Abstract
With the popularity of photo-sharing websites, the number of web images has exploded into unseen magnitude. Annotating such large-scale data will cost huge amount of human resources and is thus unaffordable. Motivated by this challenging problem, we propose a novel sparse graph based multilabel propagation (SGMP) scheme for super large scale datasets. Both the efficacy and accuracy of the image annotation are further investigated under different graph construction strategies, where Gaussian noise and non-Gaussian sparse noise are simultaneously considered in the formulations of these strategies. Our proposed approach outperforms the state-of-the-art algorithms by focusing on: (1) For large-scale graph construction, a simple yet efficient LSH (Locality Sensitive Hashing)-based sparse graph construction scheme is proposed to speed up the construction. We perform the multilabel propagation on this hashing-based graph construction, which is derived with LSH approach followed by sparse graph construction within the individual hashing buckets; (2) To further improve the accuracy, we propose a novel sparsity induced scalable graph construction scheme, which is based on a general sparse optimization framework. Sparsity essentially implies a very strong prior: for large scale optimization, the values of most variables shall be zeros when the solution reaches the optimum. By utilizing this prior, the solutions of large-scale sparse optimization problems can be derived by solving a series of much smaller scale subproblems; (3) For multilabel propagation, different from the traditional algorithms that propagate over individual label independently, our proposed propagation first encodes the label information of an image as a unit label confidence vector and naturally imposes inter-label constraints and manipulates labels interactively. Then, the entire propagation problem is formulated on the concept of Kullback-Leibler divergence defined on probabilistic distributions, which guides the propagation of the supervision information. Extensive experiments on the benchmark dataset NUS-WIDE with 270k images and its lite version NUS-WIDE-LITE with 56k images well demonstrate the effectiveness and scalability of the proposed multi-label propagation scheme.
Supplemental Material
Available for Download
Supplemental movie, appendix, image and software files for, Large-scale multilabel propagation based on efficient sparse graph construction
- Andoni, A. and Indyk, P. 2008. Near-optimal hashing algorithms for approximate nearest neighbor in high dimensions. Comm. ACM 51, 1, 117--122. Google Scholar
Digital Library
- Andrew, G. and Gao, J. 2007. Scalable training of l1-regularized log-linear models. In Proceedings of ICML. Google Scholar
Digital Library
- Beck, A. and Teboulle, M. 2009. A fast iterative shrinkage-thresholding algorithm for linear inverse problems. SIAM J. Imag. Sci. Google Scholar
Digital Library
- Boyd, S. and Vandenberghe, L. 2004. Convex Optimization. Cambridge University Press. Google Scholar
Digital Library
- Cai, J.-F., Candes, E. J., and Shen, Z. 2010. A singular value thresholding algorithm for matrix completion. SIAM J. Optim. Google Scholar
Digital Library
- Charikar, M. 2002. Similarity estimation techniques from rounding algorithms. In Proceedings of STOC. Google Scholar
Digital Library
- Chen, G., Song, Y., Wang, F., and Zhang, C. 2008. Semi-supervised multi-label learning by solving a sylvester equation. In Proceedings of the SIAM International Conference on Data Mining.Google Scholar
- Chen, S., Donoho, D., and Saunders, M. 2001. Atomic decomposition by basis pursuit. SIAM review. Google Scholar
Digital Library
- Chen, X., Mu, Y., Yan, S., and Chua, T. 2010. Efficient large-scale image annotation by probabilistic collaborative multi-label propagation. In Proceedings of ACM MM. Google Scholar
Digital Library
- Cheng, B., Yang, J., Yan, S., Fu, Y., and Huang, T. S. 2010. Learning with ℓ1-graph for image analysis. IEEE Trans. Image Proc. Google Scholar
Digital Library
- Chua, T.-S., Tang, J., Hong, R., Li, H., Luo, Z., and Zheng, Y.-T. 2009. Nus-wide: A real-world web image database from national university of singapore. In Proceedings of CIVR. Google Scholar
Digital Library
- Collobert, R., Sinz, F. H., Weston, J., and Bottou, L. 2006. Large scale transductive SVMS. J. Mach. Learn. Rese. 7, 1687--1712. Google Scholar
Digital Library
- Cover, T. M. and Thomas, J. A. 1991. Elements of Information Theory. Wiley Series in Telecommunications. Google Scholar
Digital Library
- Datar, M., Immorlica, N., Indyk, P., and Mirrokni, V. 2004. Locality-sensitive hashing scheme based on p-stable distributions. In Proceedings of SCG. Google Scholar
Digital Library
- Delalleau, O., Bengio, Y., and Le Roux, N. 2005. Efficient non-parametric function induction in semi-supervised learning. In Proceedings of the 10th International Workshop on Artificial Intelligence and Statistics. 96--103.Google Scholar
- Duda, R., Stork, D., and Hart, P. 2000. Pattern Classification. Wiley. Google Scholar
Digital Library
- Gao, Y., Tang, J., Hong, R., Yan, S., Dai, Q., Zhang, N., and Chua, T. 2012. Camera constraint-free view-based 3D object retrieval. IEEE Trans. Image Proc. 21, 4, 2269--2281.Google Scholar
Digital Library
- Hale, E. T., Yin, W., and Zhang, Y. 2008. Fixed-point continuation for l1-minimization: Methodology and convergence. SIAM J. Optim. Google Scholar
Digital Library
- Hiriart-Urruty, J. and Lemaréchal, C. 2001. Fundamentals of Convex Analysis. Springer-Verlag.Google Scholar
