Fine-grained image classification is one of the key basic tasks of computer vision. The appearance of traditional deep convolutional neural network (DCNN) combined with attention mechanism can focus on partial and local features of fine-grained images, but it still lacks the consideration of the embedding mode of different attention modules in the network, leading to the unsatisfactory result of classification model. To solve the above problems, three different attention mechanisms are introduced into the DCNN network (like ResNet, VGGNet, etc.), including SE, CBAM and ECA modules, so that DCNN could better focus on the key local features of salient regions in the image. At the same time, we adopt three different embedding modes of attention modules, including serial, residual and parallel modes, to further improve the performance of the classification model. The experimental results show that the three attention modules combined with three different embedding modes can improve the performance of DCNN network effectively. Moreover, compared with SE and ECA, CBAM has stronger feature extraction capability. Among them, the parallelly embedded CBAM can make the local information paid attention to by DCNN richer and more accurate, and bring the optimal effect for DCNN, which is 1.98% and 1.57% higher than that of original VGG16 and Resnet34 in CUB-200-2011 dataset, respectively. The visualization analysis also indicates that the attention modules can be easily embedded into DCNN networks, especially in the parallel mode, with stronger generality and universality.

Defect prediction approaches have been greatly contributing to software quality assurance activities such as code review or unit testing. Just-in-time defect prediction approaches are developed to predict whether a commit is a defect-inducing commit or not. Prior research has shown that commit-level prediction is not enough in terms of effort, and a defective commit may contain both defective and non-defective files. As the defect prediction community is promoting fine-grained granularity prediction approaches, we propose our novel class-level prediction, which is finer-grained than the file-level prediction, based on the files of the commits in this research. We designed our model for Python projects and tested it with ten open-source Python projects. We performed our experiment with two settings: setting with product metrics only and setting with product metrics plus commit information. Our investigation was conducted with three different classifiers and two validation strategies. We found that our model developed by random forest classifier performs the best, and commit information contributes significantly to the product metrics in 10-fold cross-validation. We also created a commit-based file-level prediction for the Python files which do not have the classes. The file-level model also showed a similar condition as the class-level model. However, the results showed a massive deviation in time-series validation for both levels and the challenge of predicting Python classes and files in a realistic scenario.

Fine-grained image classification is a difficult problem, and previous studies mainly overcome this problem by locating multiple discriminative regions in different scales and then aggregating complementary information explored from the located regions. However, locating discriminative regions introduces heavy overhead and is not suitable for real-world application. In this paper, we propose the recursive multi-scale channel-spatial attention module (RMCSAM) for addressing this problem. Following the experience of previous research on fine-grained image classification, RMCSAM explores multi-scale attentional information. However, the attentional information is explored by recursively refining the deep feature maps of a convolutional neural network (CNN) to better correspond to multi-scale channel-wise and spatial-wise attention, instead of localizing attention regions. In this way, RMCSAM provides a lightweight module that can be inserted into standard CNNs. Experimental results show that RMCSAM can improve the classification accuracy and attention capturing ability over baselines. Also, RMCSAM performs better than other state-of-the-art attention modules in fine-grained image classification, and is complementary to some state-of-the-art approaches for fine-grained image classification. Code is available at https://github.com/Dichao-Liu/Recursive-Multi-Scale-Channel-Spatial-Attention-Module.

Local discriminative regions play important roles in fine-grained image analysis tasks. How to locate local discriminative regions with only category label and learn discriminative representation from these regions have been hot spots. In our work, we propose Searching Discriminative Regions (SDR) and Learning Discriminative Regions (LDR) method to search and learn local discriminative regions in images. The SDR method adopts attention mechanism to iteratively search for high-response regions in images, and uses this as a clue to locate local discriminative regions. Moreover, the LDR method is proposed to learn compact within category and sparse between categories representation from the raw image and local images. Experimental results show that our proposed approach achieves excellent performance in both fine-grained image retrieval and classification tasks, which demonstrates its effectiveness.

Aiming at the problem that traditional text-level sentiment analysis methods usually ignore the emotional tendency corresponding to the object or attribute. In this paper, a novel two-stage fine-grained text-level sentiment analysis model based on syntactic rule matching and deep semantics is proposed. Based on analyzing the characteristics and difficulties of fine-grained sentiment analysis, a two-stage fine-grained sentiment analysis algorithm framework is constructed. In the first stage, the objects and its corresponding opinions are extracted based on syntactic rules matching to obtain preliminary objects and opinions. The second stage based on deep semantic network to extract more accurate objects and opinions. Aiming at the problem that the extraction result contains multiple objects and opinions to be matched, an object-opinion matching algorithm based on the minimum lexical separation distance is proposed to achieve accurate pairwise matching. Finally, the proposed algorithm is evaluated on several public datasets to demonstrate its practicality and effectiveness.

