Fluent and safe interactions of humans and robots require both partners to anticipate the others' actions. The bottleneck of most methods is the lack of an accurate model of natural human motion. In this work, we present a conditional variational autoencoder that is trained to predict a window of future human motion given a window of past frames. Using skeletal data obtained from RGB depth images, we show how this unsupervised approach can be used for online motion prediction for up to 1660 ms. Additionally, we demonstrate online target prediction within the first 300-500 ms after motion onset without the use of target specific training data. The advantage of our probabilistic approach is the possibility to draw samples of possible future motion patterns. Finally, we investigate how movements and kinematic cues are represented on the learned low dimensional manifold.

Dissimilarity measures are often used as a proxy or a handle to reason about data. This can be problematic, as the data representation is often a consequence of the capturing process or how the data is visualized, rather than a reflection of the semantics that we want to extract. Facial expressions are a subtle and essential part of human communication but they are challenging to extract from current representations. In this paper we present a method that is capable of learning semantic representations of faces in a data driven manner. Our approach uses sparse human supervision which our method grounds in the data. We provide experimental justification of our approach showing that our representation improves the performance for emotion classification.

Achilles Tendon Rupture (ATR) is one of the typical soft tissue injuries. Accurately predicting the rehabilitation outcome of ATR using noisy measurements with missing entries is crucial for treatment decision support. In this work, we design a probabilistic model that simultaneously predicts the missing measurements and the rehabilitation outcome in an end-to-end manner. We evaluate our model and compare it with multiple baselines including multi-stage methods using an ATR clinical cohort. Experimental results demonstrate the superiority of our model for ATR rehabilitation outcome prediction.

This paper presents a unique dataset consisting of 20 recordings of the same musical piece, conducted with 4 different musical intentions in mind. The upper body and baton motion of a professional conductor was recorded, as well as the sound of each instrument in a professional string quartet following the conductor. The dataset is made available for benchmarking of motion recognition algorithms. An HMM-based emotion intent classification method is trained with subsets of the data, and classification of other subsets of the data show firstly that the motion of the baton communicates energetic intention to a high degree, secondly, that the conductor’s torso, head and other arm conveys calm intention to a high degree, and thirdly, that positive vs negative sentiments are communicated to a high degree through other channels than the body and baton motion – most probably, through facial expression and muscle tension conveyed through articulated hand and finger motion. The long-term goal of this work is to develop a computer model of the entire conductor-orchestra communication pro- cess; the studies presented here indicate that computer modeling of the conductor-orchestra communication is feasible.

We study a mini-batch diversification scheme for stochastic gradient descent (SGD). While classical SGD relies on uniformly sampling data points to form a mini-batch, we propose a non-uniform sampling scheme based on the Determinantal Point Process (DPP). The DPP relies on a similarity measure between data points and gives low probabilities to mini-batches which contain redundant data, and higher probabilities to mini-batches with more diverse data. This simultaneously balances the data and leads to stochastic gradients with lower variance. We term this approach Diversified Mini-Batch SGD (DM-SGD). We show that regular SGD and a biased version of stratified sampling emerge as special cases. Furthermore, DM-SGD generalizes stratified sampling to cases where no discrete features exist to bin the data into groups. We show experimentally that our method results more interpretable and diverse features in unsupervised setups, and in better classification accuracies in supervised setups.

Mutual gaze is a powerful cue for communicating social attention and intention. A plethora of studies have demonstrated the fundamental roles of mutual gaze in establishing communicative links between humans, and enabling non-verbal communication of social attention and intention. The amount of mutual gaze between two partners regulates human-human interaction and is a sign of social engagement. This paper investigates whether implementing mutual gaze in robotic systems can achieve social effects, thus to improve human robot interaction. Based on insights from existing human face-to-face interaction studies, we implemented an interactive mutual gaze model in an embodied agent, the social robot head Furhat. We evaluated the mutual gaze prototype with 24 participants in three applications. Our results show that our mutual gaze model improves social connectedness between robots and users.

Exploring and modeling heterogeneous elastic surfaces requires multiple interactions with the environment and a complex selection of physical material parameters. The most common approaches model deformable properties from sets of offline observations using computationally expensive force-based simulators. In this work we present an online probabilistic framework for autonomous estimation of a deformability distribution map of heterogeneous elastic surfaces from few physical interactions. The method takes advantage of Gaussian Processes for constructing a model of the environment geometry surrounding a robot. A fast Position-based Dynamics simulator uses focused environmental observations in order to model the elastic behavior of portions of the environment. Gaussian Process Regression maps the local deformability on the whole environment in order to generate a deformability distribution map. We show experimental results using a PrimeSense camera, a Kinova Jaco2 robotic arm and an Optoforce sensor on different deformable surfaces.

Object tracking is a fundamental ability for a robot; manipulation as well as activity recognition relies on the robot being able to follow objects in the scene. This paper presents a tracker that adapts to changes in object appearance and is able to re-discover an object that was lost. At its core is a keypoint-based method that exploits the rigidity assumption: pairs of keypoints maintain the same relations over similarity transforms. Using a structured approach to learning, it is able to incorporate new appearances in its model for increased robustness. We show through quantitative and qualitative experiments the benefits of the proposed approach compared to the state of the art, even for objects that do not strictly follow the rigidity assumption.

This work employs an adaptive learning mechanism to perform tracking of an unknown object through RGBD cameras. We extend our previous framework to robustly track a wider range of arbitrarily shaped objects by adapting the model to the measured object size. The size is estimated as the object undergoes motion, which is done by fitting an inscribed cuboid to the measurements. The region spanned by this cuboid is used during tracking, to determine whether or not new measurements should be added to the object model. In our experiments we test our tracker with a set of objects of arbitrary shape and we show the benefit of the proposed model due to its ability to adapt to the object shape which leads to more robust tracking results.