Search results
Qiyan Li's 6 research works with 48 citations and 315 reads, including: Learned sketch for subgraph counting: a holistic approach.
About. •Top final year Electrical Engineering student with engineering project management experience. •Strong communication skills and interpersonal skills. •Knowledgeable in Microsoft...
- 422
- Hatch
Apr 25, 2024 · Guofeng Qin, Xiaoguo Wang, Li Wang, Yang Li, Qiyan Li: Remote Video Monitor of Vehicles in Cooperative Information Platform. CDVE 2009: 208-215
Subgraph counting, as a fundamental problem in network analysis, is to count the number of subgraphs in a data graph that match a given query graph by either homomorphism or subgraph isomorphism.
- Abstract
- 2 Problem statement
- (u , v) and (v, u ).
- (iii) L ((u , v)) = L (( f (u ), f
- q-error (q ) ˆ ) ) = max c (q , c ˆ ) (q )
- 4 The learned sketch
- v e
- 5 Active learning for LSS
- 6 Model transfer
- 9 Conclusion
Subgraph counting, as a fundamental problem in network analysis, is to count the number of subgraphs in a data graph that match a given query graph by either homomorphism or subgraph isomorphism. The importance of subgraph counting derives from the fact that it provides insights of a large graph, in particular a labeled graph, when a collection of ...
We model a graph as a labeled directed graph G = (V , E , L , Σ). Here, V is a set of nodes, E is a set of directed edges, Σ is a set of labels, and L is the mapping function that maps a node u V or an edge e ∈ E to a label denoted ∈ as L (u or ) L (e ). Each directed edge e (u is a link = , v) from source node u to target node v. We denote the out...
Given a data graph G (V E L and a query graph = , , , Σ) q = (Vq, Eq,L , Σ). A homomorphism of q to G is a func-tion f : Vq V such that (i) for every u → ∈ Vq, L(u ) = L ( f (u )), (ii) for every (u , v) Eq, f (u f E, ∈ ( ), (v)) ∈ and
(v))). A subgraph isomorphism of q to G is a homomorphism of q to G under the con-dition that f is an injective function, where f (u f ) = (v) for any pair of u and in Vq if u v. A homomorphism v = (or subgraph isomorphism) function f of q induces a sub-graph G f (Vf E f L in G, where Vf is = , , , Σ) the set of nodes, f (u ), for every u in Vq, an...
data, Q ∪ Q . This procedure repeats during online testing Δ and can be fully automated without human-in-the-loop. The benefit of LSS actively learned is two folds. First, the increas-ing volume of training data improves the performance of LSS. Second, the added new training data motivates LSS to adapt to a varying workload.
In this section, we discuss the learned sketch for subgraph counting (LSS) on modeling, architecture, and node encod-ing.
l∈L (l ), where e (l is the pre-trained embedding of ) (v) the label l in GL (u in q, we set if has the label
We first introduce active learning (AL) in brief. A regular ML task is to train a model Θ by minimizing the empirical loss measured by a loss function on the training data XL, and L produce the predictive value for the test data XU, where XL and XU are assumed to be randomly sampled from an under-lying distribution. The idea of AL is to enhance the...
In this section, we explore the possibility of LSS model trans-fer, i.e., reusing the learned LSS model built from a source training task with the workload Qs to a target task with the workload Qt. It is worth noting that the procedure of model transfer is different from active learning. By active learning, an LSS model is improved actively for a s...
In this paper, we propose a general neural network frame-work as a learned sketch for subgraph counting over a large data graph, which can be used for either homomorphism or subgraph isomorphism counting. The data graph and query graph can be node and/or edge labels and directed/undirected graphs. In addition, an active learning strategy is devised...
Li Qiyan ( Chinese: 李其炎; pinyin: Lǐ Qíyán) (October 1938 – 3 June 2020) served as mayor of Beijing from February 1993 to November 1996. He was originally from Qihe County in Shandong province. Li joined the Chinese Communist Party in November 1961.
Guanghua Li, Qiyan Li, Jingqiao Liu, Yuanyuan Zhu, Ming Zhong: FANE: A Fusion-Based Attributed Network Embedding Framework. APWeb/WAIM (1) 2021: 53-60
