ABSTRACT
Previous studies on multi-instance learning typically treated instances in the bags as independently and identically distributed. The instances in a bag, however, are rarely independent in real tasks, and a better performance can be expected if the instances are treated in an non-i.i.d. way that exploits relations among instances. In this paper, we propose two simple yet effective methods. In the first method, we explicitly map every bag to an undirected graph and design a graph kernel for distinguishing the positive and negative bags. In the second method, we implicitly construct graphs by deriving affinity matrices and propose an efficient graph kernel considering the clique information. The effectiveness of the proposed methods are validated by experiments.
- Amar, R. A., Dooly, D. R., Goldman, S. A., & Zhang, Q. (2001). Multiple-instance learning of real-valued data. Proc. 18th Intl. Conf. Mach. Learn. (pp. 3--10). Google ScholarDigital Library
- Andrews, S., Tsochantaridis, I., & Hofmann, T. (2003). Support vector machines for multiple-instance learning. In Adv. Neural Inf. Process. Syst. 15, 561--568. Cambridge, MA: MIT Press.Google Scholar
- Auer, P., & Ortner, R. (2004). A boosting approach to multiple instance learning. Proc. 15th Eur. Conf. Mach. Learn. (pp. 63--74).Google ScholarDigital Library
- Blockeel, H., Page, D., & Srinivasan, A. (2005). Multi-instance tree learning. Proc. 22nd Intl. Conf. Mach. Learn. (pp. 57--64). Google ScholarDigital Library
- Borgwardt, K. M., & Kriegel, H.-P. (2005). Shortestpath kernels on graphs. Proc. 5th IEEE Intl. Conf. Data Min. (pp. 74--81). Google ScholarDigital Library
- Chen, Y., Bi, J., & Wang, J. Z. (2006). MILES: Multiple-instance learning via embedded instance selection. IEEE Trans. Patt. Anal. Mach. Intell., 28, 1931--1947. Google ScholarDigital Library
- Chen, Y., & Wang, J. Z. (2004). Image categorization by learning and reasoning with regions. J. Mach. Learn. Res., 5, 913--939. Google ScholarDigital Library
- Cheung, P.-M., & Kwok, J. T. (2006). A regularization framework for multiple-instance learning. Proc. 23rd Intl. Conf. Mach. Learn. (pp. 193--200). Google ScholarDigital Library
- Chevaleyre, Y., & Zucker, J.-D. (2001). A framework for learning rules from multiple instance data. Proc. 12th Eur. Conf. Mach. Learn. (pp. 49--60). Google ScholarDigital Library
- Csurka, G., Bray, C., Dance, C., & Fan, L. (2004). Visual categorization with bags of keypoints. ECCV Workshop on Statistical Learning in Computer Vision (pp. 59--74).Google Scholar
- Dietterich, T. G., Lathrop, R. H., & Lozano-Péérez, T. (1997). Solving the multiple-instance problem with axis-parallel rectangles. Artif. Intell., 89, 31--71. Google ScholarDigital Library
- Fung, G., Dundar, M., Krishnappuram, B., & Rao, R. B. (2007). Multiple instance learning for computer aided diagnosis. In Adv. Neural Inf. Process. Syst. 19, 425--432. Cambridge, MA: MIT Press.Google Scholar
- Gärtner, T. (2003). A survey of kernels for structured data. SIGKDD Explorations, 5, 49--58. Google ScholarDigital Library
- Gärtner, T., Flach, P. A., Kowalczyk, A., & Smola, A. J. (2002). Multi-instance kernels. Proc. 19th Intl. Conf. Mach. Learn. (pp. 179--186). Google ScholarDigital Library
- Kwok, J. T., & Cheung, P.-M. (2007). Marginalized multi-instance kernels. Proc. 20th Intl. J. Conf. Artif. Intell. (pp. 901--906). Google ScholarDigital Library
- Maron, O., & Lozano-Péérez, T. (1998). A framework for multiple-instance learning. In Adv. Neural Inf. Process. Syst. 10, 570--576. Cambridge, MA: MIT Press. Google ScholarDigital Library
- McGovern, A., & Jensen, D. (2003). Identifying predictive structures in relational data using multiple instance learning. Proc. 20th Intl. Conf. Mach. Learn. (pp. 528--535).Google ScholarCross Ref
- Neuhaus, M., & Bunke, H. (2007). A quadratic programming approach to the graph edit distance problem. Proc. 6th IAPR Workshop on Graph-based Represent. in Patt. Recogn. (pp. 92--102). Google ScholarDigital Library
- Rahmani, R., & Goldman, S. A. (2006). MISSL: Multiple-instance semi-supervised learning. Proc. 23rd Intl. Conf. Mach. Learn. (pp. 705--712). Google ScholarDigital Library
- Ray, S., & Craven, M. (2005). Supervised versus multiple instance learning: An empirical comparison. Proc. 22nd Intl. Conf. Mach. Learn. (pp. 697--704). Google ScholarDigital Library
