| Presentation | 2022-04-22 A Study on Parametric Bilinear Inference via Gaussian Belief Propagation Kenta Ito, Takumi Takahashi, Koji Igarashi, Shinsuke Ibi, Seiichi Sampei, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Parametric bilinear inference estimates two unknown varables (i.e., parameters) with a symmetric linear regression structure from a multi-dimensional observation, and is expected to be applied to a wide range of fields of engineering, science, and finance. To solve this inference problem with low computational cost, parametric bilinear generalized approximate message passing (PBiGAMP), which is derived from sum-product algorithm (SPA) using the GAMP framework, has been proposed. However, the derivation relies heavily on the large-system limit assumption and the message update rule that forcibly extend the GAMP approach is not mathematically consistent, making it difficult to estimate with high accuracy even for sufficiently large systems. To circumvent this issue, we propose a novel parametric bilinear inference algorithm via Gaussian belief propagation (GaBP), which can be derived from SPA with fewer approximations while maintaining the consistency of the Bayesian update rule. The efficacy of the proposed method over the PBiGAMP is confirmed via computer simulations in terms of the normalized mean square error (NSE) performance and the symbol error rate (SER) performance. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Parametric bilinear inference / Gaussian belief propagation / generalized approximate message passing |
| Paper # | RCS2022-8 |
| Date of Issue | 2022-04-14 (RCS) |
| Conference Information | |
| Committee | RCS |
|---|---|
| Conference Date | 2022/4/21(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Osaka University, and online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Railroad Communications, Inter-Vehicle Communications, Road to Vehicle Communications, Radio Access Technologies, Wireless Communications, etc. |
| Chair | Eiji Okamoto(Nagoya Inst. of Tech.) |
| Vice Chair | Toshihiko Nishimura(Hokkaido Univ.) / Tomoya Tandai(Toshiba) / Fumihide Kojima(NICT) |
| Secretary | Toshihiko Nishimura(NEC) / Tomoya Tandai(Panasonic) / Fumihide Kojima |
| Assistant | Koichi Adachi(Univ. of Electro-Comm.) / Osamu Nakamura(Sharp) / Manabu Sakai(Mitsubishi Electric) / Masashi Iwabuchi(NTT) / Tatsuki Okuyama(NTT DOCOMO) |
| Paper Information | |
| Registration To | Technical Committee on Radio Communication Systems |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A Study on Parametric Bilinear Inference via Gaussian Belief Propagation |
| Sub Title (in English) | |
| Keyword(1) | Parametric bilinear inference |
| Keyword(2) | Gaussian belief propagation |
| Keyword(3) | generalized approximate message passing |
| 1st Author's Name | Kenta Ito |
| 1st Author's Affiliation | Osaka University(Osaka Univ.) |
| 2nd Author's Name | Takumi Takahashi |
| 2nd Author's Affiliation | Osaka University(Osaka Univ.) |
| 3rd Author's Name | Koji Igarashi |
| 3rd Author's Affiliation | Osaka University(Osaka Univ.) |
| 4th Author's Name | Shinsuke Ibi |
| 4th Author's Affiliation | Doshisha University(Doshisha Univ.) |
| 5th Author's Name | Seiichi Sampei |
| 5th Author's Affiliation | Osaka University(Osaka Univ.) |
| Date | 2022-04-22 |
| Paper # | RCS2022-8 |
| Volume (vol) | vol.122 |
| Number (no) | RCS-6 |
| Page | pp.pp.35-40(RCS), |
| #Pages | 6 |
| Date of Issue | 2022-04-14 (RCS) |