- Indyk, P. and Motwani, R. 1998. Approximate nearest neighbors: Towards removing the curse of dimensionality. In Proceedings of the Symposium on Theory Computing. Google Scholar
Digital Library
- Joachims, T. 2003. Transductive learning via spectral graph partitioning. In Proceedings of the International Conference on Machine Learning (ICML).Google Scholar
- Karlen, M., Weston, J., Erkan, A., and Collobert, R. 2008. Large-scale manifold transduction. In Proceedings of ICML. Google Scholar
Digital Library
- Lee, H., Battle, A., Raina, R., and Ng, A. 2007. Efficient sparse coding algorithms. In Proceedings of NIPS.Google Scholar
- Liu, J., Ji, S., and Ye, J. 2009. Multi-task feature learning via efficient l2, 11-norm minimization. In Proceedings of UAI. Google Scholar
Digital Library
- Liu, Y., Jin, R., and Yang, L. 2006. Semi-supervised multi-label learning by constrained non-negative matrix factorization. In Proceedings of AAAI. Google Scholar
Digital Library
- Mu, Y., Shen, J., and Yan, S. 2010. Weakly-supervised hashing in kernel space. In Proceedings of CVPR.Google Scholar
- Mu, Y., Wright, J., and Chang, S.-F. 2012. Accelerated large scale optimization by concomitant hashing. In Proceedings of European Conference on Computer Vision (ECCV). Google Scholar
Digital Library
- Perkins, S., Lacker, K., and Theiler, J. 2003. Grafting: Fast, incremental feature selection by gradient descent in function space. J. Mech. Learn. Res. Google Scholar
Digital Library
- Qi, G.-J., Hua, X.-S., Rui, Y., Tang, J., Mei, T., and Zhang, H.-J. 2007. Correlative multi-label video annotation. In Proceedings of MM. Google Scholar
Digital Library
- Roweis, S. and Saul, L. 2000. Nonlinear dimensionality reduction by locally linear embedding. Science 290, 2323--2326.Google Scholar
Cross Ref
- Russell, S. and Norvig, P. 2009. Artificial Intelligence: A Modern Approach. Prentice Hall. Google Scholar
Digital Library
- Schmidt, M., Van Den Berg, E., Friedlander, M., and Murphy, K. 2009. Optimizing costly functions with simple constraints: A limited-memory projected quasi-newton algorithm. In Proceedings of UAI.Google Scholar
- Sindhwani, V. and Keerthi, S. S. 2006. Large scale semi-supervised linear svms. In Proceedings of the 29th Annual International ACM SIGIR (SIGIR'06). Google Scholar
Digital Library
- Subramanya, A. and Bilmes, J. 2008. Soft-supervised learning for text classification. In Proceedings of EMNLP. Google Scholar
Digital Library
- Subramanya, A. and Bilmes, J. 2009. Entropic graph regularization in non-parametric semi-supervised classification. In Proceedings of NIPS.Google Scholar
- Tang, J., Yan, S., Hong, R., Qi, G.-J., and Chua, T.-S. 2009. Inferring semantic concepts from community-contributed images and noisy tags. In Proceedings of MM. Google Scholar
Digital Library
- Tibshirani, R. 1996. Regression shrinkage and selection via the lasso. J. Roy. Stat. Soc. Ser. B (Methodological).Google Scholar
- Tsang, I. W. and Kwok, J. T. 2006. Large-scale sparsified manifold regularization. In Proceedings of NIPS.Google Scholar
- Van Den Berg, E. and Friedlander, M. 2008. Probing the Pareto frontier for basis pursuit solutions. SIAM J. Sci. Comput. Google Scholar
Digital Library
- Wang, F. and Zhang, C. 2006. Label propagation through linear neighborhoods. In Proceedings of ICML. Google Scholar
Digital Library
- Wang, M., Hua, X.-S., Hong, R., Tang, J., Qi, G.-J., and Song, Y. 2012. Unified video annotation via multi-graph learning. IEEE Trans. Circ. Syst. Video Tech. 19, 5, 733--746. Google Scholar
Digital Library
- Zha, Z.-J., Mei, T., Wang, J., Wang, Z., and Hua, X.-S. 2009. Graph-based semi-supervised learning with multiple labels. J. Visual Commun. Image Represent. 20, 2, 97--103. Google Scholar
Digital Library
- Zhang, H., Zha, Z.-J., Yan, S., Wang, M., and Chua, T.-S. 2012. Robust non-negative graph embedding: Towards noisy data, unreliable graphs, and noisy labels. In Proceedings of CVPR. Google Scholar
Digital Library
- Zhu, X. 2005. Semi-Supervised Learning with Graphs. Carnegie Mellon University.Google Scholar
- Zhu, X. 2006. Semi-Supervised Learning Literature Survey. Carnegie Mellon University.Google Scholar
- Zhu, X., Ghahramani, Z., and Lafferty, J. 2003. Semi-supervised learning using Gaussian fields and harmonic functions. In Proceedings of the International Conference on Machine Learning (ICML).Google Scholar
Index Terms
Large-scale multilabel propagation based on efficient sparse graph construction
Recommendations
Efficient large-scale image annotation by probabilistic collaborative multi-label propagation
MM '10: Proceedings of the 18th ACM international conference on MultimediaAnnotating large-scale image corpus requires huge amount of human efforts and is thus generally unaffordable, which directly motivates recent development of semi-supervised or active annotation methods. In this paper we revisit this notoriously ...
Transductive Multilabel Learning via Label Set Propagation
The problem of multilabel classification has attracted great interest in the last decade, where each instance can be assigned with a set of multiple class labels simultaneously. It has a wide variety of real-world applications, e.g., automatic image ...
Graph-based semi-supervised learning with multi-modality propagation for large-scale image datasets
Semi-supervised learning (SSL) is widely-used to explore the vast amount of unlabeled data in the world. Over the decade, graph-based SSL becomes popular in automatic image annotation due to its power of learning globally based on local similarity. ...






Comments