How can we efficiently index extensive high-dimensional vector data increasing over time, enabling quick and accurate proximity searches within designated time windows? A time-restricted \( k \) -Nearest Neighbor (T \( k \) NN) query aims to identify the \( k \) -nearest vectors to a query vector within a specified time window. While high-dimensional and time-accumulating data are ubiquitous and managing such data efficiently is becoming increasingly significant, T \( k \) NN search within this context has not received much attention so far. In this article, we propose Multi-Level Block Indexing (MBI), a tailored indexing method for efficient approximate T \( k \) NN search. MBI employs an incremental hierarchical index structure that divides the data into multiple blocks based on timestamps. This structure ensures efficient query processing, irrespective of the length of the query time window, and facilitates the addition of new data over time. Furthermore, we extend T \( k \) NN beyond timestamps to the \( m \) -A \( k \) NN problem, incorporating \( m \) additional attributes—such as age, height, and weight in medical data or citation counts in academic papers—allowing queries to integrate multiple numerical constraints. Experimental results highlight MBI’s superiority over conventional methods, achieving query processing speeds up to 10.88 times faster and offering logarithmic scaling in data insertion time as the data volume grows. Additionally, in \( m \) -A \( k \) NN queries, MBI maintains stable indexing performance and achieves up to 1.99 times faster query performance, demonstrating its effectiveness in large-scale, multi-attribute \( k \) NN search.

How can we efficiently find connected components with Union-Find in a distributed system? Union-Find is the most efficient sequential algorithm for finding connected components with low memory usage and high speed. Several studies have adapted Union-Find to distributed memory systems to process large graphs quickly; however, they all suffer from load balancing problems. We notice that the leading cause of the load balancing problems is the nature of Union-Find, which gathers more and more edges to a small number of vertices as it proceeds. In this paper, we propose BTS, a new fast and scalable distributed Union-Find algorithm for finding connected components in large graphs. BTS resolves the load balancing problems by proposing Balanced Union-Find, which allocates vertices to each processor and makes edges link to vertices in the same processor as much as possible. We further optimize BTS with edge refinement to minimize network traffic and memory usage. Experimental results show that BTS efficiently resolves the load balancing problems, processing 16–1024 times larger graphs with 3.1-261.9 times faster speeds than existing algorithms.

Edge TPU (Tensor Processing Unit) is being widely utilized in various edge computing applications as a high-efficiency, low-power accelerator for deep learning computations. However, temperature rise in Edge TPU can lead to performance degradation, reduced stability, and shortened lifespan, necessitating temperature management through thermal modeling. This paper proposes a task-level thermal modeling technique for predicting Edge TPU temperature. The proposed method estimates power consumption of CPU and Edge TPU based on workloads of various deep learning tasks and predicts the convergence temperature of Edge TPU using a steady temperature model. Through experiments, we confirmed that the proposed method accurately predicts Edge TPU temperature for various workloads. The average prediction error was 0.7°C. This study is expected to serve as a foundation for developing temperature management techniques by presenting an effective temperature prediction model that considers the thermal characteristics of Edge TPU.

Deep neural networks (DNNs) have been deployed in many safety-critical real-time embedded systems. To support DNN tasks in real-time, most previous studies focused on GPU or CPU. However, Edge TPU has not yet been studied for real-time guarantees. This paper presents a real-time DNNs framework for Edge TPU to satisfy multiple DNN inference tasks’ timing requirements. The proposed framework provides 1) SRAM allocation and model partitioning techniques and 2) a MIP-based algorithm that determines the amount of SRAM and the number of segments for each task. The experiment result shows that our framework provides 79% higher schedulability than the existing Edge TPU system.