History data of edit operations are more beneficial than those stored in version control systems since they provide detailed information on how source code was changed. Meanwhile, a large number of recorded edit operations discourage developers and researchers from roughly understanding the changes. To assist with this task, it is desirable that they easily obtain traceability links for changed program elements over two source code snapshots before and after a code change. In this paper, we propose a graph representation called Operation History Graph (OHG), which presents code change information with such traceability links that are inferred from the history of edit operations. An OHG instance is generated by parsing any source code snapshot restored by edit histories and combining resultant abstract syntax trees (ASTs) into a single graph structure. To improve the performance of building graph instances, we avoided simply maintaining every program element. Any program element presenting the inner-structure of methods and non-changed elements are omitted. In addition, we adopted a lightweight static analysis for type name resolving to reduce required memory resource in the analysis while the accuracy of name resolving is preserved. Moreover, we assign a specific ID to each node and edge in the graph instance so that a part of the graph data can be separately stored and loaded on demand. These decisions make it feasible to build, manipulate, and store the graph with limited computer resources. To demonstrate the usefulness of the proposed operation history graph and verify whether detected traceability links are sufficient to reveal actual changes of program elements, we implemented tools to generate and manipulate OHG instances. The evaluation on graph generation performance shows that our tool can reduce the required computer resource as compared to another tool authors previously proposed. Moreover, the evaluation on traceability shows that OHG provides traceability links with sufficient accuracy as compared to the baseline approach using GumTree.

Location and feature representation of object's parts play key roles in fine-grained visual recognition. To promote the final recognition accuracy without any bounding boxes/part annotations, many studies adopt object location networks to propose bounding boxes/part annotations with only category labels, and then crop the images into partial images to help the classification network make the final decision. In our work, to propose more informative partial images and effectively extract discriminative features from the original and partial images, we propose a two-stage approach that can fuse the original features and partial features by evaluating and ranking the information of partial images. Experimental results show that our proposed approach achieves excellent performance on two benchmark datasets, which demonstrates its effectiveness.

Recording source code changes comes to be well recognized as an effective means for understanding the evolution of existing software and making its future changes efficient. Therefore, modern integrated development environments (IDEs) tend to employ tools that record fine-grained textual changes of source code. However, there is still no satisfactory tool that accurately records textual changes. We propose ChangeMacroRecorder that automatically and silently records all textual changes of source code and in real time correlates those textual changes with actions causing them while a programmer is writing and modifying it on the Eclipse's Java editor. The improvement with respect to the accuracy of recorded textual changes enables both programmers and researchers to exactly understand how the source code was evolved. This paper presents detailed information on how ChangeMacroRecorder achieves the accurate recording of textual changes and demonstrates how accurate textual changes were recorded in our experiment consisting of nine programming tasks.

The aim of this paper is to propose effective attentional regions for fine-grained visual recognition. Based on the Spatial Transformers' capability of spatial manipulation within networks, we propose an extension model, the Attention-Guided Spatial Transformer Networks (AG-STNs). This model can guide the Spatial Transformers with hard-coded attentional regions at first. Then such guidance can be turned off, and the network model will adjust the region learning in terms of the location and scale. Such adjustment is conditioned to the classification loss so that it is actually optimized for better recognition results. With this model, we are able to successfully capture detailed attentional information. Also, the AG-STNs are able to capture attentional information in multiple levels, and different levels of attentional information are complementary to each other in our experiments. A fusion of them brings better results.

Understanding which refactoring transformations were performed is in demand in modern software constructions. Traditionally, many researchers have been tackling understanding code changes with history data derived from version control systems. In those studies, problems of the traditional approach are pointed out, such as entanglement of multiple changes. To alleviate the problems, operation histories on IDEs' code editors are available as a new source of software evolution data nowadays. By replaying such histories, we can investigate past code changes in a fine-grained level. However, the prior studies did not provide enough evidence of their effectiveness for detecting refactoring transformations. This paper describes an experiment in which participants detect refactoring transformations performed by other participants after investigating the code changes with an operation-replay tool and diff tools. The results show that both approaches have their respective factors that pose misunderstanding and overlooking of refactoring transformations. Two negative factors on divided operations and generated compound operations were observed in the operation-based approach, whereas all the negative factors resulted from three problems on tangling, shadowing, and out-of-order of code changes in the difference-based approach. This paper also shows seven concrete examples of participants' mistakes in both approaches. These findings give us hints for improving existing tools for understanding code changes and detecting refactoring transformations.

