The maintainability of source code is a key quality characteristic for software quality. Many approaches have been proposed to quantitatively measure code maintainability. Such approaches rely heavily on code metrics, e.g., the number of Lines of Code and McCabe’s Cyclomatic Complexity. The employed code metrics are essentially statistics regarding code elements, e.g., the numbers of tokens, lines, references, and branch statements. However, natural language in source code, especially identifiers, is rarely exploited by such approaches. As a result, replacing meaningful identifiers with nonsense tokens would not significantly influence their outputs, although the replacement should have significantly reduced code maintainability. To this end, in this paper, we propose a novel approach (called DeepM) to measure code maintainability by exploiting the lexical semantics of text in source code. DeepM leverages deep learning techniques (e.g., LSTM and attention mechanism) to exploit these lexical semantics in measuring code maintainability. Another key rationale of DeepM is that measuring code maintainability is complex and often far beyond the capabilities of statistics or simple heuristics. Consequently, DeepM leverages deep learning techniques to automatically select useful features from complex and lengthy inputs and to construct a complex mapping (rather than simple heuristics) from the input to the output (code maintainability index). DeepM is evaluated on a manually-assessed dataset. The evaluation results suggest that DeepM is accurate, and it generates the same rankings of code maintainability as those of experienced programmers on 87.5% of manually ranked pairs of Java classes.

A threshold signature is a special digital signature in which the N-signer share the private key x and can construct a valid signature for any subset of the included t-signer, but less than t-signer cannot obtain any information. Considering the breakthrough achievements of threshold ECDSA signature and threshold Schnorr signature, the existing threshold SM2 signature is still limited to two parties or based on the honest majority setting, there is no more effective solution for the multiparty case. To make the SM2 signature have more flexible application scenarios, promote the application of the SM2 signature scheme in the blockchain system and secure cryptocurrency wallets. This paper designs a non-interactive threshold SM2 signature scheme based on partially homomorphic encryption and zero-knowledge proof. Only the last round requires the message input, so make our scheme non-interactive, and the pre-signing process takes 2 rounds of communication to complete after the key generation. We allow arbitrary threshold t\le n and design a key update strategy. It can achieve security with identifiable abort under the malicious majority, which means that if the signature process fails, we can find the failed party. Performance analysis shows that the computation and communication costs of the pre-signing process grows linearly with the parties, and it is only 1/3 of the Canetti’s threshold ECDSA (CCS'20).

Session-based recommendation is a popular research topic that aims to predict users’ next possible interactive item by exploiting anonymous sessions. The existing studies mainly focus on making predictions by considering users’ single interactive behavior. Some recent efforts have been made to exploit multiple interactive behaviors, but they generally ignore the influences of different interactive behaviors and the noise in interactive sequences. To address these problems, we propose a behavior-aware graph neural network for session-based recommendation. First, different interactive sequences are modeled as directed graphs. Thus, the item representations are learned via graph neural networks. Then, a sparse self-attention module is designed to remove the noise in behavior sequences. Finally, the representations of different behavior sequences are aggregated with the gating mechanism to obtain the session representations. Experimental results on two public datasets show that our proposed method outperforms all competitive baselines. The source code is available at the website of GitHub.

Attention mechanism has become a widely researched method to improve the performance of convolutional neural networks (CNNs). Most of the researches focus on designing channel-wise and spatial-wise attention modules but neglect the importance of unique information on each feature, which is critical for deciding both “what” and “where” to focus. In this paper, a feature-wise attention module is proposed, which can give each feature of the input feature map an attention weight. Specifically, the module is based on the well-known surround suppression in the discipline of neuroscience, and it consists of two sub-modules, Minus-Square-Add (MSA) operation and a group of learnable non-linear mapping functions. The MSA imitates the surround suppression and defines an energy function which can be applied to each feature to measure its importance. The group of non-linear functions refines the energy calculated by the MSA to more reasonable values. By these two sub-modules, feature-wise attention can be well captured. Meanwhile, due to the simple structure and few parameters of the two sub-modules, the proposed module can easily be almost integrated into any CNN. To verify the performance and effectiveness of the proposed module, several experiments were conducted on the Cifar10, Cifar100, Cinic10, and Tiny-ImageNet datasets, respectively. The experimental results demonstrate that the proposed module is flexible and effective for CNNs to improve their performance.

