Pose-based action recognition in 3D is the task of recognizing an action (e.g., walking or running) from a sequence of 3D skeletal poses. This is challenging because of variations due to different ways of performing the same action and inaccuracies in the estimation of the skeletal poses. The training data is usually small and hence complex classifiers risk over-fitting the data. We address this task by action-snippets which are short sequences of consecutive skeletal poses capturing the temporal relationships between poses in an action. We propose a novel representation for action-snippets, called activated simplices. Each activity is represented by a manifold which is approximated by an arrangement of activated simplices. A sequence (of action-snippets) is classified by selecting the closest manifold and outputting the corresponding activity. This is a simple classifier which helps avoid over-fitting the data but which significantly outperforms state-of-the-art methods on standard benchmarks.

Unsupervised domain adaptation (UDA) deals with the task that labeled training and unlabeled test data collected from source and target domains, respectively. In this paper, we particularly address the practical and challenging scenario of imbalanced cross-domain data. That is, we do not assume the label numbers across domains to be the same, and we also allow the data in each domain to be collected from multiple datasets/sub-domains. To solve the above task of imbalanced domain adaptation, we propose a novel algorithm of Domain-constraint Transfer Coding (DcTC). Our DcTC is able to exploit latent subdomains within and across data domains, and learns a common feature space for joint adaptation and classification purposes. Without assuming balanced cross-domain data as most existing UDA approaches do, we show that our method performs favorably against state-of-the-art methods on multiple cross-domain visual classification tasks.

Speaker identification refers to the task of localizing the face of a person who has the same identity as the ongoing voice in a video. This task not only requires collective perception over both visual and auditory signals, the robustness to handle severe quality degradations and unconstrained content variations are also indispensable. In this paper, we describe a novel multimodal Long Short-Term Memory (LSTM) architecture which seamlessly unifies both visual and auditory modalities from the beginning of each sequence input. The key idea is to extend the conventional LSTM by not only sharing weights across time steps, but also sharing weights across modalities. We show that modeling the temporal dependency across face and voice can significantly improve the robustness to content quality degradations and variations. We also found that our multimodal LSTM is robustness to distractors, namely the non-speaking identities. We applied our multimodal LSTM to The Big Bang Theory dataset and showed that our system outperforms the state-of-the-art systems in speaker identification with lower false alarm rate and higher recognition accuracy.

Viewpoint estimation, especially in case of multiple object classes, remains an important and challenging problem. First, objects under different views undergo extreme appearance variations, often making within-class variance larger than between-class variance. Second, obtaining precise ground truth for real-world images, necessary for training supervised viewpoint estimation models, is extremely difficult and time consuming. As a result, annotated data is often available only for a limited number of classes. Hence it is desirable to share viewpoint information across classes. Additional complexity arises from unaligned pose labels between classes, i.e. a side view of a car might look more like a frontal view of a toaster, than its side view. To address these problems, we propose a metric learning approach for joint class prediction and pose estimation. Our approach allows to circumvent the problem of viewpoint alignment across multiple classes, and does not require dense viewpoint labels. Moreover, we show, that the learned metric generalizes to new classes, for which the pose labels are not available, and therefore makes it possible to use only partially annotated training sets, relying on the intrinsic similarities in the viewpoint manifolds. We evaluate our approach on two challenging multi-class datasets, 3DObjects and PASCAL3D+.

In this paper, we focus on automatically detecting events in unconstrained videos without the use of any visual training exemplars. In principle, zero-shot learning makes it possible to train an event detection model based on the assumption that events (e.g. birthday party) can be described by multiple mid-level semantic concepts (e.g. ``blowing candle'', ``birthday cake''). Towards this goal, we first pre-train a bundle of concept classifiers using data from other sources. Then we evaluate the semantic correlation of each concept w.r.t. the event of interest and pick up the relevant concept classifiers, which are applied on all test videos to get multiple prediction score vectors. While most existing systems combine the predictions of the concept classifiers with fixed weights, we propose to learn the optimal weights of the concept classifiers for each testing video by exploring a set of online available videos with free-form text descriptions of their content. To validate the effectiveness of the proposed approach, we have conducted extensive experiments on the latest TRECVID MEDTest 2014, MEDTest 2013 and CCV dataset. The experimental results confirm the superiority of the proposed approach.

Human vision greatly benefits from the information about sizes of objects. The role of size in several visual reasoning tasks has been thoroughly explored in human perception and cognition. However, the impact of the information about sizes of objects is yet to be determined in AI. We postulate that this is mainly attributed to the lack of a comprehensive repository of size information. In this paper, we introduce a method to automatically infer object sizes, leveraging visual and textual information from web. By maximizing the joint likelihood of textual and visual observations, our method learns reliable relative size estimates, with no explicit human supervision. We introduce the relative size dataset and show that our method outperforms competitive textual and visual baselines in reasoning about size comparisons.

CMDragons 2015 is the champion of the RoboCup Small Size League of autonomous robot soccer. The team won all of its six games, scoring a total of 48 goals and conceding 0. This unprecedented dominant performance is the result of various features, but we particularly credit our novel offense multi-robot coordination. This paper thus presents our Selectively Reactive Coordination (SRC) algorithm, consisting of two layers: A coordinated opponent-agnostic layer enables the team to create its own plans, setting the pace of the game in offense. An individual opponent-reactive action selection layer enables the robots to maintain reactivity to different opponents. We demonstrate the effectiveness of our coordination through results from RoboCup 2015, and through controlled experiments using a physics-based simulator and an automated referee.

Restarts are an important technique to make search more robust. This paper is concerned with how to maintain and propagate nogoods recorded from restarts efficiently. It builds on reduced nld-nogoods introduced for restarts and increasing nogoods introduced for symmetry breaking. The paper shows that reduced nld-nogoods extracted from a single restart are in fact increasing, which can thus benefit from the efficient propagation algorithm of the incNGs global constraint. We present a lighter weight filtering algorithm for incNGs in the context of restart-based search using dynamic event sets (dynamic subscriptions). We show formally that the lightweight version enforces GAC on each nogood while reducing the number of subscribed decisions. The paper also introduces an efficient approximation to nogood minimization such that all shortened reduced nld-nogoods from the same restart are also increasing and can be propagated with the new filtering algorithm. Experimental results confirm that our lightweight filtering algorithm and approximated nogood minimization successfully trade a slight loss in pruning for considerably better efficiency, and hence compare favorably against existing state-of-the-art techniques.

We present MM, the first bidirectional heuristic search algorithm whose forward and backward searches are guaranteed to ''meet in the middle'', i.e. never expand a node beyond the solution midpoint. We also present a novel framework for comparing MM, A*, and brute-force search, and identify conditions favoring each algorithm. Finally, we present experimental results that support our theoretical analysis.

The efficient extraction of one maximal information subset that does not conflict with multiple contxts or additional information sources is a key basic issue in many A.I. domains, especially when these contexts or sources can be mutually conflicting. In this paper, this question is addressed from a computational point of view in clausal Boolean logic. A new approach is introduced that experimentally outperforms the currently most efficient technique.