered a fundamental step in medical image analysis and surgical interventions. Many recent studies have made improvements based on the encoder-decoder framework, which can effectively segment diverse polyps. Such improvements mainly aim to enhance local features by using global features and applying attention methods. However, relying only on the global information of the final encoder block can result in losing local regional features in the intermediate layer. In addition, determining the edges between benign regions and polyps could be a challenging task. To address the aforementioned issues, we propose a novel separated edge-guidance transformer (SegT) network that aims to build an effective polyp segmentation model. A transformer encoder that learns a more robust representation than existing convolutional neural network-based approaches was specifically applied. To determine the precise segmentation of polyps, we utilize a separated edge-guidance module consisting of separator and edge-guidance blocks. The separator block is a two-stream operator to highlight edges between the background and foreground, whereas the edge-guidance block lies behind both streams to strengthen the understanding of the edge. Lastly, an innovative cascade fusion module was used and fused the refined multi-level features. To evaluate the effectiveness of SegT, we conducted experiments with five challenging public datasets, and the proposed model achieved state-of-the-art performance.

Rotating Lorentz Force Magnetic Bearings' Dynamics Modeling and Adaptive Controller Design.

To address the issues of our agile satellites' poor attitude maneuverability, low pointing stability, and pointing inaccuracy, this paper proposes a new type of stabilized platform based on seven-degree-of-freedom Lorentz force magnetic levitation. Furthermore, in this study, we designed an adaptive controller based on the RBF neural network for the rotating magnetic bearing, which can improve the pointing accuracy of satellite loads. To begin, the advanced features of the new platform are described in comparison with the traditional electromechanical platform, and the structural characteristics and working principle of the platform are clarified. The significance of rotating magnetic bearings in improving load pointing accuracy is also clarified, and its rotor dynamics model is established to provide the input and output equations. The adaptive controller based on the RBF neural network is designed for the needs of high accuracy of the load pointing, high stability, and strong robustness of the system, and the current feedback inner loop is added to improve the system stiffness and rapidity. The final simulation results show that, when compared to the PID controller and robust sliding mode controller, the controller's pointing accuracy and anti-interference ability are greatly improved, and the system robustness is strong, which can effectively improve the pointing accuracy and pointing stability of the satellite/payload, as well as provide a powerful means of solving related problems in the fields of laser communication, high score detection, and so on.

Topical borneol relieves nonhistaminergic pruritus via targeting TRPA1 and TRPM8 channels in peripheral nerve terminals of mice.

Borneol has been used successfully for the treatment of itchy skin in traditional Chinese medicine. However, the antipruritic effect of borneol has rarely been studied, and the mechanism is unclear. Here, we showed that topical application of borneol on skin substantially suppressed pruritogen chloroquine- and compound 48/80-induced itching in mice. The potential targets of borneol, including transient receptor potential cation channel subfamily V member 3 (TRPV3), transient receptor potential cation channel subfamily A member 1 (TRPA1), transient receptor potential cation channel subfamily M member 8 (TRPM8), and gamma-aminobutyric acid type A (GABAA) receptor were pharmacologically inhibited or genetically knocked out one by one in mouse. Itching behavior studies demonstrated that the antipruritic effect of borneol is largely independent of TRPV3 and GABAA receptor, and TRPA1 and TRPM8 channels are responsible for a major portion of the effect of borneol on chloroquine-induced nonhistaminergic itching. Borneol activates TRPM8 and inhibits TRPA1 in sensory neurons of mice. Topical co-application of TRPA1 antagonist and TRPM8 agonist mimicked the effect of borneol on chloroquine-induced itching. Intrathecal injection of a group II metabotropic glutamate receptor antagonist partially attenuated the effect of borneol and completely abolished the effect of TRPM8 agonist on chloroquine-induced itching, suggesting that a spinal glutamatergic mechanism is involved. In contrast, the effect of borneol on compound 48/80-induced histaminergic itching occurs through TRPA1-and TRPM8-independent mechanisms. Our work demonstrates that borneol is an effective topical itch reliever, and TRPA1 inhibition and TRPM8 activation in peripheral nerve terminals account for its antipruritic effect.

From De-collectivization to Re-collectivization: New Transformation Trend in Agriculture Production in Taicang City of the Eastern China from the Perspective of 'Governmentality'.

