Broad Learning System (BLS) that aims to offer an alternative way of learning in deep structure is proposed in this paper. Deep structure and learning suffer from a time-consuming training process because of a large number of connecting parameters in filters and layers. Moreover, it encounters a complete retraining process if the structure is not sufficient to model the system. The BLS is established in the form of a flat network, where the original inputs are transferred and placed as "mapped features" in feature nodes and the structure is expanded in wide sense in the "enhancement nodes." The incremental learning algorithms are developed for fast remodeling in broad expansion without a retraining process if the network deems to be expanded. Two incremental learning algorithms are given for both the increment of the feature nodes (or filters in deep structure) and the increment of the enhancement nodes. The designed model and algorithms are very versatile for selecting a model rapidly. In addition, another incremental learning is developed for a system that has been modeled encounters a new incoming input. Specifically, the system can be remodeled in an incremental way without the entire retraining from the beginning. Satisfactory result for model reduction using singular value decomposition is conducted to simplify the final structure. Compared with existing deep neural networks, experimental results on the Modified National Institute of Standards and Technology database and NYU NORB object recognition dataset benchmark data demonstrate the effectiveness of the proposed BLS.

No abstract

No abstract

Background Immune‐therapy with anti‐PD1 inhibitors, such as pembrolizumab, is revolutionizing the treatment of non‐small cell lung cancers (NSCLC). However, identifying patients for the potential therapeutic response and predicting therapy resistance and early relapse remains a challenge. Methods Between 2016 and 2018, 60 patients were treated with pembrolizumab, among who 12 NSCLC patients had both baseline (before treatment) and serial (on treatment) periodical circulating tumor DNA (ctDNA) samples. Those samples were sequenced on a 329 pan cancer‐related gene panel. Analyses of tumor burden, blood tumor mutational burden (bTMB), maximum somatic allele frequency (MSAF), and tumor clonal structure were performed in association with clinical response. Candidate resistance mutations involved in relapse and metastases were further investigated. Results ctDNA was detected and mutational profiling was performed for each patient. Those with a high baseline bTMB level showed significantly improved progression‐free survival (PFS) after pembrolizumab treatment. Tumor burden and therapeutic response significantly correlated with the MSAF instead of the bTMB. Clone analysis detected tumor progression about 2‐4 months ahead of computed tomography (CT) scan. One mutation in gene PTCH1 (Protein patched homolog 1) and two acquired anti‐PD1 candidate resistance mutations of gene B2M (β2 microglobulin) were identified in association with distant metastasis. The evolutionary tree of a representative patient was also described. Conclusion This pilot study showed that MSAF could be another good indicator of therapeutic response, and clonal analysis could be clinically useful in monitoring clonal dynamics and detecting remote metastasis and early relapse.

No abstract