- Ray, S., & Page, D. (2001). Multiple instance regression. Proc. 18th Intl. Conf. Mach. Learn. (pp. 425--432). Google ScholarDigital Library
- Ruffo, G. (2000). Learning single and multiple instance decision trees for computer security applications. Doctoral dissertation, CS Dept., Univ. Turin, Torino, Italy.Google Scholar
- Scott, S. D., Zhang, J., & Brown, J. (2003). On generalized multiple-instance learning (Technical Report UNL-CSE-2003-5). CS Dept., Univ. Nebraska, Lincoln, NE.Google Scholar
- Settles, B., Craven, M., & Ray, S. (2008). Multiple-instance active learning. In Adv. Neural Inf. Process. Syst. 20, 1289--1296. Cambridge, MA: MIT Press.Google Scholar
- Stanfill, C., & Waltz, D. (1986). Toward memory-based reasoning. Comm. ACM, 29, 1213--1228. Google ScholarDigital Library
- Tenenbaum, J. B., de Silva, V., & Langford, J. C. (2000). A global geometric framework for nonlinear dimensionality reduction. Science, 290, 2319--2323.Google ScholarCross Ref
- Viola, P., Platt, J., & Zhang, C. (2006). Multiple instance boosting for object detection. In Adv. Neural Inf. Process. Syst. 18, 1419--1426. Cambridge, MA: MIT Press.Google Scholar
- Wang, H.-Y., Yang, Q., & Zha, H. (2008). Adaptive p-posterior mixture-model kernels for multiple instance learning. Proc. 25th Intl. Conf. Mach. Learn. (pp. 1136--1143). Google ScholarDigital Library
- Wang, J., & Zucker, J.-D. (2000). Solving the multi-instance problem: A lazy learning approach. Proc. 17th Intl. Conf. Mach. Learn. (pp. 1119--1125). Google ScholarDigital Library
- Weidmann, N., Frank, E., & Pfahringer, B. (2003). A two-level learning method for generalized multi-instance problem. Proc. 14th Eur. Conf. Mach. Learn. (pp. 468--479).Google ScholarDigital Library
- Xu, X., & Frank, E. (2004). Logistic regression and boosting for labeled bags of instances. Proc. 8th Pac.-Asia Conf. Knowl. Discov. Data Min. (pp. 272--281).Google ScholarCross Ref
- Zhang, C., & Viola, P. (2008). Multiple-instance pruning for learning efficient cascade detectors. In Adv. Neural Inf. Process. Syst. 20, 1681--1688. Cambridge, MA: MIT Press.Google Scholar
- Zhang, M.-L., & Zhou, Z.-H. (2006). Adapting RBF neural networks to multi-instance learning. Neural Process. Lett., 23, 1--26. Google ScholarDigital Library
- Zhang, Q., & Goldman, S. A. (2002). EM-DD: An improved multi-instance learning technique. In Adv. Neural Inf. Process. Syst. 14, 1073--1080. Cambridge, MA: MIT Press.Google Scholar
- Zhang, Q., Yu, W., Goldman, S. A., & Fritts, J. E. (2002). Content-based image retrieval using multiple-instance learning. Proc. 19th Intl. Conf. Mach. Learn. (pp. 682--689). Google ScholarDigital Library
- Zhou, Z.-H., & Xu, J.-M. (2007). On the relation between multi-instance learning and semi-supervised learning. Proc. 24th Intl. Conf. Mach. Learn. (pp. 1167--1174). Google ScholarDigital Library
- Zhou, Z.-H., & Zhang, M.-L. (2007). Multi-instance multi-label learning with application to scene classification. In Adv. Neural Inf. Process. Syst. 19, 1609--1616. Cambridge, MA: MIT Press.Google Scholar
Index Terms
- Multi-instance learning by treating instances as non-I.I.D. samples
Recommendations
On the relation between multi-instance learning and semi-supervised learning
ICML '07: Proceedings of the 24th international conference on Machine learningMulti-instance learning and semi-supervised learning are different branches of machine learning. The former attempts to learn from a training set consists of labeled bags each containing many unlabeled instances; the latter tries to exploit abundant ...
Non-I.I.D. multi-instance dimensionality reduction by learning a maximum bag margin subspace
AAAI'10: Proceedings of the Twenty-Fourth AAAI Conference on Artificial IntelligenceMulti-instance learning, as other machine learning tasks, also suffers from the curse of dimensionality. Although dimensionality reduction methods have been investigated for many years, multi-instance dimensionality reduction methods remain untouched. ...
Diversified dictionaries for multi-instance learning
Multiple-instance learning (MIL) has been a popular topic in the study of pattern recognition for years due to its usefulness for such tasks as drug activity prediction and image/text classification. In a typical MIL setting, a bag contains a bag-level ...
Comments