Recently, numerous methods have been proposed to tackle the problem of fine-grained image classification. However, rare of them focus on the pre-processing step of image alignment. In this paper, we propose a new pre-processing method with the aim of reducing the variance of objects among the same class. As a result, the variance of objects between different classes will be more significant. The proposed approach consists of four procedures. The “parts” of the objects are firstly located. After that, the rotation angle and the bounding box could be obtained based on the spatial relationship of the “parts”. Finally, all the images are resized to similar sizes. The objects in the images possess the properties of translation, scale and rotation invariance after processed by the proposed method. Experiments on the CUB-200-2011 and CUB-200-2010 datasets have demonstrated that the proposed method could boost the recognition performance by serving as a pre-processing step of several popular classification algorithms.

In this paper, we propose a novel placement algorithm for mixed-grained reconfigurable architectures (MGRAs). MGRA consists of coarse-grained and fine-grained clusters, in order to implement a combined digital systems of high-speed data paths with multi-bit operands and random logic circuits for state machines and bit-wise operations. For accelerating simulated annealing based FPGA placement algorithm, range limiter has been proposed to control the distance of two blocks to be interchanged. However, it is not applicable to MGRAs due to the heterogeneous structure of MGRAs. Proposed range limiter using connection bounding box effectively keeps the size of range limiter to encourage moves across fine-grain blocks in non-adjacent clusters. From experimental results, the proposed method achieved 47.8% reduction of cost in the best case compared with conventional methods.

Reconfigurable architectures have emerged as an optimal choice for the hardware realization of digital signal processing (DSP) algorithms. Reconfigurable architecture is either fine-grained or coarse-grained depending on the granularity of reconfiguration used. The flexibility offered by fine-grained devices such as field programmable gate array (FPGA) comes at a significant cost of huge routing area, power consumption and speed overheads. To overcome these issues, several coarse-grained reconfigurable architectures have been proposed. In this paper, a scalable and hybrid dynamically reconfigurable architecture, HyDRA, is proposed for efficient hardware realization of computation intensive DSP algorithms. The proposed architecture is greatly influenced by reported VLSI architectures of a variety of DSP algorithms. It is designed using parameterized VHDL model which allows experimenting with a variety of design features by simply modifying some constants. The proposed architecture with 8×8 processing element array is synthesized using UMC 0.25µm and LF 150nm CMOS technologies respectively. For quantitative evaluation, the architecture is also realized using Xilinx Virtex-7 FPGA. The area and timing results are presented to provide an estimate of each block of the architecture. DSP algorithms such as 32-tap finite impulse response (FIR) filters, 16-point radix-2 single path delay feedback (R2SDF) fast fourier transform (FFT) and R2SDF discrete cosine transform (DCT) are mapped and routed on the proposed architecture.

Distributed power delivery is blooming in SoC power system because the fine-grained power management needs separate power sources to adjust each voltage island dynamically. In addition, dedicated power sources for critical circuit blocks can achieve better signal integrity. To extensively utilize the power modules when they are redundant and idle, this work applies the cooperation concept in SoC power management. The key controller is a mixed-signal estimator that executes the intelligent procedures, like real-time swap the power module depending on its loading and healthy condition, automatically configure the power system with phase interleaving, and support all the peripheral functions. To demonstrate the proposed concept, a prototype chip for voltage down-conversion is implemented. This chip contains four switched-inductor converter modules to emulate the cooperative power network. Each module is small therefore the power efficiency is not optimal for the heavy load. With the cooperation between power modules, the power efficiency is 88% for 300mA load, that is 8.5% higher than the single module operation.

