Institute of Computing Technology

Transfer learning aims at improving the performance of target learners on target domains by transferring the knowledge contained in different but related source domains. In this way, the dependence on a large number of target-domain data can be reduced for constructing target learners. Due to the wide application prospects, transfer learning has become a popular and promising area in machine learning. Although there are already some valuable and impressive surveys on transfer learning, these surveys introduce approaches in a relatively isolated way and lack the recent advances in transfer learning. Due to the rapid expansion of the transfer learning area, it is both necessary and challenging to comprehensively review the relevant studies. This survey attempts to connect and systematize the existing transfer learning research studies, as well as to summarize and interpret the mechanisms and the strategies of transfer learning in a comprehensive way, which may help readers have a better understanding of the current research status and ideas. Unlike previous surveys, this survey article reviews more than 40 representative transfer learning approaches, especially homogeneous transfer learning approaches, from the perspectives of data and model. The applications of transfer learning are also briefly introduced. In order to show the performance of different transfer learning models, over 20 representative transfer learning models are used for experiments. The models are performed on three different data sets, that is, Amazon Reviews, Reuters-21578, and Office-31, and the experimental results demonstrate the importance of selecting appropriate transfer learning models for different applications in practice.

Internet of Things (IoT) has provided a promising opportunity to build powerful industrial systems and applications by leveraging the growing ubiquity of radio-frequency identification (RFID), and wireless, mobile, and sensor devices. A wide range of industrial IoT applications have been developed and deployed in recent years. In an effort to understand the development of IoT in industries, this paper reviews the current research of IoT, key enabling technologies, major IoT applications in industries, and identifies research trends and challenges. A main contribution of this review paper is that it summarizes the current state-of-the-art IoT in industries systematically.

The genome of the japonica subspecies of rice, an important cereal and model monocot, was sequenced and assembled by whole-genome shotgun sequencing. The assembled sequence covers 93% of the 420-megabase genome. Gene predictions on the assembled sequence suggest that the genome contains 32,000 to 50,000 genes. Homologs of 98% of the known maize, wheat, and barley proteins are found in rice. Synteny and gene homology between rice and the other cereal genomes are extensive, whereas synteny with Arabidopsis is limited. Assignment of candidate rice orthologs to Arabidopsis genes is possible in many cases. The rice genome sequence provides a foundation for the improvement of cereals, our most important crops.

In 2008 we published the first set of guidelines for standardizing research in autophagy. Since then, research on this topic has continued to accelerate, and many new scientists have entered the field. Our knowledge base and relevant new technologies have also been expanding. Accordingly, it is important to update these guidelines for monitoring autophagy in different organisms. Various reviews have described the range of assays that have been used for this purpose. Nevertheless, there continues to be confusion regarding acceptable methods to measure autophagy, especially in multicellular eukaryotes. A key point that needs to be emphasized is that there is a difference between measurements that monitor the numbers or volume of autophagic elements (e.g., autophagosomes or autolysosomes) at any stage of the autophagic process vs. those that measure flux through the autophagy pathway (i.e., the complete process); thus, a block in macroautophagy that results in autophagosome accumulation needs to be differentiated from stimuli that result in increased autophagic activity, defined as increased autophagy induction coupled with increased delivery to, and degradation within, lysosomes (in most higher eukaryotes and some protists such as Dictyostelium) or the vacuole (in plants and fungi). In other words, it is especially important that investigators new to the field understand that the appearance of more autophagosomes does not necessarily equate with more autophagy. In fact, in many cases, autophagosomes accumulate because of a block in trafficking to lysosomes without a concomitant change in autophagosome biogenesis, whereas an increase in autolysosomes may reflect a reduction in degradative activity. Here, we present a set of guidelines for the selection and interpretation of methods for use by investigators who aim to examine macroautophagy and related processes, as well as for reviewers who need to provide realistic and reasonable critiques of papers that are focused on these processes. These guidelines are not meant to be a formulaic set of rules, because the appropriate assays depend in part on the question being asked and the system being used. In addition, we emphasize that no individual assay is guaranteed to be the most appropriate one in every situation, and we strongly recommend the use of multiple assays to monitor autophagy. In these guidelines, we consider these various methods of assessing autophagy and what information can, or cannot, be obtained from them. Finally, by discussing the merits and limits of particular autophagy assays, we hope to encourage technical innovation in the field.

Siamese network based trackers formulate tracking as convolutional feature cross-correlation between target template and searching region. However, Siamese trackers still have accuracy gap compared with state-of-the-art algorithms and they cannot take advantage of feature from deep networks, such as ResNet-50 or deeper. In this work we prove the core reason comes from the lack of strict translation invariance. By comprehensive theoretical analysis and experimental validations, we break this restriction through a simple yet effective spatial aware sampling strategy and successfully train a ResNet-driven Siamese tracker with significant performance gain. Moreover, we propose a new model architecture to perform depth-wise and layer-wise aggregations, which not only further improves the accuracy but also reduces the model size. We conduct extensive ablation studies to demonstrate the effectiveness of the proposed tracker, which obtains currently the best results on four large tracking benchmarks, including OTB2015, VOT2018, UAV123, and LaSOT. Our model will be released to facilitate further studies based on this problem.

It is a challenge to classify protein-coding or non-coding transcripts, especially those re-constructed from high-throughput sequencing data of poorly annotated species. This study developed and evaluated a powerful signature tool, Coding-Non-Coding Index (CNCI), by profiling adjoining nucleotide triplets to effectively distinguish protein-coding and non-coding sequences independent of known annotations. CNCI is effective for classifying incomplete transcripts and sense-antisense pairs. The implementation of CNCI offered highly accurate classification of transcripts assembled from whole-transcriptome sequencing data in a cross-species manner, that demonstrated gene evolutionary divergence between vertebrates, and invertebrates, or between plants, and provided a long non-coding RNA catalog of orangutan. CNCI software is available at http://www.bioinfo.org/software/cnci.

We present an integrated stand-alone software package named KaKs_Calculator 2.0 as an updated version. It incorporates 17 methods for the calculation of nonsynonymous and synonymous substitution rates; among them, we added our modified versions of several widely used methods as the gamma series including gamma-NG, gamma-LWL, gamma-MLWL, gamma-LPB, gamma-MLPB, gamma-YN and gamma-MYN, which have been demonstrated to perform better under certain conditions than their original forms and are not implemented in the previous version. The package is readily used for the identification of positively selected sites based on a sliding window across the sequences of interests in 5' to 3' direction of protein-coding sequences, and have improved the overall performance on sequence analysis for evolution studies. A toolbox, including C++ and Java source code and executable files on both Windows and Linux platforms together with a user instruction, is downloadable from the website for academic purpose at https://sourceforge.net/projects/kakscalculator2/.

Gene set enrichment (GSE) analysis plays an essential role in extracting biological insight from genome-scale experiments. ORA (overrepresentation analysis), FCS (functional class scoring), and PT (pathway topology) approaches are three generations of GSE methods along the timeline of development. Previous versions of KOBAS provided services based on just the ORA method. Here we presented version 3.0 of KOBAS, which is named KOBAS-i (short for KOBAS intelligent version). It introduced a novel machine learning-based method we published earlier, CGPS, which incorporates seven FCS tools and two PT tools into a single ensemble score and intelligently prioritizes the relevant biological pathways. In addition, KOBAS has expanded the downstream exploratory visualization for selecting and understanding the enriched results. The tool constructs a novel view of cirFunMap, which presents different enriched terms and their correlations in a landscape. Finally, based on the previous version's framework, KOBAS increased the number of supported species from 1327 to 5944. For an easier local run, it also provides a prebuilt Docker image that requires no installation, as a supplementary to the source code version. KOBAS can be freely accessed at http://kobas.cbi.pku.edu.cn, and a mirror site is available at http://bioinfo.org/kobas.

Y Artificial intelligence (AI) coupled with promising machine learning (ML) techniques well known from computer science is broadly affecting many aspects of various fields including science and technology, industry, and even our day-to-day life. The ML techniques have been developed to analyze high-throughput data with a view to obtaining useful insights, categorizing, predicting, and making evidence-based decisions in novel ways, which will promote the growth of novel applications and fuel the sustainable booming of AI. This paper undertakes a comprehensive survey on the development and application of AI in different aspects of fundamental sciences, including information science, mathematics, medical science, materials science, geoscience, life science, physics, and chemistry. The challenges that each discipline of science meets, and the potentials of AI techniques to handle these challenges, are discussed in detail. Moreover, we shed light on new research trends entailing the integration of AI into each scientific discipline. The aim of this paper is to provide a broad research guideline on fundamental sciences with potential infusion of AI, to help motivate researchers to deeply understand the state-of-the-art applications of AI-based fundamental sciences, and thereby to help promote the continuous development of these fundamental sciences.

Abstract KaKs_Calculator is a software package that calculates nonsynonymous (Ka) and synonymous (Ks) substitution rates through model selection and model averaging. Since existing methods for this estimation adopt their specific mutation (substitution) models that consider different evolutionary features, leading to diverse estimates, KaKs_Calculator implements a set of candidate models in a maximum likelihood framework and adopts the Akaike information criterion to measure fitness between models and data, aiming to include as many features as needed for accurately capturing evolutionary information in protein-coding sequences. In addition, several existing methods for calculating Ka and Ks are also incorporated into this software. KaKs_Calculator, including source codes, compiled executables, and documentation, is freely available for academic use at http://evolution.genomics.org.cn/software.htm.

For a target task where the labeled data are unavailable, domain adaptation can transfer a learner from a different source domain. Previous deep domain adaptation methods mainly learn a global domain shift, i.e., align the global source and target distributions without considering the relationships between two subdomains within the same category of different domains, leading to unsatisfying transfer learning performance without capturing the fine-grained information. Recently, more and more researchers pay attention to subdomain adaptation that focuses on accurately aligning the distributions of the relevant subdomains. However, most of them are adversarial methods that contain several loss functions and converge slowly. Based on this, we present a deep subdomain adaptation network (DSAN) that learns a transfer network by aligning the relevant subdomain distributions of domain-specific layer activations across different domains based on a local maximum mean discrepancy (LMMD). Our DSAN is very simple but effective, which does not need adversarial training and converges fast. The adaptation can be achieved easily with most feedforward network models by extending them with LMMD loss, which can be trained efficiently via backpropagation. Experiments demonstrate that DSAN can achieve remarkable results on both object recognition tasks and digit classification tasks. Our code will be available at https://github.com/easezyc/deep-transfer-learning.

Existing state-of-the-art salient object detection networks rely on aggregating multi-level features of pre-trained convolutional neural networks (CNNs). However, compared to high-level features, low-level features contribute less to performance. Meanwhile, they raise more computational cost because of their larger spatial resolutions. In this paper, we propose a novel Cascaded Partial Decoder (CPD) framework for fast and accurate salient object detection. On the one hand, the framework constructs partial decoder which discards larger resolution features of shallow layers for acceleration. On the other hand, we observe that integrating features of deep layers will obtain relatively precise saliency map. Therefore we directly utilize generated saliency map to recurrently optimize features of deep layers. This strategy efficiently suppresses distractors in the features and significantly improves their representation ability. Experiments conducted on five benchmark datasets exhibit that the proposed model not only achieves state-of-the-art but also runs much faster than existing models. Besides, we apply the proposed framework to optimize existing multi-level feature aggregation models and significantly improve their efficiency and accuracy.