Heterogeneous information network (HIN) has recently been widely adopted to describe complex graph structure in recommendation systems, proving its effectiveness in modeling complex graph data. Although existing HIN-based recommendation studies have achieved great success by performing message propagation between connected nodes on the defined metapaths, they have the following major limitations. Existing works mainly convert heterogeneous graphs into homogeneous graphs via defining metapaths, which are not expressive enough to capture more complicated dependency relationships involved on the metapath. Besides, the heterogeneous information is more likely to be provided by item attributes while social relations between users are not adequately considered. To tackle these limitations, we propose a novel social recommendation model MPISR, which models MetaPath Interaction for Social Recommendation on heterogeneous information network. Specifically, our model first learns the initial node representation through a pretraining module, and then identifies potential social friends and item relations based on their similarity to construct a unified HIN. We then develop the two-way encoder module with similarity encoder and instance encoder to capture the similarity collaborative signals and relational dependency on different metapaths. Extensive experiments on five real datasets demonstrate the effectiveness of our method.

Oblivious Cross-Tags (OXT) [1] is the first efficient searchable encryption (SE) protocol for conjunctive queries in a single-writer single-reader framework. However, it also has a trade-off between security and efficiency by leaking partial database information to the server. Recent attacks on these SE schemes show that the leakages from these SE schemes can be used to recover the content of queried keywords. To solve this problem, Lai et al. [2] propose Hidden Cross-Tags (HXT), which reduces the access pattern leakage from Keyword Pair Result Pattern (KPRP) to Whole Result Pattern (WRP). However, the WRP leakage can also be used to recover some additional contents of queried keywords. This paper proposes Improved Cross-Tags (IXT), an efficient searchable encryption protocol that achieves access and searches pattern hiding based on the labeled private set intersection. We also prove the proposed labeled private set intersection (PSI) protocol is secure against semi-honest adversaries, and IXT is L^{\prime }-semi-honest secure (L^{\prime } is leakage function). Finally, we do experiments to compare IXT with HXT. The experimental results show that the storage overhead and computation overhead of the search phase at the client-side in IXT is much lower than those in HXT. Meanwhile, the experimental results also show that IXT is scalable and can be applied to various sizes of datasets.

Multiparty private set intersection (PSI) allows several parties, each holding a set of elements, to jointly compute the intersection without leaking any additional information. With the development of cloud computing, PSI has a wide range of applications in privacy protection. However, it is complex to build an efficient and reliable scheme to protect user privacy.

To address this issue, we propose EMPSI, an efficient PSI (with cardinality) protocol in a multiparty setting. EMPSI avoids using heavy cryptographic primitives (mainly rely on symmetric-key encryption) to achieve better performance. In addition, both PSI and PSI with the cardinality of EMPSI are secure against semi-honest adversaries and allow any number of colluding clients (at least one honest client). We also do experiments to compare EMPSI with some state-of-the-art works. The experimental results show that proposed EMPSI(-CA) has better performance and is scalable in the number of clients and the set size.

Community search is an important problem in network analysis, which has attracted much attention in recent years. As a query-oriented variant of community detection problem, community search starts with some given nodes, pays more attention to local network structures, and gets personalized resultant communities quickly. The existing community search method typically returns a single target community containing query nodes by default. This is a strict requirement and does not allow much flexibility. In many real-world applications, however, query nodes are expected to be located in multiple communities with different semantics. To address this limitation of existing methods, an efficient spectral-based Multi-Scale Community Search method (MSCS) is proposed, which can simultaneously identify the multi-scale target local communities to which query node belong. In MSCS, each node is equipped with a graph Fourier multiplier operator. The access of the graph Fourier multiplier operator helps nodes to obtain feature representations at various community scales. In addition, an efficient algorithm is proposed for avoiding the large number of matrix operations due to spectral methods. Comprehensive experimental evaluations on a variety of real-world datasets demonstrate the effectiveness and efficiency of the proposed method.

