Modeling occupant behavior in buildings

Set intersection is a fundamental operation in information retrieval and database systems. This paper introduces linear space data structures to represent sets such that their intersection can be computed in a worst-case efficient way. In general, given k (preprocessed) sets, with totally n elements, we will show how to compute their intersection in expected time O(n/sqrt(w)+kr), where r is the intersection size and w is the number of bits in a machine-word. In addition,we introduce a very simple version of this algorithm that has weaker asymptotic guarantees but performs even better in practice; both algorithms outperform the state of the art techniques in terms of execution time for both synthetic and real data sets and workloads.

Fast Set Intersection in Memory

Recently, recommender systems play a pivotal role in alleviating the problem of information overload. Latent factor models have been widely used for recommendation. Most existing latent factor models mainly utilize the interaction information between users and items, although some recently extended models utilize some auxiliary information to learn a unified latent factor for users and items. The unified latent factor only represents the characteristics of users and the properties of items from the aspect of purchase history. However, the characteristics of users and the properties of items may stem from different aspects, e.g., the brand-aspect and category-aspect of items. Moreover, the latent factor models usually use the shallow projection, which cannot capture the characteristics of users and items well. Deep neural network has shown tremendous potential to model the non-linearity relationship between users and items. It can be used to replace shallow projection to model the complex correlation between users and items. In this paper, we propose a Neural network based Aspect-level Collaborative Filtering model (NeuACF) to exploit different aspect latent factors. Through modelling the rich object properties and relations in recommender system as a heterogeneous information network, NeuACF first extracts different aspect-level similarity matrices of users and items, respectively, through different meta-paths, and then feeds an elaborately designed deep neural network with these matrices to learn aspect-level latent factors. Finally, the aspect-level latent factors are fused for the top-N recommendation. Moreover, to fuse information from different aspects more effectively, we further propose NeuACF++ to fuse aspect-level latent factors with self-attention mechanism. Extensive experiments on three real world datasets show that NeuACF and NeuACF++ significantly outperform both existing latent factor models and recent neural network models.

Deep Collaborative Filtering with Multi-Aspect Information in Heterogeneous Networks

Smart home is an emerging area that opens up a diverse set of downstream applications, such as dynamic pricing and demand response techniques, whose goals are typically to lower power consumption while providing comfortable and convenient services. Smart home has been extensively studied, and shown to be beneficial in people's real lives. Although useful, typical smart home platforms work at the "servant" level -- highly dependent on user inputs with no predictions of human demands -- which, if well addressed, would significantly widen their applicability. In this article, we propose a human-centric smart home energy management system (SHE) that works at the "butler" level. The system integrates ubiquitous sensing data from the physical and cyber spaces to discover the patterns of power usage and cognitively understand the behaviors of human beings. The relationship between them is established to dynamically infer users' demands for electricity, and then the optimal scheduling of the home energy system is triggered to respond to both the users' demands and electricity rates. Based on the novel framework, our SHE system provides intelligent services to satisfy the requirements of users as a butler -- aiming not only to save the electricity cost or reduce the peak load, but also to predict users' demands and managing "servants."

Butler, Not Servant: A Human-Centric Smart Home Energy Management System

We study the performance of eight representative in-memory subgraph matching algorithms. Specifically, we put QuickSI, GraphQL, CFL, CECI, DP-iso, RI and VF2++ in a common framework to compare them on the following four aspects: (1) method of filtering candidate vertices in the data graph; (2) method of ordering query vertices; (3) method of enumerating partial results; and (4) other optimization techniques. Then, we compare the overall performance of these algorithms with Glasgow, an algorithm based on the constraint programming. Through experiments, we find that (1) the filtering method of GraphQL is competitive to that of the latest algorithms CFL, CECI and DP-iso in terms of pruning power; (2) the ordering methods in GraphQL and RI are usually the most effective; (3) the set intersection based local candidate computation in CECI and DP-iso performs the best in the enumeration; and (4) the failing sets pruning in DP-iso can significantly improve the performance when queries become large. Our source code is publicly available at https://github.com/RapidsAtHKUST/SubgraphMatching.

In-Memory Subgraph Matching: An In-depth Study

A subgraph query searches for all embeddings in a data graph that are identical to a query graph. Two kinds of algorithms, either graph exploration based or join based, have been developed for processing subgraph queries. Due to algorithmic and implementational differences, join-based systems can handle query graphs of a few vertices efficiently whereas exploration-based approaches typically process up to several tens of vertices in the query graph. In this paper, we first compare these two kinds of methods and prove that the complexity of result enumeration in state-of-the-art exploration-based methods matches that of the worst-case optimal join. Furthermore, we propose RapidMatch, a holistic subgraph query processing framework integrating the two approaches. Specifically, RapidMatch not only runs relational operators such as selections and joins, but also utilizes graph structural information, as in graph exploration, for filtering and join plan generation. Consequently, it outperforms the state of the art in both approaches on a wide range of query workloads. PVLDB Reference Format: Shixuan Sun, Xibo Sun, Yulin Che, Qiong Luo, and Bingsheng He. RapidMatch: A Holistic Approach to Subgraph Query Processing. PVLDB, 14(2): 176-188, 2021. doi:10.14778/3425879.3425888 PVLDB Availability Tag: The source code of this research paper has been made publicly available at https://github.com/RapidsAtHKUST/RapidMatch.