Fine-grained visual categorization (FGVC) has drawn increasing attention as an emerging research field in recent years. In contrast to generic-domain visual recognition, FGVC is characterized by high intra-class and subtle inter-class variations. To distinguish conceptually and visually similar categories, highly discriminative visual features must be extracted. Moreover, FGVC has highly specialized and task-specific nature. It is not always easy to obtain a sufficiently large-scale training dataset. Therefore, the key to success in practical FGVC systems is to efficiently exploit discriminative features from a limited number of training examples. In this paper, we propose an efficient two-step dimensionality compression method to derive compact middle-level part-based features. To do this, we compare both space-first and feature-first convolution schemes and investigate their effectiveness. Our approach is based on simple linear algebra and analytic solutions, and is highly scalable compared with the current one-vs-one or one-vs-all approach, making it possible to quickly train middle-level features from a number of pairwise part regions. We experimentally show the effectiveness of our method using the standard Caltech-Birds and Stanford-Cars datasets.

What should be the ultimate form of the cloud computing environment? The solution should have two important features; “Fine-Granularity” and “Participation.” To realize an attractive and feasible solution with these features, we propose a “participating fine-grained cloud computing platform” that a large number of personal or small-company resource suppliers participate in, configure and provide cloud computing on. This enables users to be supplied with smaller units of resources such as computing, memory, content, and applications, in comparison with the traditional Infrastructure as a Service (IaaS). Furthermore, to search for nearby resources efficiently among the many available on the platform, we also propose Resource Breadcrumbs (RBC) as a key technology of our proposed platform to provide in-network guidance capability autonomously for users' queries. RBC allows supplier-nodes to distribute guidance information directed to themselves with dedicated control messages; in addition, the information can be logged along the trail of message from supplier to user. With this distributed information, users can to autonomously locate nearby resources. Distributed management also reduces computational load on the central database and enables a participating fine-grained cloud platform at lower cost.

In this paper, we propose a new low power nonvolatile counter unit based on Magnetic Tunnel Junction (MTJ) with fine-grained power gating. The proposed counter unit consists of only a single latch with two MTJs. We verify the basic operation and estimate the power consumption of the proposed counter unit. The operating power consumption of the proposed nonvolatile counter unit is smaller than the conventional one below 140 kHz. The power of the proposed unit is 74.6% smaller than the conventional one at low frequency.

With the aim to improve the effectiveness of SAWSDL service discovery, this paper proposes a novel discovery method for SAWSDL services, which is based on the matchmaking of so-called fine-grained data semantics that is defined via sets of atomic elements with built-in data types. The fine-grained data semantics can be obtained by a transformation algorithm that decomposes parameters at message level into a set of atomic elements, considering the characteristics of SAWSDL service structure and semantic annotations. Then, a matchmaking algorithm is proposed for the matching of fine-grained data semantics, which avoids the complex and expensive structural matching at the message level. The fine-grained data semantics is transparent to the specific data structure of message-level parameters, therefore, it can help to match successfully similar Web services with different data structures of parameters. Moreover, a comprehensive measure is proposed by considering together several important components of SAWSDL service descriptions at the same time. Finally, this method is evaluated on SAWSDL service discovery test collection SAWSDL-TC2 and compared with other SAWSDL matchmakers. The experimental results show that our method can improve the effectiveness of SAWSDL service discovery with low average query response time. The results imply that fine-grained parameters fit to represent the data semantics of SAWSDL services, especially when data structures of parameters are not important for semantics.

The security hash algorithm 512 (SHA-512), which is used to verify the integrity of a message, involves computational iterations on data. The huge computation delay generated in such iterations limits the entire throughput of the system and makes it difficult to pipeline the computation. We describe a way to pipeline the computation using fine-grained pipelining with balanced critical paths. In this method, one critical path is broken into two stages by using data forwarding. The other critical path is broken into three stages by using computation postponement. The resulting critical paths all have two adder-layers with some data movements, and thus are balanced. In addition, the method also allows register reduction. Also, the similarity in SHA-384 and SHA-512 are used for a multi-mode design, which can generate a message digest for both versions with the same throughput, but with only a small increase in hardware size. Experimental results show that our implementation achieved not only the best area performance rate (throughput divided by area), but also a higher throughput than almost all related work.

This paper describes the design and evaluation of a massively parallel processor base on Matrix architecture which is suitable for portable multimedia applications. The proposed architecture in this paper achieves 40 GOPS of 16-bit fixed-point additions at 200 MHz clock frequency and 250 mW power dissipation. In addition, 1 M-bit SRAM for data registers and 2,048 2-bit processing elements connected by a flexible switching network are integrated in 3.1 mm2 in 90 nm low-power CMOS technology. The energy-efficient Matrix architecture supports 2,048-way parallel operations and the programmable functions required for multimedia SoCs.