Uncovering the topics within short texts, such as tweets and instant messages, has become an important task for many content analysis applications. However, directly applying conventional topic models (e.g. LDA and PLSA) on such short texts may not work well. The fundamental reason lies in that conventional topic models implicitly capture the document-level word co-occurrence patterns to reveal topics, and thus suffer from the severe data sparsity in short documents. In this paper, we propose a novel way for modeling topics in short texts, referred as biterm topic model (BTM). Specifically, in BTM we learn the topics by directly modeling the generation of word co-occurrence patterns (i.e. biterms) in the whole corpus. The major advantages of BTM are that 1) BTM explicitly models the word co-occurrence patterns to enhance the topic learning; and 2) BTM uses the aggregated patterns in the whole corpus for learning topics to solve the problem of sparse word co-occurrence patterns at document-level. We carry out extensive experiments on real-world short text collections. The results demonstrate that our approach can discover more prominent and coherent topics, and significantly outperform baseline methods on several evaluation metrics. Furthermore, we find that BTM can outperform LDA even on normal texts, showing the potential generality and wider usage of the new topic model.

Most of existing salient object detection models have achieved great progress by aggregating multi-level features extracted from convolutional neural networks. However, because of the different receptive fields of different convolutional layers, there exists big differences between features generated by these layers. Common feature fusion strategies (addition or concatenation) ignore these differences and may cause suboptimal solutions. In this paper, we propose the F3Net to solve above problem, which mainly consists of cross feature module (CFM) and cascaded feedback decoder (CFD) trained by minimizing a new pixel position aware loss (PPA). Specifically, CFM aims to selectively aggregate multi-level features. Different from addition and concatenation, CFM adaptively selects complementary components from input features before fusion, which can effectively avoid introducing too much redundant information that may destroy the original features. Besides, CFD adopts a multi-stage feedback mechanism, where features closed to supervision will be introduced to the output of previous layers to supplement them and eliminate the differences between features. These refined features will go through multiple similar iterations before generating the final saliency maps. Furthermore, different from binary cross entropy, the proposed PPA loss doesn't treat pixels equally, which can synthesize the local structure information of a pixel to guide the network to focus more on local details. Hard pixels from boundaries or error-prone parts will be given more attention to emphasize their importance. F3Net is able to segment salient object regions accurately and provide clear local details. Comprehensive experiments on five benchmark datasets demonstrate that F3Net outperforms state-of-the-art approaches on six evaluation metrics. Code will be released at https://github.com/weijun88/F3Net.

We report improved whole-genome shotgun sequences for the genomes of indica and japonica rice, both with multimegabase contiguity, or almost 1,000-fold improvement over the drafts of 2002. Tested against a nonredundant collection of 19,079 full-length cDNAs, 97.7% of the genes are aligned, without fragmentation, to the mapped super-scaffolds of one or the other genome. We introduce a gene identification procedure for plants that does not rely on similarity to known genes to remove erroneous predictions resulting from transposable elements. Using the available EST data to adjust for residual errors in the predictions, the estimated gene count is at least 38,000-40,000. Only 2%-3% of the genes are unique to any one subspecies, comparable to the amount of sequence that might still be missing. Despite this lack of variation in gene content, there is enormous variation in the intergenic regions. At least a quarter of the two sequences could not be aligned, and where they could be aligned, single nucleotide polymorphism (SNP) rates varied from as little as 3.0 SNP/kb in the coding regions to 27.6 SNP/kb in the transposable elements. A more inclusive new approach for analyzing duplication history is introduced here. It reveals an ancient whole-genome duplication, a recent segmental duplication on Chromosomes 11 and 12, and massive ongoing individual gene duplications. We find 18 distinct pairs of duplicated segments that cover 65.7% of the genome; 17 of these pairs date back to a common time before the divergence of the grasses. More important, ongoing individual gene duplications provide a never-ending source of raw material for gene genesis and are major contributors to the differences between members of the grass family.

