Model combination, often regarded as a key sub-field of ensemble learning, has been widely used in both academic research and industry applications. To facilitate this process, we propose and implement an easy-to-use Python toolkit, combo, to aggregate models and scores under various scenarios, including classification, clustering, and anomaly detection. In a nutshell, combo provides a unified and consistent way to combine both raw and pretrained models from popular machine learning libraries, e.g., scikit-learn, XGBoost, and LightGBM. With accessibility and robustness in mind, combo is designed with detailed documentation, interactive examples, continuous integration, code coverage, and maintainability check; it can be installed easily through Python Package Index (PyPI) or {https://github.com/yzhao062/combo}.

Although several automatic computer systems have been proposed to address facial expression recognition problems, the majority of them still fail to cope with some requirements of many practical application scenarios. In this paper, one of the most influential and common issues raised in practical application scenarios when applying automatic facial expression recognition system, head pose variation, is comprehensively explored and investigated. In order to do this, two novel texture feature representations are proposed for implementing multi-view facial expression recognition systems in practical environments. These representations combine the block-based techniques with Local Ternary Pattern-based features, providing a more informative and efficient feature representation of the facial images. In addition, an in-house multi-view facial expression database has been designed and collected to allow us to conduct a detailed research study of the effect of out-of-plane pose angles on the performance of a multi-view facial expression recognition system. Along with the proposed in-house dataset, the proposed system is tested on two well-known facial expression databases, CK+ and BU-3DFE datasets. The obtained results shows that the proposed system outperforms current state-of-the-art 2D facial expression systems in the presence of pose variations.

In the context of facial expression recognition (FER), this paper reviews the fundamental theories of emotions and further explains the key dimensions of a defined emotional space. The main contribution of this paper is to propose a set of novel categorization methods for facial expressions to be used in the design of an automatic FER system. This novel categorization enables the facial expression to be interpreted in a better way that and to be more effective in practical applications of automatic FER systems. In order to validate the feasibility of the proposed categorization methods, a set of experiments is reported which investigates and analyzes the influence that the novel categorization brings to a multi-view FER system.

Given a labeled dataset that contains a rare (or minority) class containing of-interest instances, as well as a large class of instances that are not of interest, how can we learn to recognize future of-interest instances over a continuous stream? The setting is different from traditional classification in that instances from novel minority subclasses might continually emerge over time—and hence is often referred as continual, life-long, or open-world classification. We introduce RaRecognize, which ( i ) estimates a general decision boundary between the rare class and the majority class, ( ii ) learns to recognize the individual rare subclasses that exist within the training data, as well as ( iii ) flags instances from previously unseen rare subclasses as newly emerging (i.e., novel). The learner in (i) is general in the sense that by construction it is dissimilar to the specialized learners in (ii) , thus distinguishes minority from the majority without overly tuning to what is only seen in the training data. Thanks to this generality, RaRecognize ignores all future instances that it labels as majority and recognizes the recurring as well as emerging rare subclasses only. This saves effort at test time as well as ensures that the model size grows moderately over time as it only maintains specialized minority learners. Overall, we build an end-to-end system which consists of (1) a representation learning component that transforms data instances into suitable vector inputs; (2) a continual classifier that labels incoming instances as majority (not of interest), rare recurrent, or rare emerging; and (3) a clustering component that groups the rare emerging instances into novel subclasses for expert vetting and model re-training. Through extensive experiments, we show that RaRecognize outperforms state-of-the art baselines on three real-world datasets that contain documents related to corporate-risk and (natural and man-made) disasters as rare classes.

Reinforcement learning (RL) has recently been introduced to interactive recommender systems (IRS) because of its nature of learning from dynamic interactions and planning for long-run performance. As IRS is always with thousands of items to recommend (i.e., thousands of actions), most existing RL-based methods, however, fail to handle such a large discrete action space problem and thus become inefficient. The existing work that tries to deal with the large discrete action space problem by utilizing the deep deterministic policy gradient framework suffers from the inconsistency between the continuous action representation (the output of the actor network) and the real discrete action. To avoid such inconsistency and achieve high efficiency and recommendation effectiveness, in this paper, we propose a Tree-structured Policy Gradient Recommendation (TPGR) framework, where a balanced hierarchical clustering tree is built over the items and picking an item is formulated as seeking a path from the root to a certain leaf of the tree. Extensive experiments on carefully-designed environments based on two real-world datasets demonstrate that our model provides superior recommendation performance and significant efficiency improvement over state-of-the-art methods.

Link prediction is a fundamental problem with a wide range of applications in various domains, which predicts the links that are not yet observed or the links that may appear in the future. Most existing works in this field only focus on modeling a single network, while real-world networks are actually aligned with each other. Network alignments contain valuable additional information for understanding the networks, and provide a new direction for addressing data insufficiency and alleviating cold start problem. However, there are rare works leveraging network alignments for better link prediction. Besides, neural network is widely employed in various domains while its capability of capturing high-level patterns and correlations for link prediction problem has not been adequately researched yet. Hence, in this paper we target atlink prediction over aligned networks using neural networks. The major challenge is the heterogeneousness of the considered networks, as the networks may have different characteristics, link purposes, etc. To overcome this, we propose a novel multi-neural-network framework MNN, where we have one individual neural network for each heterogeneous target or feature while the vertex representations are shared. We further discuss training methods for the multi-neural-network framework. Extensive experiments demonstrate that MNN outperforms the state-of-the-art methods and achieves 3% to 5% relative improvement of AUC score across different settings, particularly over 8% for cold start scenarios.

As aggregators, online news portals face great challenges in continuously selecting a pool of candidate articles to be shown to their users. Typically, those candidate articles are recommended manually by platform editors from a much larger pool of articles aggregated from multiple sources. Such a hand-pick process is labor intensive and time-consuming. In this paper, we study the editor article selection behavior and propose a learning by demonstration system to automatically select a subset of articles from the large pool. Our data analysis shows that (i) editors' selection criteria are non-explicit, which are less based only on the keywords or topics, but more depend on the quality and attractiveness of the writing from the candidate article, which is hard to capture based on traditional bag-of-words article representation. And (ii) editors' article selection behaviors are dynamic: articles with different data distribution come into the pool everyday and the editors' preference varies, which are driven by some underlying periodic or occasional patterns. To address such problems, we propose a meta-attention model across multiple deep neural nets to (i) automatically catch the editors' underlying selection criteria via the automatic representation learning of each article and its interaction with the meta data and (ii) adaptively capture the change of such criteria via a hybrid attention model. The attention model strategically incorporates multiple prediction models, which are trained in previous days. The system has been deployed in a commercial article feed platform. A 9-day A/B testing has demonstrated the consistent superiority of our proposed model over several strong baselines.