As a carrier of knowledge, papers have been a popular choice since ancient times for documenting everything from major historical events to breakthroughs in science and technology. With the booming development of science and technology, the number of papers has been growing exponentially. Just like the fact that Internet of Things (IoT) allows the world to be connected in a flatter way, how will the network formed by massive academic papers look like? Most existing visualization methods can only handle up to hundreds of thousands of node size, which is much smaller than that of academic networks which are usually composed of millions or even more nodes. In this paper, we are thus motivated to break this scale limit and design a new visualization method particularly for super-large-scale academic networks (VSAN). Nodes can represent papers or authors while the edges means the relation (e.g., citation, coauthorship) between them. In order to comprehensively improve the visualization effect, three levels of optimization are taken into account in the whole design of VSAN in a progressive manner, i.e., bearing scale, loading speed, and effect of layout details. Our main contributions are two folded: 1) We design an equivalent segmentation layout method that goes beyond the limit encountered by state-of-the-arts, thus ensuring the possibility of visually revealing the correlations of larger-scale academic entities. 2) We further propose a hierarchical slice loading approach that enables users to observe the visualized graphs of the academic network at both macroscopic and microscopic levels, with the ability to quickly zoom between different levels. In addition, we propose a “jumping between nebula graphs” method that connects the static pages of many academic graphs and helps users to form a more systematic and comprehensive understanding of various academic networks. Applying our methods to three academic paper citation datasets in the AceMap database confirms the visualization scalability of VSAN in the sense that it can visualize academic networks with more than 4 million nodes. The super-large-scale visualization not only allows a galaxy-like scholarly picture unfolding that were never discovered previously, but also returns some interesting observations that may drive extra attention from scientists.

Cloud vendors are actively adopting FPGAs into their infrastructures for enhancing performance and efficiency. As cloud services continue to evolve, FPGA (field programmable gate array) systems would play an even important role in the future. In this context, FPGA sharing in multi-tenancy scenarios is crucial for the wide adoption of FPGA in the cloud. Recently, many works have been done towards effective FPGA sharing at different layers of the cloud computing stack.

In this work, we provide a comprehensive survey of recent works on FPGA sharing. We examine prior art from different aspects and encapsulate relevant proposals on a few key topics. On the one hand, we discuss representative papers on FPGA resource sharing schemes; on the other hand, we also summarize important SW/HW techniques that support effective sharing. Importantly, we further analyze the system design cost behind FPGA sharing. Finally, based on our survey, we identify key opportunities and challenges of FPGA sharing in future cloud scenarios.

Agile development aims at rapidly developing software while embracing the continuous evolution of user requirements along the whole development process. User stories are the primary means of requirements collection and elicitation in the agile development. A project can involve a large amount of user stories, which should be clustered into different groups based on their functionality’s similarity for systematic requirements analysis, effective mapping to developed features, and efficient maintenance. Nevertheless, the current user story clustering is mainly conducted in a manual manner, which is time-consuming and subjective to human bias. In this paper, we propose a novel approach for clustering the user stories automatically on the basis of natural language processing. Specifically, the sentence patterns of each component in a user story are first analysed and determined such that the critical structure in the representative tasks can be automatically extracted based on the user story meta-model. The similarity of user stories is calculated, which can be used to generate the connected graph as the basis of automatic user story clustering. We evaluate the approach based on thirteen datasets, compared against ten baseline techniques. Experimental results show that our clustering approach has higher accuracy, recall rate and F1-score than these baselines. It is demonstrated that the proposed approach can significantly improve the efficacy of user story clustering and thus enhance the overall performance of agile development. The study also highlights promising research directions for more accurate requirements elicitation.

The deep learning methods based on syntactic dependency tree have achieved great success on Aspect-based Sentiment Analysis (ABSA). However, the accuracy of the dependency parser cannot be determined, which may keep aspect words away from its related opinion words in a dependency tree. Moreover, few models incorporate external affective knowledge for ABSA. Based on this, we propose a novel architecture to tackle the above two limitations, while fills up the gap in applying heterogeneous graphs convolution network to ABSA. Specially, we employ affective knowledge as an sentiment node to augment the representation of words. Then, linking sentiment node which have different attributes with word node through a specific edge to form a heterogeneous graph based on dependency tree. Finally, we design a multi-level semantic heterogeneous graph convolution network (Semantic-HGCN) to encode the heterogeneous graph for sentiment prediction. Extensive experiments are conducted on the datasets SemEval 2014 Task 4, SemEval 2015 task 12, SemEval 2016 task 5 and ACL 14 Twitter. The experimental results show that our method achieves the state-of-the-art performance.