The continuous development of agricultural technologies and produces trade and updated state reforms strongly shape the dominant organizing styles of local agriculture production. Since the end of the 1970s, rural China has witnessed a drastic agriculture transformation featured by de-collectivization, and recently in Taicang, a developed city in the eastern China, a kind of new cooperative farms have replaced the smallholders by pooling households' contracted farmland, signaling a new shift tendency towards agriculture recollectivization that is distinctive compared to other countries. Given the state governance's leadership in promoting this latest transformation, drawing on the theoretical concept of 'governmentality' that is powerful in understanding how to govern the society to meet certain objectives, this paper examines the accurate processes of agriculture re-collectivization in Taicang City based on the methods of semi-structured interview and participant observation. Our research shows that the China's tax-sharing and relevant reforms in the 1990s framed a new governance structure in line with the new mode of 'governmentality' closely linked to 'advanced liberalism', creating the responsibility shift from central to local governments and a set of evaluation technologies, and guiding various actors to engage in the agriculture re-collectivization practice in Taicang. Importantly, the considerable subsidies and the impressive extension services in terms of farming personnel training and agricultural techniques diffusion underpin the normal running of cooperative farms and contribute to the local implementation of national strategy of agriculture modernization. We argue that the combination of distinctive institutional arrangement of rural land owned by village collectives in China and the burgeoning local economy allows the developed areas to re-collectivizing the agriculture production for ensuring grain self-sufficiency rather than capturing more share in global agriculture products trade, and that the cooperative farms in our case have been deeply integrated into the 'modern procurement system' and become more sensitive to market fluctuation, requiring rural extension service to transform towards farms' capacity building for maintaining the long-term development momentum.

An invasive zone in human liver cancer identified by Stereo-seq promotes hepatocyte-tumor cell crosstalk, local immunosuppression and tumor progression.

Dissecting and understanding the cancer ecosystem, especially that around the tumor margins, which have strong implications for tumor cell infiltration and invasion, are essential for exploring the mechanisms of tumor metastasis and developing effective new treatments. Using a novel tumor border scanning and digitization model enabled by nanoscale resolution-SpaTial Enhanced REsolution Omics-sequencing (Stereo-seq), we identified a 500 µm-wide zone centered around the tumor border in patients with liver cancer, referred to as "the invasive zone". We detected strong immunosuppression, metabolic reprogramming, and severely damaged hepatocytes in this zone. We also identified a subpopulation of damaged hepatocytes with increased expression of serum amyloid A1 and A2 (referred to collectively as SAAs) located close to the border on the paratumor side. Overexpression of CXCL6 in adjacent malignant cells could induce activation of the JAK-STAT3 pathway in nearby hepatocytes, which subsequently caused SAAs' overexpression in these hepatocytes. Furthermore, overexpression and secretion of SAAs by hepatocytes in the invasive zone could lead to the recruitment of macrophages and M2 polarization, further promoting local immunosuppression, potentially resulting in tumor progression. Clinical association analysis in additional five independent cohorts of patients with primary and secondary liver cancer (n = 423) showed that patients with overexpression of SAAs in the invasive zone had a worse prognosis. Further in vivo experiments using mouse liver tumor models in situ confirmed that the knockdown of genes encoding SAAs in hepatocytes decreased macrophage accumulation around the tumor border and delayed tumor growth. The identification and characterization of a novel invasive zone in human cancer patients not only add an important layer of understanding regarding the mechanisms of tumor invasion and metastasis, but may also pave the way for developing novel therapeutic strategies for advanced liver cancer and other solid tumors.

Immune checkpoint B7-H3 is a therapeutic vulnerability in prostate cancer harboring PTEN and TP53 deficiencies.