Inspired by Weber's Law, this paper proposes a simple, yet very powerful and robust local descriptor, called the Weber Local Descriptor (WLD). It is based on the fact that human perception of a pattern depends not only on the change of a stimulus (such as sound, lighting) but also on the original intensity of the stimulus. Specifically, WLD consists of two components: differential excitation and orientation. The differential excitation component is a function of the ratio between two terms: One is the relative intensity differences of a current pixel against its neighbors, the other is the intensity of the current pixel. The orientation component is the gradient orientation of the current pixel. For a given image, we use the two components to construct a concatenated WLD histogram. Experimental results on the Brodatz and KTH-TIPS2-a texture databases show that WLD impressively outperforms the other widely used descriptors (e.g., Gabor and SIFT). In addition, experimental results on human face detection also show a promising performance comparable to the best known results on the MIT+CMU frontal face test set, the AR face data set, and the CMU profile test set.

In this paper, we describe the acquisition and contents of a large-scale Chinese face database: the CAS-PEAL face database. The goals of creating the CAS-PEAL face database include the following: 1) providing the worldwide researchers of face recognition with different sources of variations, particularly pose, expression, accessories, and lighting (PEAL), and exhaustive ground-truth information in one uniform database; 2) advancing the state-of-the-art face recognition technologies aiming at practical applications by using off-the-shelf imaging equipment and by designing normal face variations in the database; and 3) providing a large-scale face database of Mongolian. Currently, the CAS-PEAL face database contains 99 594 images of 1040 individuals (595 males and 445 females). A total of nine cameras are mounted horizontally on an arc arm to simultaneously capture images across different poses. Each subject is asked to look straight ahead, up, and down to obtain 27 images in three shots. Five facial expressions, six accessories, and 15 lighting changes are also included in the database. A selected subset of the database (CAS-PEAL-R1, containing 30 863 images of the 1040 subjects) is available to other researchers now. We discuss the evaluation protocol based on the CAS-PEAL-R1 database and present the performance of four algorithms as a baseline to do the following: 1) elementarily assess the difficulty of the database for face recognition algorithms; 2) preference evaluation results for researchers using the database; and 3) identify the strengths and weaknesses of the commonly used algorithms.

Since its renaissance, deep learning has been widely used in various medical imaging tasks and has achieved remarkable success in many medical imaging applications, thereby propelling us into the so-called artificial intelligence (AI) era. It is known that the success of AI is mostly attributed to the availability of big data with annotations for a single task and the advances in high performance computing. However, medical imaging presents unique challenges that confront deep learning approaches. In this survey paper, we first present traits of medical imaging, highlight both clinical needs and technical challenges in medical imaging, and describe how emerging trends in deep learning are addressing these issues. We cover the topics of network architecture, sparse and noisy labels, federating learning, interpretability, uncertainty quantification, etc. Then, we present several case studies that are commonly found in clinical practice, including digital pathology and chest, brain, cardiovascular, and abdominal imaging. Rather than presenting an exhaustive literature survey, we instead describe some prominent research highlights related to these case study applications. We conclude with a discussion and presentation of promising future directions.

In recent years, neural network accelerators have been shown to achieve both high energy efficiency and high performance for a broad application scope within the important category of recognition and mining applications.

In this paper, we present a new hashing method to learn compact binary codes for highly efficient image retrieval on large-scale datasets. While the complex image appearance variations still pose a great challenge to reliable retrieval, in light of the recent progress of Convolutional Neural Networks (CNNs) in learning robust image representation on various vision tasks, this paper proposes a novel Deep Supervised Hashing (DSH) method to learn compact similarity-preserving binary code for the huge body of image data. Specifically, we devise a CNN architecture that takes pairs of images (similar/dissimilar) as training inputs and encourages the output of each image to approximate discrete values (e.g. +1/-1). To this end, a loss function is elaborately designed to maximize the discriminability of the output space by encoding the supervised information from the input image pairs, and simultaneously imposing regularization on the real-valued outputs to approximate the desired discrete values. For image retrieval, new-coming query images can be easily encoded by propagating through the network and then quantizing the network outputs to binary codes representation. Extensive experiments on two large scale datasets CIFAR-10 and NUS-WIDE show the promising performance of our method compared with the state-of-the-arts.