The Internet of Everything (IoE) based cloud computing is one of the most prominent areas in the digital big data world. This approach allows efficient infrastructure to store and access big real-time data and smart IoE services from the cloud. The IoE-based cloud computing services are located at remote locations without the control of the data owner. The data owners mostly depend on the untrusted Cloud Service Provider (CSP) and do not know the implemented security capabilities. The lack of knowledge about security capabilities and control over data raises several security issues. Deoxyribonucleic Acid (DNA) computing is a biological concept that can improve the security of IoE big data. The IoE big data security scheme consists of the Station-to-Station Key Agreement Protocol (StS KAP) and Feistel cipher algorithms. This paper proposed a DNA-based cryptographic scheme and access control model (DNACDS) to solve IoE big data security and access issues. The experimental results illustrated that DNACDS performs better than other DNA-based security schemes. The theoretical security analysis of the DNACDS shows better resistance capabilities.

Model-based diagnosis (MBD) with multiple observations shows its significance in identifying fault location. The existing approaches for MBD with multiple observations use observations which is inconsistent with the prediction of the system. In this paper, we proposed a novel diagnosis approach, namely, the Diagnosis with Different Observations (DiagDO), to exploit the diagnosis when given a set of pseudo normal observations and a set of abnormal observations. Three ideas are proposed in this paper. First, for each pseudo normal observation, we propagate the value of system inputs and gain fanin-free edges to shrink the size of possible faulty components. Second, for each abnormal observation, we utilize filtered nodes to seek surely normal components. Finally, we encode all the surely normal components and parts of dominated components into hard clauses and compute diagnosis using the MaxSAT solver and MCS algorithm. Extensive tests on the ISCAS'85 and ITC'99 benchmarks show that our approach performs better than the state-of-the-art algorithms.

Commonsense question answering (CQA) requires understanding and reasoning over QA context and related commonsense knowledge, such as a structured Knowledge Graph (KG). Existing studies combine language models and graph neural networks to model inference. However, traditional knowledge graph are mostly concept-based, ignoring direct path evidence necessary for accurate reasoning. In this paper, we propose MRGNN (Meta-path Reasoning Graph Neural Network), a novel model that comprehensively captures sequential semantic information from concepts and paths. In MRGNN, meta-paths are introduced as direct inference evidence and an original graph neural network is adopted to aggregate features from both concepts and paths simultaneously. We conduct sufficient experiments on the CommonsenceQA and OpenBookQA datasets, showing the effectiveness of MRGNN. Also, we conduct further ablation experiments and explain the reasoning behavior through the case study.

BERT is a representative pre-trained language model that has drawn extensive attention for significant improvements in downstream Natural Language Processing (NLP) tasks. The complex architecture and massive parameters bring BERT competitive performance but also result in slow speed at model inference time. To speed up BERT inference, FastBERT realizes adaptive inference with an acceptable drop in accuracy based on knowledge distillation and the early-exit technique. However, many factors may limit the performance of FastBERT, such as the teacher classifier that is not knowledgeable enough, the batch size shrinkage and the redundant computation of student classifiers. To overcome these limitations, we propose a new BERT inference method with GPU-Efficient Exit Prediction (GEEP). GEEP leverages the shared exit loss to simplify the training process of FastBERT from two steps into only one step and makes the teacher classifier more knowledgeable by feeding diverse Transformer outputs to the teacher classifier. In addition, the exit layer prediction technique is proposed to utilize a GPU hash table to handle the token-level exit layer distribution and to sort test samples by predicted exit layers. In this way, GEEP can avoid batch size shrinkage and redundant computation of student classifiers. Experimental results on twelve public English and Chinese NLP datasets prove the effectiveness of the proposed approach. The source codes of GEEP will be released to the public upon paper acceptance.

Session-based recommendation (SBR) and multi-behavior recommendation (MBR) are both important problems and have attracted the attention of many researchers and practitioners. Different from SBR that solely uses one single type of behavior sequences and MBR that neglects sequential dynamics, heterogeneous SBR (HSBR) that exploits different types of behavioral information (e.g., examinations like clicks or browses, purchases, adds-to-carts and adds-to-favorites) in sequences is more consistent with real-world recommendation scenarios, but it is rarely studied. Early efforts towards HSBR focus on distinguishing different types of behaviors or exploiting homogeneous behavior transitions in a sequence with the same type of behaviors. However, all the existing solutions for HSBR do not exploit the rich heterogeneous behavior transitions in an explicit way and thus may fail to capture the semantic relations between different types of behaviors. However, all the existing solutions for HSBR do not model the rich heterogeneous behavior transitions in the form of graphs and thus may fail to capture the semantic relations between different types of behaviors. The limitation hinders the development of HSBR and results in unsatisfactory performance. As a response, we propose a novel behavior-aware graph neural network (BGNN) for HSBR. Our BGNN adopts a dual-channel learning strategy for differentiated modeling of two different types of behavior sequences in a session. Moreover, our BGNN integrates the information of both homogeneous behavior transitions and heterogeneous behavior transitions in a unified way. We then conduct extensive empirical studies on three real-world datasets, and find that our BGNN outperforms the best baseline by 21.87%, 18.49%, and 37.16% on average correspondingly. A series of further experiments and visualization studies demonstrate the rationality and effectiveness of our BGNN. An exploratory study on extending our BGNN to handle more than two types of behaviors show that our BGNN can easily and effectively be extended to multi-behavior scenarios.