Checkpoint immunotherapy has yielded meaningful responses across many cancers but has shown modest efficacy in advanced prostate cancer. B7 homolog 3 protein (B7-H3/CD276) is an immune checkpoint molecule and has emerged as a promising therapeutic target. However, much remains to be understood regarding B7-H3's role in cancer progression, predictive biomarkers for B7-H3-targeted therapy, and combinatorial strategies. Our multi-omics analyses identified B7-H3 as one of the most abundant immune checkpoints in prostate tumors containing PTEN and TP53 genetic inactivation. Here, we sought in vivo genetic evidence for, and mechanistic understanding of, the role of B7-H3 in PTEN/TP53-deficient prostate cancer. We found that loss of PTEN and TP53 induced B7-H3 expression by activating transcriptional factor Sp1. Prostate-specific deletion of Cd276 resulted in delayed tumor progression and reversed the suppression of tumor-infiltrating T cells and NK cells in Pten/Trp53 genetically engineered mouse models. Furthermore, we tested the efficacy of the B7-H3 inhibitor in preclinical models of castration-resistant prostate cancer (CRPC). We demonstrated that enriched regulatory T cells and elevated programmed cell death ligand 1 (PD-L1) in myeloid cells hinder the therapeutic efficacy of B7-H3 inhibition in prostate tumors. Last, we showed that B7-H3 inhibition combined with blockade of PD-L1 or cytotoxic T lymphocyte-associated protein 4 (CTLA-4) achieved durable antitumor effects and had curative potential in a PTEN/TP53-deficient CRPC model. Given that B7-H3-targeted therapies have been evaluated in early clinical trials, our studies provide insights into the potential of biomarker-driven combinatorial immunotherapy targeting B7-H3 in prostate cancer, among other malignancies.

METTL3-mediated m6A modification of lnc KCNQ1OT1 promotes doxorubicin resistance in breast cancer by regulating miR-103a-3p/MDR1 axis.

Doxorubicin (DOX) resistance in breast cancer (BC) poses a huge challenge for the therapeutic effect on BC. Lnc KCNQ1OT1 play crucial roles in chemotherapy resistance. However, the role and mechanism of lnc KCNQ1OT1 in DOX resistance BC have not been investigated, which merits further exploration. Based on MCF-7 and MDA-MB-231 cells, MCF-7/DOX and MDA-MB-231/DOX cells were established using gradient concentrations of DOX. IC50 values and cell viability were determined using MTT. Cell proliferation was investigated by colony formation. Flow cytometry was performed to examine cell apoptosis and cell cycle. Gene expression was examined using qRT-PCR and western blot. The interactions among METTL3, lnc KCNQ1OT1, miR-103a-3p, and MDR1 were validated with MeRIP-qPCR, RIP, and dual-luciferase reporter gene assays. The results showed that Lnc KCNQ1OT1 was highly expressed in DOX-resistant BC cells, and lnc KCNQ1OT1 depletion could enhance DOX sensitivity in BC cells and DOX-resistant BC cells. Besides, lnc KCNQ1OT1 was modulated by MELLT3 in the manner of m6A modification. MiR-103a-3p could interact with lnc KCNQ1OT1 and MDR1. Overexpression of MDR1 abolished the impacts of lnc KCNQ1OT1 depletion on DOX resistance in BC. In conclusion, our results unveiled that in BC cells and DOX-resistant BC cells, lnc KCNQ1OT1 could be mediated by METTL3 through m6A modification to elevate and stabilize its expression, further inhibiting miR-103a-3p/MDR1 axis to promote DOX resistance, which might provide novel thought to overcome DOX resistance in BC.

Evaluating prognostic value of biliary stone in intrahepatic cholangiocarcinoma by propensity score matching analysis.

Background: To explore biliary tract stone (BTS) as prognostic factors of intrahepatic cholangiocarcinoma (ICC). Methods: Clinical data of 985 ICC patients were classified into no BTS group and BTS group-subgrouped into hepatolithiasis (HL) and non-hepatolithiasis (NHL) group. Propensity score matching was utilized to mitigate baseline characteristics. Preoperative peripheral inflammation parameters (PPIP) were further investigated. Immunostaining of CD3, CD4, CD8, CD68, PD1 and PD-L1 were conducted. Results: Overall survival (OS) of patients without BTS surpassed BTS group (P = 0.040) while no difference of time to recurrence (TTR) was observed (P = 0.146). HL group had shorter OS and TTR than HL-matched group (P < 0.001 and P = 0.017, respectively) and survival time of NHL group differed not with NHL-matched group (P > 0.05). PPIP like neutrophils to lymphocytes ratio (NLR), platelet to lymphocyte ratio (PLR) and systemic immune inflammation (SII) of HL group exceeded no BTS group or NHL group (all P < 0.05). Associations of PPIP and tumorous immunocytes differed vastly among HL group, NHL group and no BTS group. Tumorous CD4+/CD3+ ratio and PD1+/CD3+ ratio of HL group surpassed those in no BTS group (P = 0.036 and P < 0.001, respectively) and NHL group (P = 0.015 and 0.002, respectively). Para-tumorous CD68+ macrophages exceeded that in tumor samples of HL group (P < 0.001). No difference of CD8+/CD3+ lymphocyte ratio and PD-L1 rank were detected. Conclusions: Hepatolithiasis, rather than extra-hepatic biliary stone, is a poor prognostic indicator of ICC. Immunotherapy is promising in treating HL-related ICC.

