Previous works utilized "smaller-norm-less-important" criterion to prune filters with smaller norm values in a convolutional neural network. In this paper, we analyze this norm-based criterion and point out that its effectiveness depends on two requirements that are not always met: (1) the norm deviation of the filters should be large; (2) the minimum norm of the filters should be small. To solve this problem, we propose a novel filter pruning method, namely Filter Pruning via Geometric Median (FPGM), to compress the model regardless of those two requirements. Unlike previous methods, FPGM compresses CNN models by pruning filters with redundancy, rather than those with "relatively less" importance. When applied to two image classification benchmarks, our method validates its usefulness and strengths. Notably, on CIFAR-10, FPGM reduces more than 52% FLOPs on ResNet-110 with even 2.69% relative accuracy improvement. Moreover, on ILSVRC-2012, FPGM reduces more than 42% FLOPs on ResNet-101 without top-5 accuracy drop, which has advanced the state-of-the-art. Code is publicly available on GitHub: https://github.com/he-y/filter-pruning-geometric-median

This review summarises the recent advances of covalent organic framework photocatalysts including structures and applications.

ObjectiveTo evaluate rates of serious organ specific immune-related adverse events, general adverse events related to immune activation, and adverse events consistent with musculoskeletal problems for anti-programmed cell death 1 (PD-1) drugs overall and compared with control treatments.DesignSystematic review and meta-analysis.Data sourcesMedline, Embase, Cochrane Library, Web of Science, and Scopus searched to 16 March 2017 and combined with data from ClinicalTrials.gov.Study selectionEligible studies included primary clinical trial data on patients with cancer with recurrent or metastatic disease.Data extractionThree independent investigators extracted data on adverse events from ClinicalTrials.gov and the published studies. Risk of bias was assessed using the Cochrane tool by three independent investigators.Results13 relevant studies were included; adverse event data were available on ClinicalTrials.gov for eight. Studies compared nivolumab (n=6), pembrolizumab (5), or atezolizumab (2) with chemotherapy (11), targeted drugs (1), or both (1). Serious organ specific immune-related adverse events were rare, but compared with standard treatment, rates of hypothyroidism (odds ratio 7.56, 95% confidence interval 4.53 to 12.61), pneumonitis (5.37, 2.73 to 10.56), colitis (2.88, 1.30 to 6.37), and hypophysitis (3.38, 1.02 to 11.08) were increased with anti-PD-1 drugs. Of the general adverse events related to immune activation, only the rate of rash (2.34, 2.73 to 10.56) increased. Incidence of fatigue (32%) and diarrhea (19%) were high but similar to control. Reporting of adverse events consistent with musculoskeletal problems was inconsistent; rates varied but were over 20% in some studies for arthraligia and back pain.ConclusionsOrgan specific immune-related adverse events are uncommon with anti-PD-1 drugs but the risk is increased compared with control treatments. General adverse events related to immune activation are largely similar. Adverse events consistent with musculoskeletal problems are inconsistently reported but adverse events may be common.

Traumatic articular cartilage injuries heal poorly and may predispose patients to the early onset of osteoarthritis. One current treatment relies on surgical delivery of autologous chondrocytes that are prepared, prior to implantation, through ex vivo cell expansion of cartilage biopsy cells. The requirement for cell expansion, however, is both complex and expensive and has proven to be a major hurdle in achieving a widespread adoption of the treatment. This study presents evidence that autologous chondrocyte implantation can be delivered without requiring ex vivo cell expansion. The proposed improvement relies on mechanical fragmentation of cartilage tissue sufficient to mobilize embedded chondrocytes via increased tissue surface area. Our outgrowth study, which was used to demonstrate chondrocyte migration and growth, indicated that fragmented cartilage tissue is a rich source for chondrocyte redistribution. The chondrocytes outgrown into 3-D scaffolds also formed cartilage-like tissue when implanted in SCID mice. Direct treatment of fullthickness chondral defects in goats using cartilage fragments on a resorbable scaffold produced hyaline-like repair tissue at 6 months. Thus, delivery of chondrocytes in the form of cartilage tissue fragments in conjunction with appropriate polymeric scaffolds provides a novel intraoperative approach for cell-based cartilage repair. ß

This article is the series of methodology of clinical prediction model construction (total 16 sections of this methodology series). The first section mainly introduces the concept, current application status, construction methods and processes, classification of clinical prediction models, and the necessary conditions for conducting such researches and the problems currently faced. The second episode of these series mainly concentrates on the screening method in multivariate regression analysis. The third section mainly introduces the construction method of prediction models based on Logistic regression and Nomogram drawing. The fourth episode mainly concentrates on Cox proportional hazards regression model and Nomogram drawing. The fifth Section of the series mainly introduces the calculation method of C-Statistics in the logistic regression model. The sixth section mainly introduces two common calculation methods for C-Index in Cox regression based on R. The seventh section focuses on the principle and calculation methods of Net Reclassification Index (NRI) using R. The eighth section focuses on the principle and calculation methods of IDI (Integrated Discrimination Index) using R. The ninth section continues to explore the evaluation method of clinical utility after predictive model construction: Decision Curve Analysis. The tenth section is a supplement to the previous section and mainly introduces the Decision Curve Analysis of survival outcome data. The eleventh section mainly discusses the external validation method of Logistic regression model. The twelfth mainly discusses the in-depth evaluation of Cox regression model based on R, including calculating the concordance index of discrimination (C-index) in the validation data set and drawing the calibration curve. The thirteenth section mainly introduces how to deal with the survival data outcome using competitive risk model with R. The fourteenth section mainly introduces how to draw the nomogram of the competitive risk model with R. The fifteenth section of the series mainly discusses the identification of outliers and the interpolation of missing values. The sixteenth section of the series mainly introduced the Zhou et al. Clinical prediction models with R