Reachability query plays a vital role in many graph analysis tasks. Previous researches proposed many methods to efficiently answer reachability queries between vertex pairs. Since many real graphs are labeled graph, it highly demands Label-Constrained Reachability (LCR) query in which constraint includes a set of labels besides vertex pairs. Recent researches proposed several methods for answering some LCR queries which require appearance of some labels specified in constraints in the path. Besides that constraint may be a label set, query constraint may be ordered labels, namely OLCR (Ordered-Label-Constrained Reachability) queries which retrieve paths matching a sequence of labels. Currently, no solutions are available for OLCR. Here, we propose DHL, a novel bloom filter based indexing technique for answering OLCR queries. DHL can be used to check reachability between vertex pairs. If the answers are not no, then constrained DFS is performed. So, we employ DHL followed by performing constrained DFS to answer OLCR queries. We show that DHL has a bounded false positive rate, and it’s powerful in saving indexing time and space. Extensive experiments on 10 real-life graphs and 12 synthetic graphs demonstrate that DHL achieves about 4.8–22.5 times smaller index space and 4.6–114 times less index construction time than two state-of-art techniques for LCR queries, while achieving comparable query response time. The results also show that our algorithm can answer OLCR queries effectively.

When one enterprise acquires another, the electronic data of the acquired enterprise will be transferred to the acquiring enterprise. In particular, if the data system of acquired enterprise contains a searchable encryption mechanism, the corresponding searchability will also be transferred. In this paper, we introduce the concept of Searchable Encryption with Ownership Transfer (SEOT), and propose a secure SEOT scheme. Based on the new structure of polling pool, our proposed searchable encryption scheme not only achieves efficient transfer of outsourced data, but also implements secure transfer of data searchability. Moreover, we optimize the storage cost for user to a desirable value. We prove our scheme can achieve the secure characteristics, then carry out the performance evaluation and experiments. The results demonstrate that our scheme is superior in efficiency and practicability.

With the benefits of reducing time and workforce, automated testing has been widely used for the quality assurance of mobile applications (APPs). Compared with automated testing, manual testing can achieve higher coverage in complex interactive Activities. And the effectiveness of manual testing is highly dependent on the user operation process (UOP) of experienced testers. Based on the UOP, we propose an iterative Android automated testing (IAAT) method that automatically records, extracts, and integrates UOPs to guide the test logic of the tool across the complex Activity iteratively. The feedback test results can train the UOPs to achieve higher coverage in each iteration. We extracted 50 UOPs and conducted experiments on 10 popular mobile APPs to demonstrate IAAT’s effectiveness compared with Monkey and the initial automated tests. The experimental results show a noticeable improvement in the IAAT compared with the test logic without human knowledge. Under the 60 minutes test time, the average code coverage is improved by 13.98% to 37.83%, higher than the 27.48% of Monkey under the same conditions.

It is a significant and challenging task to detect the informative features to carry out explainable analysis for high dimensional data, especially for those with very small number of samples. Feature selection especially the unsupervised ones are the right way to deal with this challenge and realize the task. Therefore, two unsupervised spectral feature selection algorithms are proposed in this paper. They group features using advanced Self-Tuning spectral clustering algorithm based on local standard deviation, so as to detect the global optimal feature clusters as far as possible. Then two feature ranking techniques, including cosine-similarity-based feature ranking and entropy-based feature ranking, are proposed, so that the representative feature of each cluster can be detected to comprise the feature subset on which the explainable classification system will be built. The effectiveness of the proposed algorithms is tested on high dimensional benchmark omics datasets and compared to peer methods, and the statistical test are conducted to determine whether or not the proposed spectral feature selection algorithms are significantly different from those of the peer methods. The extensive experiments demonstrate the proposed unsupervised spectral feature selection algorithms outperform the peer ones in comparison, especially the one based on cosine similarity feature ranking technique. The statistical test results show that the entropy feature ranking based spectral feature selection algorithm performs best. The detected features demonstrate strong discriminative capabilities in downstream classifiers for omics data, such that the AI system built on them would be reliable and explainable. It is especially significant in building transparent and trustworthy medical diagnostic systems from an interpretable AI perspective.

