Recent advances in domain adaptation show that deep self-training presents a powerful means for unsupervised domain adaptation. These methods often involve an iterative process of predicting on target domain and then taking the confident predictions as pseudo-labels for retraining. However, since pseudo-labels can be noisy, self-training can put overconfident label belief on wrong classes, leading to deviated solutions with propagated errors. To address the problem, we propose a confidence regularized self-training (CRST) framework, formulated as regularized self-training. Our method treats pseudo-labels as continuous latent variables jointly optimized via alternating optimization. We propose two types of confidence regularization: label regularization (LR) and model regularization (MR). CRST-LR generates soft pseudo-labels while CRST-MR encourages the smoothness on network output. Extensive experiments on image classification and semantic segmentation show that CRSTs outperform their non-regularized counterpart with state-of-the-art performance. The code and models of this work are available at https://github.com/yzou2/CRST. Network retraining (b) Label regularized self-training (c) Model regularized self-training (a) Self-training without confidence regularization Backbone Network Car Person Bus Network retraining Network retraining Figure 1: Illustration of proposed confidence regularization. (a) Self-training without confidence regularization generates and retrains with hard pseudo-labels, resulting in sharp network output. (b) Label regularized self-training introduces soft pseudo-labels, therefore enabling outputs to be smooth. (c) Model regularized self-training also retrains with hard pseudo-labels, but incorporates a regularizer to directly promote output smoothness.More recently, self-training with networks emerged as a promising alternative towards domain adaptation [4,5,25,29,49,54,69]. Self-training iteratively generates a set of one-hot (or hard) pseudo-labels corresponding to large selection scores (i.e., prediction confidence) in target domain, and then retrains network based on these pseudo-labels with target data. Recently, [69] proposes class-balanced selftraining (CBST) and formulates self-training as a unified loss minimization with pseudo-labels that can be solved in an end-to-end manner. Instead of reducing domain gap by minimizing both the task loss and domain adversarial loss, the self-training loss implicitly encourages cross-domain feature alignment for each class by learning from both labeled source data and pseudo-labeled target data.Early work [29] shows that the essence of deep selftraining is entropy minimization -pushing network output to be as sharp as hard pseudo-label. However, 100% accuracy cannot always be guaranteed for pseudo-labels. Trusting all selected pseudo-labels as "ground truth" by encoding

The PharmMapper online tool is a web server for potential drug target identification by reversed pharmacophore matching the query compound against an in-house pharmacophore model database. The original version of PharmMapper includes more than 7000 target pharmacophores derived from complex crystal structures with corresponding protein target annotations. In this article, we present a new version of the PharmMapper web server, of which the backend pharmacophore database is six times larger than the earlier one, with a total of 23 236 proteins covering 16 159 druggable pharmacophore models and 51 431 ligandable pharmacophore models. The expanded target data cover 450 indications and 4800 molecular functions compared to 110 indications and 349 molecular functions in our last update. In addition, the new web server is united with the statistically meaningful ranking of the identified drug targets, which is achieved through the use of standard scores. It also features an improved user interface. The proposed web server is freely available at http://lilab.ecust.edu.cn/pharmmapper/.

In silico drug target identification, which includes many distinct algorithms for finding disease genes and proteins, is the first step in the drug discovery pipeline. When the 3D structures of the targets are available, the problem of target identification is usually converted to finding the best interaction mode between the potential target candidates and small molecule probes. Pharmacophore, which is the spatial arrangement of features essential for a molecule to interact with a specific target receptor, is an alternative method for achieving this goal apart from molecular docking method. PharmMapper server is a freely accessed web server designed to identify potential target candidates for the given small molecules (drugs, natural products or other newly discovered compounds with unidentified binding targets) using pharmacophore mapping approach. PharmMapper hosts a large, in-house repertoire of pharmacophore database (namely PharmTargetDB) annotated from all the targets information in TargetBank, BindingDB, DrugBank and potential drug target database, including over 7000 receptor-based pharmacophore models (covering over 1500 drug targets information). PharmMapper automatically finds the best mapping poses of the query molecule against all the pharmacophore models in PharmTargetDB and lists the top N best-fitted hits with appropriate target annotations, as well as respective molecule’s aligned poses are presented. Benefited from the highly efficient and robust triangle hashing mapping method, PharmMapper bears high throughput ability and only costs 1 h averagely to screen the whole PharmTargetDB. The protocol was successful in finding the proper targets among the top 300 pharmacophore candidates in the retrospective benchmarking test of tamoxifen. PharmMapper is available at http://59.78.96.61/pharmmapper.

Ferritins are spherical, cage-like proteins with nanocavities formed by multiple polypeptide subunits (four-helix bundles) that manage iron/oxygen chemistry. Catalytic coupling yields diferric oxo/hydroxo complexes at ferroxidase sites in maxi-ferritin subunits (24 subunits, 480 kDa; plants, animals, microorganisms). Oxidation occurs at the cavity surface of mini-ferritins/Dps proteins (12 subunits, 240 kDa; bacteria). Oxidation products are concentrated as minerals in the nanocavity for iron-protein cofactor synthesis (maxi-ferritins) or DNA protection (mini-ferritins). The protein cage and nanocavity characterize all ferritins, although amino acid sequences diverge, especially in bacteria. Catalytic oxidation/di-iron coupling in the protein cage (maxi-ferritins, 480 kDa; plants, bacteria and animal cell-specific isoforms) or on the cavity surface (mini-ferritins/Dps proteins, 280 kDa; bacteria) initiates mineralization. Gated pores (eight or four), symmetrically arranged, control iron flow. The multiple ferritin functions combine pore, channel, and catalytic functions in compact protein structures required for life and disease response.

Despite tremendous recent efforts, noninvasive sweat monitoring is still far from delivering its early analytical promise. Here, we describe a flexible epidermal microfluidic detection platform fabricated through hybridization of lithographic and screen-printed technologies, for efficient and fast sweat sampling and continuous, real-time electrochemical monitoring of glucose and lactate levels. This soft, skin-mounted device judiciously merges lab-on-a-chip and electrochemical detection technologies, integrated with a miniaturized flexible electronic board for real-time wireless data transmission to a mobile device. Modeling of the device design and sweat flow conditions allowed optimization of the sampling process and the microchannel layout for achieving attractive fluid dynamics and rapid filling of the detection reservoir (within 8 min from starting exercise). The wearable microdevice thus enabled efficient natural sweat pumping to the electrochemical detection chamber containing the enzyme-modified electrode transducers. The fabricated device can be easily mounted on the epidermis without hindrance to the wearer and displays resiliency against continuous mechanical deformation expected from such epidermal wear. Amperometric biosensing of lactate and glucose from the rapidly generated sweat, using the corresponding immobilized oxidase enzymes, was wirelessly monitored during cycling activity of different healthy subjects. This ability to monitor sweat glucose levels introduces new possibilities for effective diabetes management, while similar lactate monitoring paves the way for new wearable fitness applications. The new epidermal microfluidic electrochemical detection strategy represents an attractive alternative to recently reported colorimetric sweat-monitoring methods, and hence holds considerable promise for practical fitness or health monitoring applications.