Multi-utility Learning: Structured-Output Learning with Multiple Annotation-Specific Loss Functions
Structured-output learning is a challenging problem; particularly so because of the difficulty in obtaining large datasets of fully labelled instances for training. In this paper we try to overcome this difficulty by presenting a multi-utility learning framework for structured prediction that can learn from training instances with different forms of supervision. We propose a unified technique for inferring the loss functions most suitable for quantifying the consistency of solutions with the given weak annotation. We demonstrate the effectiveness of our framework on the challenging semantic image segmentation problem for which a wide variety of annotations can be used. For instance, the popular training datasets for semantic segmentation are composed of images with hard-to-generate full pixel labellings, as well as images with easy-to-obtain weak annotations, such as bounding boxes around objects, or image-level labels that specify which object categories are present in an image. Experimental evaluation shows that the use of annotation-specific loss functions dramatically improves segmentation accuracy compared to the baseline system where only one type of weak annotation is used.
In the modern Web, it is common for an active person to have several profiles in different online social networks. As new general-purpose and niche social network services arise every year, the problem of social data integration will likely remain actual in the nearest future. Discovering multiple profiles of a single person across different social networks allows to merge all user's contacts from different social services or compose more complete social graph that is helpful in many social-powered applications. In this paper we propose a new approach for user profile matching based on Conditional Random Fields that extensively combines usage of profile attributes and social linkage. It is extremely suitable for cases when profile data is poor, incomplete or hidden due to privacy settings. Evaluation on Twitter and Facebook sample datasets showed that our solution significatnly outperforms common attribute-based approach and is able to find matches that are not discoverable by using only profile information. We also demonstrate the importance of social links for identity resolution task and show that certain profiles can be matched based only on social relationships between OSN users.
The Shape Boltzmann Machine (SBM) and its multilabel version MSBM have been recently introduced as deep generative models that capture the variations of an object shape. While being more flexible MSBM requires datasets with labeled parts of the objects for training. In the paper we present an algorithm for training MSBM using binary masks of objects and the seeds which approximately correspond to the locations of objects parts. The latter can be obtained from part-based detectors in an unsupervised manner. We derive a latent variable model and an EM-like training procedure for adjusting the weights of MSBM using a deep learning framework. We show that the model trained by our method outperforms SBM in the tasks related to binary shapes and is very close to the original MSBM in terms of quality of multilabel shapes.
In this paper we consider the Shape Boltzmann Machine(SBM) and its multi-label version MSBM. We present an algorithm for training MSBM using only binary masks of objects and the seeds which approximately correspond to the locations of objects parts.
We present a new click model for processing click logs and predicting relevance and appeal for query–document pairs in search results. Our model is a simplified version of the task-centric click model but outperforms it in an experimental comparison.
In the paper we present a new framework for dealing with probabilistic graphical models. Our approach relies on the recently proposed Tensor Train format (TT-format) of a tensor that while being compact allows for efficient application of linear algebra operations. We present a way to convert the energy of a Markov random field to the TT-format and show how one can exploit the properties of the TT-format to attack the tasks of the partition function estimation and the MAP-inference. We provide theoretical guarantees on the accuracy of the proposed algorithm for estimating the partition function and compare our methods against several state-of-the-art algorithms.
In this paper we address the problem of finding the most probable state of a discrete Markov random field (MRF), also known as the MRF energy minimization problem. The task is known to be NP-hard in general and its practical importance motivates numerous approximate algorithms. We propose a submodular relaxation approach (SMR) based on a Lagrangian relaxation of the initial problem. Unlike the dual decomposition approach of Komodakis et al., 2011 SMR does not decompose the graph structure of the initial problem but constructs a submodular energy that is minimized within the Lagrangian relaxation. Our approach is applicable to both pairwise and high-order MRFs and allows to take into account global potentials of certain types. We study theoretical properties of the proposed approach and evaluate it experimentally.