| Presentation | 2012-11-08 Multi-Instance SVM using Parametric Programming Naoki ISHIHARA, Saori KURUMI, Ichiro TAKEUCHI, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In this paper, we propose a new optimization algorithm for multiple instance learning (MIL). MIL is a supervised learning paradigm in which supervised labels are only assigned to a set of instances called bags. MIL algorithms are usually formulated as non-convex optimization problems because the labels of instances are unknown. In this study, we propose a homotopy-based non-convex optimization algorithm for finding local optimal solutions of a well-known MIL algorithm called Multi-Instance SVM (MI-SVM). To this end, we introduce a parametric program (a family of parametrized optimization problems) which contains both of convex standard SVM and non-convex MI-SVM. Then, we propose an algorithm that can compute a path of solutions of the parametric program when the former standard SVM problem is gradually modified to the latter MI-SVM problem. We demonstrate the behavior of our proposed approach through simple experiments. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Multi-Instance Learning / Non-convex Optimization / Parametric Optimization / Homotopy Method |
| Paper # | IBISML2012-72 |
| Date of Issue |
| Conference Information | |
| Committee | IBISML |
|---|---|
| Conference Date | 2012/10/31(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Information-Based Induction Sciences and Machine Learning (IBISML) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Multi-Instance SVM using Parametric Programming |
| Sub Title (in English) | |
| Keyword(1) | Multi-Instance Learning |
| Keyword(2) | Non-convex Optimization |
| Keyword(3) | Parametric Optimization |
| Keyword(4) | Homotopy Method |
| 1st Author's Name | Naoki ISHIHARA |
| 1st Author's Affiliation | Department of Engineering, Nagoya Institute of Technology() |
| 2nd Author's Name | Saori KURUMI |
| 2nd Author's Affiliation | Department of Engineering, Nagoya Institute of Technology |
| 3rd Author's Name | Ichiro TAKEUCHI |
| 3rd Author's Affiliation | Department of Engineering, Nagoya Institute of Technology |
| Date | 2012-11-08 |
| Paper # | IBISML2012-72 |
| Volume (vol) | vol.112 |
| Number (no) | 279 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |