The most critical moves that determine the success of a manipulation task are performed within the close vicinities of the object prior to grasping, and the goal prior to the final placement. Memorizing these state-action sequences and reusing them can dramatically improve the task efficiency, whereas even the state-of-the-art planning algorithms may require significant amount of time and computational resources to generate a solution from scratch depending on the complexity and the constraints of the task. In this paper, we propose a hybrid approach that combines the reliability of the past experiences gained through demonstration and the flexibility of a generative motion planning algorithm, namely RRT*, to improve task execution efficiency. As a side benefit of reusing these final moves, we can dramatically reduce the number of nodes used by the generative planner, hence the planning time, by either early-terminating the planner when the generated plan reaches a "recalled state", or deliberately biasing it towards the memorized state-action sequences that are feasible at the moment. This complementary combination of the already available partial plans and the generated ones yield to fast, reliable, and repeatable solutions.

General representation, abstraction and exchange definitions are crucial for dynamically configurable context recognition. However, to evaluate potential definitions, suitable standard datasets are needed. This paper presents our effort to create and maintain large scale, multimodal standard datasets for context recognition research. We ourselves used these datasets in previous research to deal with placement effects and presented low-level sensor abstractions in motion based on-body sensing. Researchers, conducting novel data collections, can rely on the toolchain and the the low-level sensor abstractions summarized in this paper. Additionally, they can draw from our experiences developing and conducting context recognition experiments. Our toolchain is already a valuable rapid prototyping tool. Still, we plan to extend it to crowd-based sensing, enabling the general public to gather context data, learn more about their lives and contribute to context recognition research. Applying higher level context reasoning on the gathered context data is a obvious extension to our work.

Pervasive, context-aware computing technologies are essential drivers of next generation applications and appliances that will profoundly impact the way we work and play, conduct research, impart education, govern ourselves and care for our health. In order to support context-aware applications and achieve multidevice interoperability, it is important to develop an effective framework and representation language(s) for capturing and representing activity and context information, reasoning about the information and moving such information across devices in a secure and efficient manner. Our intent was as follows: First, discuss and review existing and novel Activity Context Representation and Exchange Languages. Discuss results from creation of solution architectures and proposals for languages, data structures, operations to enable top use-case categories. Second, discuss papers and proposals for new research areas and review work building on key research themes with specific opportunities for collaborative work in the next two-three years in this academically and commercially important area, with topics including, but not limited to semantic computing, task modeling, context representation, and activity recognition.

A cognitive robot should construct a plan to attain its goals. While it executes the actions in its plan, it may face several failures due to both internal and external issues. We present a taxonomy to classify these failures that may be encountered during the execution of cognitive tasks. The taxonomy presents a wide range of failure types. To recover from most of these failures presented in this taxonomy, we propose a Robust Planning Framework for cognitive robots. Our framework combines planning, reasoning and learning procedures into each other for robust execution of cognitive tasks. Failures can be detected and handled by reasoning and replanning, respectively. The framework also facilitates learning new hypotheses incrementally based on experience. It can successfully detect and recover from temporary failures on a selected set of actions executed by a Pioneer3DX robot. It has been shown that our preliminary results for hypothesis learning in failure scenarios are promising.

In this study, we propose a method for learning interactions among different types of objects to devise new plans using these objects. Learning is accomplished by observing a given sequence of events with their timestamps and using spatial information on the initial state of the objects in the environment. We assume that no intermediate state information is available about the states of objects. We have used the Incredible Machine game as a suitable domain for analyzing and learning object interactions. When a knowledge base about relations among objects is provided, interactions to devise new plans are learned to a desired extent. Moreover, using spatial information of objects or temporal information of events makes it feasible to learn the conditional effects of objects on each other. Our analyses show that, integrating spatial and temporal data in a spatio-temporal learning approach gives closer results to that of the knowledge-based approach by providing applicable event models for planning. This is promising because gathering spatio-temporal information does not require great amount of knowledge.

Crowdsourcing services are often used to collect a large amount of labeled data for machine learning. Although they provide us an easy way to get labels at very low cost in a short period, they have serious limitations. One of them is the variable quality of the crowd-generated data. There have been many attempts to increase the reliability of crowd-generated data and the quality of classifiers obtained from such data. However, in these problem settings, relatively few researchers have tried using expert-generated data to achieve further improvements. In this paper, we extend three models that deal with the problem of learning from crowds to utilize ground truths: a latent class model, a personal classifier model, and a data-dependent error model. We evaluate the proposed methods against two baseline methods on a real data set to demonstrate the effectiveness of combining crowd-generated data and expert-generated data.

This paper outlines a study that investigates spatial language for use in human-robot communication. The scenario studied is a home setting in which the elderly resident has misplaced an object, such as eyeglasses, and the robot will help the resident find the object. We present results from phase I of the study in which we investigate spatial language generated to a human addressee or a robot addressee in a virtual environment and highlight differences between younger and older adults. Drawn from these results, a discussion is included of needed robot capabilities, such as an approach that addresses varying perspectives used and recognition of furniture items for use as spatial references.

We have explored with three notions: conceptualization, and contextualization from situated cognition, and psychic reflection from activity theory for identifying activities into a method called the activity states framework (ASF). The purpose of our work is to build an AI system based on ASF for the identification of collaborators activities during situated context, e.g., collaborators are engaged in a tutorial activity. In this paper, we will introduce and propose how Web-mediated collaborative activities can be identified from collaborators communication exchanges by applying the ASF.

In this paper, we present a simplified shallow semantic parsing approach to extracting opinion targets. This is done by formulating opinion target extraction (OTE) as a shallow semantic parsing problem with the opinion expression as the predicate and the corresponding targets as its arguments. In principle, our parsing approach to OTE differs from the state-of-the-art sequence labeling one in two aspects. First, we model OTE from parse tree level, where abundant structured syntactic information is available for use, instead of word sequence level, where only lexical information is available. Second, we focus on determining whether a constituent, rather than a word, is an opinion target or not, via a simplified shallow semantic parsing framework. Evaluation on two datasets shows that structured syntactic information plays a critical role in capturing the domination relationship between an opinion expression and its targets. It also shows that our parsing approach much outperforms the state-of-the-art sequence labeling one.

Social events are events that occur between people where at least one person is aware of the other and of the event taking place. Extracting social events can play an important role in a wide range of applications, such as the construction of social network. In this paper, we introduce the task of social event extraction for tweets, an important source of fresh events. One main challenge is the lack of information in a single tweet, which is rooted in the short and noise-prone nature of tweets. We propose to collectively extract social events from multiple similar tweets using a novel factor graph, to harvest the redundance in tweets, i.e., the repeated occurrences of a social event in several tweets. We evaluate our method on a human annotated data set, and show that it outperforms all baselines, with an absolute gain of 21% in F1.