The contribution of this paper is twofold. First, we present a new method for collective intention recognition based on mainstream philosophical accounts. Second, we extend our previous Elder Care system with collective intention recognition ability for assisting a couple of elderly people. The previous system was just capable of individual intention recognition, and so it has now been enabled to deal with situations where the elders intend to do things together.

To assist scientists in data analysis tasks, we have developed semantic workflow representations that support automatic constraint propagation and reasoning algorithms to manage constraints among the individual workflow steps. Semantic constraints can be used to represent requirements of input datasets as well as best practices for the method represented in a workflow. We demonstrate how the Wings workflow system uses semantic workflows to assist users in creating workflows while validating that the workflows comply with the requirements of the software components and datasets. Wings reasons over semantic workflow representations that consist of both a traditional dataflow graph as well as a network of constraints on the data and components of the workflow.

We propose a learning framework, which is based on diffusionmethodology, that performs data fusion and anomalydetection in multi-dimensional time series data. Real lifeapplications and processes usually contain a large numberof sensors that generate parameters (features), where eachsensor collects partial information about the running process.These input sensors are fused to describe the behaviorof the whole process. The proposed data fusing algorithmis done in an hierarchial fashion: first it re-scales the inputsensors. Then, the re-formulated inputs are fused togetherby the application of the diffusion maps to reveal the nonlinearrelationships among them. This process constructsby embedding a low-dimensional description of the system.The embedding separates between sensors (parameters) thatcause stable and instable behavior of the system.This unsupervised algorithm first studies the system’sprofile from a training dataset by reducing its dimensions.Then, the coordinates of newly arrived data points are determinedby the application of multi-scale Gaussian approximation.To achieve this, an hierarchial processing of theincoming data is introduced.

Compressive spectral clustering combines the distance preserving measurements of compressed sensing with the power of spectral clustering. Our analysis provides rigorous bounds on how small errors in the affinity matrix can affect the spectral coordinates and clusterability. This work generalizes the current perturbation results of two-class spectral clustering to incorporate multiclass clustering using k eigenvectors.

The emergence of low-cost sensing architectures for diverse modalities has made it possible to deploy sensor networks that acquire large amounts of very high-dimensional data. To cope with such a data deluge, manifold models are often developed that provide a powerful theoretical and algorithmic framework for capturing the intrinsic structure of data governed by a low-dimensional set of parameters. However, these models do not typically take into account dependencies among multiple sensors. We thus propose a new joint manifold framework for data ensembles that exploits such dependencies. We show that joint manifold structure can lead to improved performance for manifold learning. Additionally, we leverage recent results concerning random projections of manifolds to formulate a universal, network-scalable dimensionality reduction scheme that efficiently fuses the data from all sensors.

In this paper, we present a method for isometric correction of manifold learning techniques. We first present an isometric nonlinear dimension reduction method. Our proposed method overcomes the issues associated with well-known isometric embedding techniques such as ISOMAP and maximum variance unfolding (MVU), i.e., computational complexity and the geodesic convexity requirement. Based on the proposed algorithm, we derive our isometric correction method. Our approach follows an isometric solution to the problem of local tangent space alignment. We provide a derivation of a fast iterative solution. The performance of our algorithm is illustrated on both synthetic and real datasets compared to other methods.

Both Lawrence Carin tools utilize the same transformed Robust PCA model for the visual data: D A E, and use practically the same Hierarchical Bayesian methods are employed to learn a reversible algorithm for extracting the low-rank structures A from the statistical embedding. The proposed embedding visual data D, despite image domain transformation T and procedure is connected to spectral embedding methods (e.g., corruptions E. We will show how these two seemingly simple diffusion maps and Isomap), yielding a new statistical spectral tools can help unleash tremendous information in images framework. The proposed approach allows one to discard and videos that we used to struggle to get. We believe these the training data when embedding new data, allows synthesis new tools will bring disruptive changes to many challenging of high-dimensional data from the embedding space, tasks in computer vision and image processing, including and provides accurate estimation of the latent-space dimensionality.

Researchers in many fields such as machine learning, computer vision, bioinformatics and robotics often observe that high dimensional data samples have low degrees of freedom in local neighborhoods, but a more complicated global structure. In many cases, there is enough structure in the data so the degrees of freedom can be described by a lower dimensional object such as a manifold. The goal of manifold learning research is to discover techniques that exploit local structure in data to learn better models, learn better input-output relationships and reduce the computational complexity of learning. The field of manifold learning is truly cross-disciplinary, involving researchers from such varied fields as topology, geometry, machine learning, statistics, computer vision, robotics and many others. This has led to an accelerating pace of research and applications in recent years.

Researchers can use simulators Figure 1: The Siros client/server architecture of the Konbini to collect data to build and evaluate interaction models at system. the same time as core components of the real-world robot are built and integrated. Once the real robot becomes robust enough, the models trained on simulators can be applied for Clients are in charge of rendering a given view of the virtual further experiments.

In order to design robots suited to engage in cooperative manipulation tasks with humans, we study human-human teams as they work together to move a heavy object across a room. We are interested in several questions. First, do two person, gross motion tasks follow the same sinusoidal pattern, one person fine motion tasks do? Does performance improve when audio or visual communication is permitted? How does performance correlate with an individual's perception of performance? Non-physiological, or performance based, studies of human-human cooperation can be divided into two categories: Haptic and Non-Haptic (audio, visual, etc). The first category, involves physical interaction through the object being manipulated via force, pressure, and tactile sensations (Jones and Sarter 2008), (Reed and Peshkin 2008). Most of the non-haptic experiments are virtual setups where individuals are moving an object together on a computer screen via two controllers (Basdogan, Ho, and Srinivasan 2000), (Sallnas and Zhai 2003). A survey on the role of communication between people appears in (Whitaker, 2003). The novelty of our work is to investigate non-haptic communication in haptic manipulation tasks.