Consensus inference with multilayer graphs for multi-modal data

Published on Nov 1, 2014 in ASILOMAR (Asilomar Conference on Signals, Systems and Computers)
· DOI :10.1109/ACSSC.2014.7094679
Karthikeyan Natesan Ramamurthy14
Estimated H-index: 14
Jayaraman J. Thiagarajan15
Estimated H-index: 15
(LLNL: Lawrence Livermore National Laboratory)
+ 2 AuthorsRamanathan Nachiappan1
Estimated H-index: 1
(SSN: Sri Sivasubramaniya Nadar College of Engineering)
Emergence of numerous modalities for data generation necessitates the development of machine learning techniques that can perform efficient inference with multi-modal data. In this paper, we present an approach to learn discriminant low-dimensional projections from supervised multi-modal data. We construct intra- and inter-class similarity graphs for each modality and optimize for consensus projections in the kernel space. Features obtained with these projections can then be used to train a classifier for consensus inference. We also provide methods for out-of-sample extensions with novel test data. Classification results with standard multi-modal data sets demonstrate the efficacy of our method.
  • References (18)
  • Citations (1)
📖 Papers frequently viewed together
2012ICDM: International Conference on Data Mining
15 Citations
1 Author (Petra Perner)
4 Authors (Qi Tian, ..., Huang SiLiang)
15 Citations
78% of Scinapse members use related papers. After signing in, all features are FREE.
We resolve two problems regarding subspace distances that have arisen considerably often in applications: How could one define a notion of distance between (i) two linear subspaces of different dimensions, or (ii) two affine subspaces of the same dimension, in a way that generalizes the usual Grassmann distance between equidimensional linear subspaces? We show that (i) is the distance of a point to a Schubert variety, and (ii) is the distance in the Grassmannian of affine subspaces, both regarde...
17 Citations
#1Fayao Liu (University of Adelaide)H-Index: 11
#2Luping Zhou (UOW: University of Wollongong)H-Index: 17
Last. Jianping Yin (National University of Defense Technology)H-Index: 24
view all 4 authors...
To achieve effective and efficient detection of Alzheimer's disease (AD), many machine learning methods have been introduced into this realm. However, the general case of limited training samples, as well as different feature representations typically makes this problem challenging. In this paper, we propose a novel multiple kernel-learning framework to combine multimodal features for AD classification, which is scalable and easy to implement. Contrary to the usual way of solving the problem in ...
58 CitationsSource
#2Aleksandr Y. Aravkin (UW: University of Washington)H-Index: 8
Last. Jayaraman J. Thiagarajan (LLNL: Lawrence Livermore National Laboratory)H-Index: 15
view all 3 authors...
Incorporating sparsity priors in learning tasks can give rise to simple, and interpretable models for complex high dimensional data. Sparse models have found widespread use in structure discovery, recovering data from corruptions, and a variety of large scale unsupervised and supervised learning problems. Assuming the availability of sufficient data, these methods infer dictionaries for sparse representations by optimizing for high-fidelity reconstruction. In most scenarios, the reconstruction q...
1 Citations
#1Xiaowen Dong ('ENS Paris': École Normale Supérieure)H-Index: 13
#2Pascal Frossard ('ENS Paris': École Normale Supérieure)H-Index: 43
Last. Nikolai Nefedov (ETH Zurich)H-Index: 8
view all 4 authors...
Relationships between entities in datasets are often of multiple nature, like geographical distance, social relationships, or common interests among people in a social network, for example. This information can naturally be modeled by a set of weighted and undirected graphs that form a global multi-layer graph, where the common vertex set represents the entities and the edges on different layers capture the similarities of the entities in term of the different modalities. In this paper, we addre...
100 CitationsSource
Jun 16, 2013 in ICML (International Conference on Machine Learning)
#1Quoc V. Le (Google)H-Index: 62
#2Tamas Sarlos (Google)H-Index: 16
Last. Alexander J. Smola (Google)H-Index: 99
view all 3 authors...
