A5103063161- Parallel Computing and Optimization Techniques
- Embedded Systems Design Techniques
- Advanced Neural Network Applications
- Advanced Memory and Neural Computing
- Access Control and Trust
- Advanced Data Storage Technologies
- Cryptography and Data Security
- Interconnection Networks and Systems
- VLSI and Analog Circuit Testing
- Data Management and Algorithms
- Ferroelectric and Negative Capacitance Devices
- Cloud Computing and Resource Management
- Advanced Image and Video Retrieval Techniques
- CCD and CMOS Imaging Sensors
- Adversarial Robustness in Machine Learning
- Arsenic contamination and mitigation
- Image Retrieval and Classification Techniques
- Multi-Agent Systems and Negotiation
- Augmented Reality Applications
- Heavy metals in environment
- VLSI and FPGA Design Techniques
- Heavy Metal Exposure and Toxicity
- Semiconductor materials and devices
- Internet Traffic Analysis and Secure E-voting
- Distributed systems and fault tolerance
Chinese Academy of Sciences
2014-2025
Institute of Computing Technology
2016-2025
Xinjiang Technical Institute of Physics & Chemistry
2015-2024
Xinjiang University
2024
First Affiliated Hospital of Zhengzhou University
2024
Tsinghua University
2024
Brigham Young University
2023
Changchun Institute of Technology
2021
Minzu University of China
2021
Dalian University of Technology
2020
Neural networks (NNs) have been demonstrated to be useful in a broad range of applications such as image recognition, automatic translation and advertisement recommendation. State-of-the-art NNs are known both computationally memory intensive, due the ever-increasing deep structure, i.e., multiple layers with massive neurons connections (i.e., synapses). Sparse neural emerged an effective solution reduce amount computation required. Though existing NN accelerators able efficiently process...
Devising a complete and correct set of roles has been recognized as one the most important challenging tasks in implementing role based access control. A key problem related to this is notion goodness/interestingness -- when good/interesting? In paper, we define mining (RMP) discovering an optimal from existing user permissions. The main contribution paper formally RMP, analyze its theoretical bounds. addition above basic introduce two different variations called δ-approx RMP Minimal Noise...
Devising a complete and correct set of roles has been recognized as one the most important challenging tasks in implementing role based access control. A key problem related to this is notion goodness - when good? Recently, mining (RMP) defined discovering an optimal from existing user permissions. Several different objectives for optimality have proposed. However, with these definitions that often organizations already deployed wish optimize set. Even if discovered, widely different, it...
Multiple myeloma (MM) is the second most common haematological malignancy in UK. We present a case series of three patients with light chain only who had normal serum protein electrophoretic studies at screening and were diagnosed using urine free assessment. This reiterates importance thorough robust for MM presenting renal disease. review up to date literature we highlight need screen combination electrophoresis/immunofixation either urinary or measurement maintain high index suspicion...
Devising a complete and correct set of roles has been recognized as one the most important challenging tasks in implementing role-based access control. A key problem related to this is notion goodness/interestingness—when role good/interesting? In article, we define Role Mining Problem (RMP) discovering an optimal from existing user permissions. The main contribution article formally RMP analyze its theoretical bounds. addition above basic RMP, introduce two different variations called...
Due to the "curse of dimensionality" problem, it is very expensive process nearest neighbor (NN) query in high-dimensional spaces; and hence, approximate approaches, such as Locality-Sensitive Hashing (LSH), are widely used for their theoretical guarantees empirical performance. Current LSH-based approaches target at L1 L2 spaces, while shown previous work, fractional distance metrics (Lp with 0 < p 1) can provide more insightful results than usual data mining multimedia applications....
Role hierarchies are fundamental to the role based access control (RBAC) model. The notion of hierarchy is a well understood concept that allows senior roles inherit permissions corresponding junior roles. further ease burden security administration, as there no need explicitly specify and maintain large number permissions. Given set or user permissions, one may construct alternative hierarchies. However, does not exist an optimal hierarchy. Optimality helps in maximizing benefit employing...
Today, role-based access control (RBAC) has become a well-accepted paradigm for implementing because of its convenience and ease administration. However, in order to realize the full benefits RBAC paradigm, one must first define roles accurately. This task defining associating permissions with them, also known as role engineering, is typically accomplished either top-down or bottom-up manner. Under approach, careful analysis business processes done job functions then specify appropriate from...
Deep Learning Accelerators (DLAs) are effective to improve both performance and energy efficiency of compute-intensive deep learning algorithms. A flexible portable mean exploit DLAs is using high-performance software libraries with well-established APIs, which typically either manually implemented or automatically generated by exploration-based compilation approaches. Though approaches significantly reduce programming efforts, they fail find optimal near-optimal programs from a large but...
Graph neural networks (GNNs), which extend traditional for processing graph-structured data, have been widely used in many fields. The GNN computation mainly consists of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">edge processing</i> to generate messages by combining edge/vertex features and xmlns:xlink="http://www.w3.org/1999/xlink">vertex update vertex with aggregated messages. In addition nontrivial vector operations edge...
Sampling-based methods, such as SimPoint, are widely used for efficient pre-silicon \mu Arch evaluations, where the costs number of simulation points multiplied by evaluated designs. However, these keep growing with an increasing and expanding design space. Although techniques have been developed to accelerate space exploration, less attention has given further reducing budget each evaluation. Common strategies like coverage or sampling fewer typically compromise estimation accuracy....
Because of its ease administration, role-based access control (RBAC) has become the norm to enforcing security in most today's organizations. For implementing RBAC, it is important devise a complete and correct set roles. This task, known
During the design of a microprocessor, Design Space Exploration (DSE) is critical step which determines appropriate configuration microprocessor. In computer architecture community, supervised learning techniques have been applied to DSE build models for predicting qualities configurations. For learning, however, considerable simulation costs are required attaining labeled Given limited resources, it difficult achieve high accuracy. this paper, inspired by recent advances in semi-supervised...
Ever-increasing design complexity and advances of technology impose great challenges on the modern microprocessors. One such challenge is to determine promising microprocessor configurations meet specific constraints, which called Design Space Exploration (DSE). In computer architecture community, supervised learning techniques have been applied DSE build regression models for predicting qualities configurations. For learning, however, considerable simulation costs are required attaining...
As a fundamental task in computer architecture research, performance comparison has been continuously hampered by the variability of performance. In traditional comparisons, impact is usually ignored (i.e., means observations are compared regardless variability), or few cases directly addressed with <inline-formula><tex-math>$t$</tex-math></inline-formula> -statistics without checking number and normality observations. this paper, we formulate as statistical task, empirically illustrate why...
In this talk, we present the overall system design and architecture, challenges encountered in practice, lessons learned from production deployment of talent search recommendation systems at LinkedIn. By presenting our experiences applying techniques intersection recommender systems, information retrieval, machine learning, statistical modeling a large-scale industrial setting highlighting open problems, hope to stimulate further research collaborations within SIGIR community.