Wavelet transform is being widely used in the field of information processing. One-dimension and two-dimension quantum wavelet transforms have been investigated as important tool algorithms. However, three-dimensional quantum wavelet transforms have not been reported. This paper proposes a multi-level three-dimensional quantum wavelet transform theory to implement the wavelet transform for quantum videos. Then, we construct the iterative formulas for the multi-level three-dimensional Haar and Daubechies D4 quantum wavelet transforms, respectively. Next, we design quantum circuits of the two wavelet transforms using iterative methods. Complexity analysis shows that the proposed wavelet transforms offer exponential speed-up over their classical counterparts. Finally, the proposed quantum wavelet transforms are selected to realize quantum video compression as a primary application. Simulation results reveal that the proposed wavelet transforms have better compression performance for quantum videos than two-dimension quantum wavelet transforms.

Identity-based threshold signature (IDTS) is a forceful primitive to protect identity and data privacy, in which parties can collaboratively sign a given message as a signer without reconstructing a signing key. Nevertheless, most IDTS schemes rely on a trusted key generation center (KGC). Recently, some IDTS schemes can achieve escrow-free security against corrupted KGC, but all of them are vulnerable to denial-of-service attacks in the dishonest majority setting, where cheaters may force the protocol to abort without providing any feedback. In this work, we present a fully decentralized IDTS scheme to resist corrupted KGC and denial-of-service attacks. To this end, we design threshold protocols to achieve distributed key generation, private key extraction, and signing generation which can withstand the collusion between KGCs and signers, and then we propose an identification mechanism that can detect the identity of cheaters during key generation, private key extraction and signing generation. Finally, we formally prove that the proposed scheme is threshold unforgeability against chosen message attacks. The experimental results show that the computation time of both key generation and signing generation is <1 s, and private key extraction is about 3 s, which is practical in the distributed environment.

Uncertain Knowledge Graphs (UKGs) are used to characterize the inherent uncertainty of knowledge and have a richer semantic structure than deterministic knowledge graphs. The research on the embedding of UKG has only recently begun, Uncertain Knowledge Graph Embedding (UKGE) model has a certain effect on solving this problem. However, there are still unresolved issues. On the one hand, when reasoning the confidence of unseen relation facts, the introduced probabilistic soft logic cannot be used to combine multi-path and multi-step global information, leading to information loss. On the other hand, the existing UKG embedding model can only model symmetric relation facts, but the embedding problem of asymmetric relation facts has not be addressed. To address the above issues, a Multiplex Uncertain Knowledge Graph Embedding (MUKGE) model is proposed in this paper. First, to combine multiple information and achieve more accurate results in confidence reasoning, the Uncertain ResourceRank (URR) reasoning algorithm is introduced. Second, the asymmetry in the UKG is defined. To embed asymmetric relation facts of UKG, a multi-relation embedding model is proposed. Finally, experiments are carried out on different datasets via 4 tasks to verify the effectiveness of MUKGE. The results of experiments demonstrate that MUKGE can obtain better overall performance than the baselines, and it helps advance the research on UKG embedding.

The flourish of deep learning frameworks and hardware platforms has been demanding an efficient compiler that can shield the diversity in both software and hardware in order to provide application portability. Among the existing deep learning compilers, TVM is well known for its efficiency in code generation and optimization across diverse hardware devices. In the meanwhile, the Sunway many-core processor renders itself as a competitive candidate for its attractive computational power in both scientific computing and deep learning workloads. This paper combines the trends in these two directions. Specifically, we propose swTVM that extends the original TVM to support ahead-of-time compilation for architecture requiring cross-compilation such as Sunway. In addition, we leverage the architecture features during the compilation such as core group for massive parallelism, DMA for high bandwidth memory transfer and local device memory for data locality, in order to generate efficient codes for deep learning workloads on Sunway. The experiment results show that the codes generated by swTVM achieve 1.79\times improvement of inference latency on average compared to the state-of-the-art deep learning framework on Sunway, across eight representative benchmarks. This work is the first attempt from the compiler perspective to bridge the gap of deep learning and Sunway processor particularly with productivity and efficiency in mind. We believe this work will encourage more people to embrace the power of deep learning and Sunway many-core processor.

