Skip to main content

Hitachi

Corporate InformationResearch & Development

Sample Design Method for Adaptive Group Testing

— Presentation at APSIPA ASC 2016 —

February 15, 2017

Report from Presenter

The Asia-Pacific Signal and Information Processing Association (APSIPA) Annual Summit and Conference (ASC) were held by APSIPA in Jeju, Korea during 13-16 December 2016. APSIPA ASC is a major conference on signal processing in the Asia-Pacific region. A lot of research results on wide areas are presented in APSIPA ASC every year. For example, research topics are theory and methods of signal processing, audio and speech signal processing, image and video signal processing, signal processing for wireless communication, bio-signal processing, etc. In APSIPA ASC 2016, 242 papers were presented, and the cutting edge of signal processing was actively discussed.

In APSIPA ASC 2016, Hitachi, Ltd. Research & Development Group made a presentation titled "Adaptive Boolean compressive sensing by sequential pool-design". This presentation is about a method for solving "group-testing". Group-testing is a problem that appears in many areas such as genetic screening, anomaly detection in network, substance monitoring in wide-area, etc. So, a method for solving group-testing can be also used to many applications. The goal of group-testing is to find a small number of anomalies from among a lot of samples. If each sample is tested individually, a large number of tests will be necessary. Therefore, in group-testing, to reduce the number of tests, multiple samples are mixed into one mixed-sample, and the mixed sample is tested instead of testing each sample individually. By repeating these steps changing the combination of mixing, a sequence of test results is obtained. Then, based on the sequence of test results, the anomalies can be found through only a small number of tests.


Fig. 1 Proposed approach
Enlarge


Fig. 2 Simulation results
Enlarge

We had already proposed a method for controlling the number of samples selected for mixing in order to make group-testing possible regardless of the number of the anomalies in IEEE ICASSP 2016 on March 2016, and the conventional method is based on "compressive sensing" and "multi-armed bandit". The method proposed in APSIPA ASC 2016 is a combination of the conventional method and a new procedure that selects samples for mixing such that the next selection is not similar to the past ones (Figure 1). The proposed method can find the anomalies through smaller number of tests than the conventional method (Figure 2).

(By KAWAGUCHI Yohei)

Related Papers

  • Y. Kawaguchi and M. Togami, "Adaptive Boolean Compressive Sensing by Sequential Pool-Design", in Proc. APSIPA ASC, 2016.
  • Y. Kawaguchi and M. Togami, "Adaptive Boolean Compressive Sensing by Using Multi-Armed Bandit", in Proc. IEEE ICASSP, 2016.
  • Page top