Ising-Based Consensus Clustering on Specialized Hardware

Cohen, Eldan; Mandal, Avradip; Ushijima‐Mwesigwa, Hayato

doi:10.1007/978-3-030-44584-3_9

Cited by 10 publications

(7 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We use the second generation of the DA that is capable of solving problems with up to 8192 variables and up to 64 bits of precision. The DA has previously been used in different areas such as communication [20], signal processing [28], and data mining [4,5].…”

Section: Solving Community Detection On the Fujitsu Digital Annealermentioning

confidence: 99%

“…Examples of these novel computing hardware include, adiabatic quantum computers, CMOS annealers, memristive circuits, and optical parametric oscillators, that are designed to solve optimization problems formulated as an Ising or QUBO mathematical model. Given that many well-known problems in graphs can easily be modeled in this form, there has been a growing interest in formulating and evaluating these problem and their subsequent QUBO models on the different specialized hardware platforms [4,15,21,32,34]. However, many real-world problems require significantly more variables than these devices can handle, thus hybrid methods are usually used.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Ising-Based Louvain Method: Clustering Large Graphs with Specialized Hardware

Kalehbasti

Ushijima‐Mwesigwa

Mandal

et al. 2021

Advances in Intelligent Data Analysis XIX

Self Cite

View full text Add to dashboard Cite

Recent advances in specialized hardware for solving optimization problems such quantum computers, quantum annealers, and CMOS annealers give rise to new ways for solving real-word complex problems. However, given current and near-term hardware limitations, the number of variables required to express a large real-world problem easily exceeds the hardware capabilities, thus hybrid methods are usually developed in order to utilize the hardware. In this work, we advocate for the development of hybrid methods that are built on top of the frameworks of existing state-of-art heuristics, thereby improving these methods. We demonstrate this by building on the so called Louvain method, which is one of the most popular algorithms for the Community detection problem and develop and Ising-based Louvain method. The proposed method outperforms two state-of-the-art community detection algorithms in clustering several small to large-scale graphs. The results show promise in adapting the same optimization approach to other unsupervised learning heuristics to improve their performance.

show abstract

Section: Solving Community Detection On the Fujitsu Digital Annealermentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Ising-Based Louvain Method: Clustering Large Graphs with Specialized Hardware

Kalehbasti

Ushijima‐Mwesigwa

Mandal

et al. 2021

Advances in Intelligent Data Analysis XIX

Self Cite

View full text Add to dashboard Cite

show abstract

“…Therefore, Ising-based constrained clustering can easily provide several different supervisory information simultaneously. Cohen et al also proposed Ising-based consensus clustering that is a method for obtaining a new optimal clustering result by integrating multiple known clustering results [40]. Ising-based consensus clustering also uses the one-hot constraint.…”

Section: Related Workmentioning

confidence: 99%

An External Definition of the One-Hot Constraint and Fast QUBO Generation for High-Performance Combinatorial Clustering

Kumagai

Komatsu

Takano³

et al. 2021

IJNC

View full text Add to dashboard Cite

Recently, a clustering method using a combinatorial optimization problem, called combinatorial clustering, has been drawing attention due to the rapid spreads of quantum annealing and simulated annealing. Combinatorial clustering is performed by minimizing an objective function under a condition to satisfy a one-hot constraint. The objective function and the constraint function are generally formulated to a unified objective function of a QUBO (Quadratic Unconstrained Binary Optimization) problem using the method of the Lagrange multiplier. The coefficients of the QUBO function can be represented by a square matrix, which is called the QUBO matrix.Although the Lagrange multiplier needs to be large enough to avoid violating the constraint, it is usually hard to be set appropriately due to the limitation of the bit precision. For example, the latest quantum annealer can handle values represented by only six or fewer bits. Even conventional computing systems cannot control the larger value of the Lagrange multiplier as the number of data points increases. Besides, the execution time for combinatorial clustering increases exponentially as the problem size increases. This is because the time for the QUBO matrix generation is long and a dominant factor of the total execution time when the problem size is large.To solve these problems, this paper proposes combinatorial clustering that overcomes the limitation of the method of the Lagrange multiplier. The proposed method uses a QUBO solver

show abstract

“…We use the second generation of the DA that is capable of solving problems with up to 8192 variables and up to 64 bits of precision. The DA has previously been used in different areas such as communication [13], signal processing [14], and data mining [8,15].…”

Section: Solving Unified Clustering and Outlier Detection Problems On The Fujitsu Digital Annealermentioning

confidence: 99%

“…Consequently, there is a growing interest in formulating and evaluating QUBO models for key computational problems. In particular, significant attention was given to a range of data mining problems, e.g., clustering, community detection, and graph partitioning [5,6,7,8]. * Work done while at Fujitsu Laboratories of America.…”

Section: Introductionmentioning

confidence: 99%

Unified Clustering and Outlier Detection on Specialized Hardware

Cohen

Ushijima‐Mwesigwa

Mandal

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

Clustering and outlier detection are often studied as separate problems. However, previous work has shown that a unified approach can lead to better performance. Unified clustering and outlier detection is a hard combinatorial problem that has received significant attention in recent years. The recent emergence of specialized optimization hardware capable of solving combinatorial problems formulated as Quadratic Unconstrained Binary Optimization (QUBO) models has led to increased interest in harnessing these platforms in core data mining tasks. In this work, we present a novel QUBO formulation of the unified clustering and outlier detection problem and use the Fujitsu Digital Annealer, a specialized CMOS hardware, to solve it. Experiments on synthetic and real datasets demonstrate the effectiveness of our approach.

show abstract

Ising-Based Consensus Clustering on Specialized Hardware

Cited by 10 publications

References 27 publications

Ising-Based Louvain Method: Clustering Large Graphs with Specialized Hardware

Ising-Based Louvain Method: Clustering Large Graphs with Specialized Hardware

An External Definition of the One-Hot Constraint and Fast QUBO Generation for High-Performance Combinatorial Clustering

Unified Clustering and Outlier Detection on Specialized Hardware

Contact Info

Product

Resources

About