Despite their successes, what makes kernel methods difficult to use in many large scale problems is the fact that computing the decision function is typically expensive, especially at prediction time. In this paper, we overcome this difficulty by proposing Fastfood, an approximation that accelerates such computation significantly. Key to Fastfood is the observation that Hadamard matrices when combined with diagonal Gaussian matrices exhibit properties similar to dense Gaussian random matrices. Y...
279 Citations
#1Chang XuH-Index: 23
#2Dacheng Tao (UTS: University of Technology, Sydney)H-Index: 101
Last. Chao Xu (PKU: Peking University)H-Index: 28
view all 3 authors...
In recent years, a great many methods of learning from multi-view data by considering the diversity of different views have been proposed. These views may be obtained from multiple sources or different feature subsets. In trying to organize and highlight similarities and differences between the variety of multi-view learning approaches, we review a number of representative multi-view learning algorithms in different areas and classify them into three groups: 1) co-training, 2) multiple kernel le...
492 Citations
#1Jun Yu (Ha Tai: Xiamen University)H-Index: 31
#2Meng Wang (Hefei University of Technology)H-Index: 13
Last. Dacheng Tao (UTS: University of Technology, Sydney)H-Index: 101
view all 3 authors...
In image processing, cartoon character classification, retrieval, and synthesis are critical, so that cartoonists can effectively and efficiently make cartoons by reusing existing cartoon data. To successfully achieve these tasks, it is essential to extract visual features that comprehensively represent cartoon characters and to construct an accurate distance metric to precisely measure the dissimilarities between cartoon characters. In this paper, we introduce three visual features, color histo...
190 CitationsSource
Jun 28, 2011 in ICML (International Conference on Machine Learning)
#1Abhishek Kumar (UMD: University of Maryland, College Park)H-Index: 21
#2Hal Daumé (UMD: University of Maryland, College Park)H-Index: 45
We propose a spectral clustering algorithm for the multi-view setting where we have access to multiple views of the data, each of which can be independently used for clustering. Our spectral clustering algorithm has a flavor of co-training, which is already a widely used idea in semi-supervised learning. We work on the assumption that the true underlying clustering would assign a point to the same cluster irrespective of the view. Hence, we constrain our approach to only search for the clusterin...
401 Citations
Jun 28, 2011 in ICML (International Conference on Machine Learning)
#1Jiquan Ngiam (Stanford University)H-Index: 13
#2Aditya Khosla (Stanford University)H-Index: 36
Last. Andrew Y. Ng (Stanford University)H-Index: 105
view all 6 authors...
Deep networks have been successfully applied to unsupervised feature learning for single modalities (e.g., text, images or audio). In this work, we propose a novel application of deep networks to learn features over multiple modalities. We present a series of tasks for multimodal learning and show how to train deep networks that learn features to address these tasks. In particular, we demonstrate cross modality feature learning, where better features for one modality (e.g., video) can be learned...
1,523 Citations
Last. Petros DarasH-Index: 24
view all 3 authors...
In this paper a novel approach for multimodal search and retrieval is introduced. The searchable items are media representations consisting of multiple modalities, such as 2D images and 3D objects, which share a common semantic concept. The proposed method combines the low-level feature distances of each separate modality to construct a new low-dimensional feature space, where all media objects are mapped irrespective of their constituting modalities. While most of the existing state-of-the-art ...
6 CitationsSource
Cited By1
Mar 20, 2016 in ICASSP (International Conference on Acoustics, Speech, and Signal Processing)
#1Huan Songg (ASU: Arizona State University)H-Index: 1
#2Jayaraman J. Thiagarajan (LLNL: Lawrence Livermore National Laboratory)H-Index: 15
Last. Pavan Turaga (ASU: Arizona State University)H-Index: 26
view all 5 authors...
The pervasive use of wearable sensors in activity and health monitoring presents a huge potential for building novel data analysis and prediction frameworks. In particular, approaches that can harness data from a diverse set of low-cost sensors for recognition are needed. Many of the existing approaches rely heavily on elaborate feature engineering to build robust recognition systems, and their performance is often limited by the inaccuracies in the data. In this paper, we develop a novel two-st...
6 CitationsSource