/pdf/rapidmatch-a-holistic-approach-to-subgraph-query-processing-2d70wqij3i.pdf

RapidMatch: a holistic approach to subgraph query processing

Truss decomposition is to divide a graph into a hierarchy of subgraphs, or trusses. A subgraph is a k-truss (k 2) if each edge is in at least k --- 2 triangles in the subgraph. Existing algorithms ...

/pdf/accelerating-truss-decomposition-on-heterogeneous-processors-41blbecj5f.pdf

Accelerating Truss Decomposition on Heterogeneous Processors

Subgraph enumeration finds all subgraphs in an unlabeled graph that are isomorphic to another unlabeled graph. Existing depth-first search (DFS) based algorithms work on a single machine, but they are slow on large graphs due to the large search space. In contrast, distributed algorithms on clusters adopt a parallel breadth-first search (BFS) and improve the performance at the cost of large amounts of hardware resources, since the BFS approach incurs expensive data transfer and space cost due to the exponential number of intermediate results. In this paper, we develop an efficient parallel subgraph enumeration algorithm for a single machine, named LIGHT. Our algorithm reduces redundant computation in DFS by deferring the materialization of pattern vertices until necessary and converting the candidate set computation into finding a minimum set cover. Moreover, we parallelize our algorithm with both SIMD (Single-Instruction-Multiple-Data) instructions and SMT (Simultaneous Multi-Threading) technologies in modern CPUs. Our experimental results show that LIGHT running on a single machine outperforms existing single-machine DFS algorithms by more than three orders of magnitude, and is up to two orders of magnitude faster than the state-of-the-art distributed algorithms running on 12 machines. Additionally, LIGHT completed all test cases, whereas the existing algorithms fail in some cases due to either running out of time or running out of available hardware resources.

Efficient Parallel Subgraph Enumeration on a Single Machine

The invention discloses an intelligent household energy management method based on intelligent wearable equipment behavior perception. The method comprises steps of S101, collecting sensor data and external environmental data of wearable equipment; S102, preprocessing data; S103, determining the sensor type; S104, perceiving motion behaviors of a user; S105, perceiving body states of the user; S106, estimating living environment requirements of the user; S107, judging user requirement update; S108, optimizing control strategies of electrical equipment in an event-driven manner; and S109, updating control strategies of an intelligent household control module. According to the invention, requirements for current and future living environments of the user can be estimated and updated by monitoring various sensors on wearable equipment of the user, so that scheduling strategies of various intelligent household electrical appliances are made, household electrical appliances can be actively controlled, requirements for the comfort level of the household environment of the user can be met, and the power utilization cost is reduced.

Intelligent household energy management method based on intelligent wearable equipment behavior perception

Measuring similarities among different vertices is a fundamental problem in graph analysis. Among different similarity measurements, SimRank is one of the most promising and popular. In reality, instead of computing the whole similarity matrix, people often issue SimRank queries in a pairwise manner, each of which needs to estimate an approximate SimRank value within a specified accuracy for a given pair of nodes. These pairwise SimRank queries are often processed on real-life graphs, which typically evolve over time, requiring efficient algorithms that can query pairwise SimRank under dynamic graph updates. However, current single-pair SimRank solutions are either static or inefficient in handling dynamic cases with good-quality results. Observing that the sample size is the major factor that determines the efficiency and the accuracy in Monte Carlo methods to estimate pairwise SimRank, in this paper, we propose three algorithms to query pairwise SimRank over static and dynamic graphs efficiently, by using different sample reduction strategies. The accuracy of our algorithms is guaranteed by the different invariants we propose for pairwise SimRank. We show that our algorithms outperform the state-of-the-art static and dynamic solutions for pairwise SimRank estimation.

https://link.springer.com/content/pdf/10.1007%2Fs00778-018-0521-x.pdf

Yulin Che

Papers

RapidMatch: a holistic approach to subgraph query processing

Accelerating Truss Decomposition on Heterogeneous Processors

Efficient Parallel Subgraph Enumeration on a Single Machine

Intelligent household energy management method based on intelligent wearable equipment behavior perception

Accelerating pairwise SimRank estimation over static and dynamic graphs