Stavros Vassos

Planning with a task modeling framework in manufacturing robotics

In this paper we present the idea that by using AI planning in concert with formal task modeling, the overhead associated with plan creation for complex tasks can be reduced. The proposed approach uses a SysML taxonomy to model the system capabilities and the process specification, and the PDDL planning language to determine acceptable objective solutions. This idea is applied to the manufacturing domain, and examples are shown modeling a multi-robot system in an automobile manufacturing environment. A discussion is given regarding the merits of the demonstrated approach.

Building Virtual Earth Observatories Using Ontologies, Linked Geospatial Data and Knowledge Discovery Algorithms

Advances in remote sensing technologies have allowed us to send an ever-increasing number of satellites in orbit around Earth. As a result, satellite image archives have been constantly increasing in size in the last few years (now reaching petabyte sizes), and have become a valuable source of information for many science and application domains (environment, oceanography, geology, archaeology, security, etc.). TELEIOS is a recent European project that addresses the need for scalable access to petabytes of Earth Observation data and the discovery of knowledge that can be used in applications. To achieve this, TELEIOS builds on scientific databases, linked geospatial data, ontologies and techniques for discovering knowledge from satellite images and auxiliary data sets. In this paper we outline the vision of TELEIOS (now in its second year), and give details of its original contributions on knowledge discovery from satellite images and auxiliary datasets, ontologies, and linked geospatial data.

How to progress a database III

In a seminal paper, Lin and Reiter introduced a model-theoretic definition for the progression of a basic action theory in the situation calculus, and proved that it implies the intended properties. They also showed that this definition comes with a strong negative result, namely that for certain cases first-order logic is not expressive enough to correctly characterize the progressed theory and second-order axioms are necessary. However, they also considered an alternative simpler definition according to which the progressed theory is always first-order definable. They conjectured that this alternative definition is incorrect in the sense that the progressed theory is too weak and may sometimes lose information. This conjecture and the status of the definability of progression in first-order logic has remained open since. In this paper we present two significant results about this alternative definition of progression. First, we prove the Lin and Reiter conjecture by presenting a case where the progressed theory indeed does lose information, thus closing a question that has remained open for more than ten years. Second, we prove that the alternative definition is nonetheless correct for reasoning about a large class of sentences, including some that quantify over situations.

Building virtual earth observatories using ontologies and linked geospatial data

Advances in remote sensing technologies have enabled public and commercial organizations to send an ever-increasing number of satellites in orbit around Earth. As a result, Earth Observation (EO) data has been constantly increasing in volume in the last few years, and is currently reaching petabytes in many satellite archives. For example, the multi-mission data archive of the TELEIOS partner German Aerospace Center (DLR) is expected to reach 2PB next year, while ESA estimates that it will be archiving 20PB of data before the year 2020. As the volume of data in satellite archives has been increasing, so have the scientific and commercial applications of EO data. Nevertheless, it is estimated that up to 95% of the data present in existing archives has never been accessed, so the potential for increasing exploitation is very big.

Art-Bots: Toward Chat-Based Conversational Experiences in Museums

In this work we explore how a widely used conversational interface can be employed to offer engaging experiences in museums. “Art-bots” interact with visitors through chat and convey information about the museum artifacts in the form of short stories. The wide adoption of chat platforms such as Facebook messenger offers new ground to revisit approaches on avatars and virtual guides and build interactive dialogues to engage visitors. We take a practical approach based on a set of responses that are triggered by certain keywords and a curated story that guides visitors to a gamified information hunt.

Synthesizing daily life logs through gaming and simulation

In the recent years there has been a growing interest in the design and implementation of smart homes, and smart buildings in general. The evaluation of approaches in this area typically requires massive datasets of measurements from deployed sensors in real prototypes. While a few datasets obtained by real smart homes are freely available, they are not sufficient for comparing different approaches and techniques in a variety of configurations. In this work, we propose a smart home dataset generation strategy based on a simulated environment populated with virtual autonomous agents, sensors and devices which allow to customize and reproduce a smart space using a series of useful parameters. The simulation is based on declarative process models for modeling habits performed by agents, an action theory for realizing low-level atomic actions, and a 3D virtual execution environment. We show how different configurations generate a variety of sensory logs that can be used as input to a state-of-the-art activity recognition technique in order to evaluate its performance under parametrized scenarios, as well as provide guidelines for actually building real smart homes.

Touching Notes: A Gesture-Based Game for Teaching Music to Children

In this work we present some findings on how gesture-based interfaces can stimulate and motivate children into learning the basics of music notation. We report on a serious game we developed that relies on a hand detection interface using motion-sensing camera technology. Our initial experiments indicate the accuracy, effectiveness, and user acceptance of the gesture-based interface among children, thereby validating the idea that such interfaces can be used to facilitate the teaching of music concepts.

Controlling Logistics Robots with the Action-Based Language YAGI

To achieve any meaningful tasks, a robot needs some form of task-level executive which acquires knowledge, reasons or plans, and performs and monitors actions. A formal approach for such agent programming is the Golog agent programming language. Golog is based on a first-order logic representation, and a drawback of common implementations is that in order to program agents, also knowledge of Prolog functionality is typically needed. In this paper, we present a prototype implementation of YAGI, a language rooted in Golog that offers a practical subset of the rich Golog framework in a more familiar syntax. Bridging imperative-style programming with an action-based specification, YAGI is more accessible to developers and provides a better ground for robot task-level executives. Moreover, we developed bindings for popular robotics frameworks such as ROS and Fawkes. As a proof of concept we present a YAGI-based agent for the RoboCup Logistics League which shows the expressiveness and the possibility to easily embed YAGI into robot applications.

Accessing Government Open Data Through Chatbots

In this paper, we propose to employ chatbots as an interface for open data published by organizations, specifically focusing on public administrations. Open data are especially useful in e-Government initiatives but their exploitation is currently hampered to end users by the lack of user-friendly access methods. On the other hand, current UX in social networks have made people used to chatting. Building on cognitive technologies, we prototyped a chatbot on top of the OpenCantieri dataset published by the Italian Ministero delle Infrastrutture e Trasporti, and we argue that such a model can be extended as a generally available access method to open data.

Collecting Human Habit Datasets for Smart Spaces Through Gamification and Crowdsourcing

A lot of research in the last years has focused on smart spaces, covering aspects related to ambient intelligence, activity monitoring and mining, etc. All these efforts require datasets to be used for experimental purposes and as benchmarks for novel techniques. Such datasets are today difficult to obtain as, on the one hand, building smart facilities is expensive, requiring considerable costs for maintenance and extension, and, on the other hand, freely available datasets are scarce, not continuously updated and contain a limited set of sensors, thus not allowing the evaluation of algorithms that require the availability of specific categories of sensors. To this aim, we have built a prototype smart virtual environment producing sensor logs on the basis of activities performed by users as if they were really acting in a physical smart space.