The temporal and spatial characteristics and influencing factors of CO2 emissions from municipal solid waste in China.

Understanding the temporal and spatial characteristics of carbon dioxide (CO2) emissions from municipal solid waste (MSW) and a quantitative evaluation of the contribution rate of the factors influencing the changes in CO2 emissions are important for pollution and emission reduction and the realization of the "double carbon" goal. This study analyzed the spatial and temporal evolution of waste generation and treatment based on panel data from 31 Chinese provinces over the past 15 years and then applied the logarithmic mean Divisia index (LMDI) model to study the driving factors of CO2 emissions from MSW. China's MSW production and CO2 emissions displayed a rising trend, and the overall CO2 emissions showed a geographical pattern of being high in the east and low in the west. Carbon emission intensity, economic output, urbanization level, and population size were positive factors that increased CO2 emissions. The most important factors driving CO2 emissions were carbon emission intensity and economic output, with cumulative contribution rates of 55.29% and 47.91%, respectively. Solid waste emission intensity was a negative factor in reducing CO2 emissions, with a cumulative contribution rate of -24.52%. These results have important implications for the design of policies to reduce CO2 emissions from MSW.

Synthetic miR-26a mimics delivered by tumor exosomes repress hepatocellular carcinoma through downregulating lymphoid enhancer factor 1.

BACKGROUND: The dysregulation of exosomal microRNAs plays an important role in the progression of hepatocarcinogenesis. In this study, we investigated the therapeutic potential of synthetic exosomal miR-26a against HCC cells and explored the feasibility of tumor-derived exosomes as drug delivery vehicles. METHODS: Proliferation and migration assays were performed to examine the effects of miR-26a on HCC in vitro. The direct target gene of miR-26a was identified through miRecords analysis and target validation. The transferring efficiency and anti-HCC effect of exosomes with different origin were studied and the optimal miR-26a delivery method was established and verified in vitro and in vivo. In addition, the relationships between prognosis of HCC patients and miR-26a expression in HCC serum and exosomes were retrospectively analyzed. RESULTS: Here, we found that tumor cell-derived exosomes were taken in preferentially by HCC cells and promoted HCC progression through Wnt pathway by low-density lipoprotein receptor-related protein 6 (LRP6). HCC cells with vacuolar protein sorting-associated protein 35 knocked down were adopted to generate engineered LRP6-exosomes. The engineered HCC-derived exosomes loading miR-26a inhibited HCC progression in vitro and in vivo effectively. Overexpression of miR-26a impaired the growth and migration of HCC by targeting lymphoid enhancer factor 1 (LEF1). Moreover, low expression of exosomal miR-26a was an independent prognostic factor for recurrence and survival in HCC patients. CONCLUSIONS: Our findings suggested the exosomal miR-26a could serve as a non-invasive prognostic marker for HCC patients. Genetically modified tumor-derived exosomes showed preferable transfection efficiency but reduced Wnt activity, which provides a novel therapeutic strategy for HCC.

Neural pathway connectivity and discharge changes between M1 and STN in hemiparkinsonian rats.