Transformer architectures have facilitated the development of large-scale and general-purpose sequence models for prediction tasks in natural language processing and computer vision, e.g., GPT-3 and Swin Transformer. Although originally designed for prediction problems, it is natural to inquire about their suitability for sequential decision-making and reinforcement learning problems, which are typically beset by long-standing issues involving sample efficiency, credit assignment, and partial observability. In recent years, sequence models, especially the Transformer, have attracted increasing interest in the RL communities, spawning numerous approaches with notable effectiveness and generalizability. This survey presents a comprehensive overview of recent works aimed at solving sequential decision-making tasks with sequence models such as the Transformer, by discussing the connection between sequential decision-making and sequence modeling, and categorizing them based on the way they utilize the Transformer. Moreover, this paper puts forth various potential avenues for future research intending to improve the effectiveness of large sequence models for sequential decision-making, encompassing theoretical foundations, network architectures, algorithms, and efficient training systems.

In this paper, we introduce a sub-Nyquist sampling-based receiver architecture and method for wideband spectrum sensing. Instead of recovering the original wideband analog signal, the proposed method aims to directly reconstruct the power spectrum of the wideband analog signal from sub-Nyquist samples. Note that power spectrum alone is sufficient for wideband spectrum sensing. Since only the covariance matrix of the wideband signal is needed, the proposed method, unlike compressed sensing-based methods, does not need to impose any sparsity requirement on the frequency domain. The proposed method is based on a multi-coset sampling architecture. By exploiting the inherent sampling structure, a fast compressed power spectrum estimation method whose primary computational task consists of fast Fourier transform (FFT) is proposed. Simulation results are presented to show the effectiveness of the proposed method.

Recently, segmentation-based scene text detection has drawn a wide research interest due to its flexibility in describing scene text instance of arbitrary shapes such as curved texts. However, existing methods usually need complex post-processing stages to process ambiguous labels, i.e., the labels of the pixels near the text boundary, which may belong to the text or background. In this paper, we present a framework for segmentation-based scene text detection by learning from ambiguous labels. We use the label distribution learning method to process the label ambiguity of text annotation, which achieves a good performance without using additional post-processing stage. Experiments on benchmark datasets demonstrate that our method produces better results than state-of-the-art methods for segmentation-based scene text detection.

Multi-task learning is to improve the performance of the model by transferring and exploiting common knowledge among tasks. Existing MTL works mainly focus on the scenario where label sets among multiple tasks (MTs) are usually the same, thus they can be utilized for learning across the tasks. However, the real world has more general scenarios in which each task has only a small number of training samples and their label sets are just partially overlapped or even not. Learning such MTs is more challenging because of less correlation information available among these tasks. For this, we propose a framework to learn these tasks by jointly leveraging both abundant information from a learnt auxiliary big task with sufficiently many classes to cover those of all these tasks and the information shared among those partially-overlapped tasks. In our implementation of using the same neural network architecture of the learnt auxiliary task to learn individual tasks, the key idea is to utilize available label information to adaptively prune the hidden layer neurons of the auxiliary network to construct corresponding network for each task, while accompanying a joint learning across individual tasks. Extensive experimental results demonstrate that our proposed method is significantly competitive compared to state-of-the-art methods.

Crowdsourcing provides an effective and low-cost way to collect labels from crowd workers. Due to the lack of professional knowledge, the quality of crowdsourced labels is relatively low. A common approach to addressing this issue is to collect multiple labels for each instance from different crowd workers and then a label integration method is used to infer its true label. However, to our knowledge, almost all existing label integration methods merely make use of the original attribute information and do not pay attention to the quality of the multiple noisy label set of each instance. To solve these issues, this paper proposes a novel three-stage label integration method called attribute augmentation-based label integration (AALI). In the first stage, we design an attribute augmentation method to enrich the original attribute space. In the second stage, we develop a filter to single out reliable instances with high-quality multiple noisy label sets. In the third stage, we use majority voting to initialize integrated labels of reliable instances and then use cross-validation to build multiple component classifiers on reliable instances to predict all instances. Experimental results on simulated and real-world crowdsourced datasets demonstrate that AALI outperforms all the other state-of-the-art competitors.