Alterations of electrophysiological activities, such as changed spike firing rates, reshaping the firing patterns, and aberrant frequency oscillations between the subthalamic nucleus (STN) and the primary motor cortex (M1), are thought to contribute to motor impairment in Parkinson's disease (PD). However, the alterations of electrophysiological characteristics of STN and M1 in PD are still unclear, especially under specific treadmill movement. To examine the relationship between electrophysiological activity in the STN-M1 pathway, extracellular spike trains and local field potential (LFPs) of STN and M1 were simultaneously recorded during resting and movement in unilateral 6-hydroxydopamine (6-OHDA) lesioned rats. The results showed that the identified STN neurons and M1 neurons exhibited abnormal neuronal activity after dopamine loss. The dopamine depletion altered the LFP power in STN and M1 whatever in rest or movement states. Furthermore, the enhanced synchronization of LFP oscillations after dopamine loss was found in 12-35 Hz (beta frequencies) between the STN and M1 during rest and movement. In addition, STN neurons were phase-locked firing to M1 oscillations at 12-35 Hz during rest epochs in 6-OHDA lesioned rats. The dopamine depletion also impaired the anatomical connectivity between the M1 and STN by injecting anterograde neuroanatomical tracing virus into M1 in control and PD rats. Collectively, impairment of' electrophysiological activity and anatomical connectivity in the M1-STN pathway may be the basis for dysfunction of the cortico-basal ganglia circuit, correlating with motor symptoms of PD.

Radiative Decay Width of J/ψ→γη_{(2)} from N_{f}=2 Lattice QCD.

The large radiative production rate for pseudoscalar mesons in the J/ψ radiative decay remains elusive. We present the first lattice QCD calculation of partial decay width of J/ψ radiatively decaying into η_{(2)}, the SU(2) flavor singlet pseudoscalar meson, which confirms QCD U_{A}(1) anomaly enhancement to the coupling of gluons with flavor singlet pseudoscalar mesons. The lattice simulation is carried out using N_{f}=2 lattice QCD gauge configurations at the pion mass m_{π}≈350 MeV. In particular, the distillation method has been utilized to calculate light quark loops. The results are reported here with the mass m_{η_{(2)}}=718(8) MeV and the decay width Γ(J/ψ→γη_{(2)})=0.385(45) keV. By assuming the dominance of U_{A}(1) anomaly and flavor singlet-octet mixing angle θ=-24.5°, the production rates for the physical η and η^{'} in J/ψ radiative decay are predicted to be 1.15(14)×10^{-3} and 4.49(53)×10^{-3}, respectively, which agree well with the experimental measurement data. Our study manifests the potential of lattice QCD studies on the light hadron production in J/ψ radiative decays.

A numerical tool for assessing human thermal safety and thermal comfort in cold-weather activities.

This paper describes a newly developed software tool to evaluate human thermal safety and thermal comfort in cold-weather activities aimed at guiding users to arrange activity plans and select appropriate clothing ensembles. The software inputs include conditions of activity, environment, human body, and clothing ensemble. It outputs physiological temperatures, cold injury risks, thermal sensations, and thermal comforts in intuitive ways like cloud maps and curves. The software tool is characterized by (1) integration of a thermoregulatory model that predicts human thermophysiological responses under exercise conditions in cold environments, (2) the functions of clothing ensemble database and individual parameter database, (3) the human centric outputs that directly reflect human physiological and mental status, and (4) the user-friendly operation interface and output interface, as well as a wide applicability. The software is validated with human test studies covering ambient temperatures from - 30.6 to 5 °C, clothing ensembles from 1.34 to 3.20 clo, and activity intensities from 2 to 9 Mets. The average prediction RMSEs of core temperature, mean skin temperature, thermal sensation, and thermal comfort are 0.16 °C, 0.45 °C, 0.58, and 1.41, respectively. The software is an advanced expansion to current standards and guidance of cold exposure assessment and a meaningful tool for the fields of occupational health care, cold protection, and environmental ergonomics.

Empowering the Diversity and Individuality of Option: Residual Soft Option Critic Framework.

Extracting temporal abstraction (option), which empowers the action space, is a crucial challenge in hierarchical reinforcement learning. Under a well-structured action space, decision-making agents can probe more deeply in the searching or plan efficiently through pruning irrelevant action candidates. However, automatically capturing a well-performed temporal abstraction is a nontrivial challenge due to its insufficient exploration and inadequate functionality. We consider alleviating this challenge from two perspectives, i.e., diversity and individuality. For the aspect of diversity, we propose a maximum entropy model based on ensembled options to encourage exploration. For the aspect of individuality, we propose to distinguish each option accurately, utilizing mutual formation minimization, so that each option can better express and function. We name our framework as an ensemble with soft option (ESO) critics. Furthermore, the residual algorithm (RA) with a bidirectional target network is introduced to stabilize bootstrapping, yielding a residual version of ESO. We provide detailed analysis for extensive experiments, which shows that our method boosts performance in commonly used continuous control tasks.