Pavel Smrž

Indexing and search methods for spoken documents

This paper presents two approaches to spoken document retrieval—search in LVCSR recognition lattices and in phoneme lattices For the former one, an efficient method of indexing and search of multi-word queries is discussed In phonetic search, the indexation of tri-phoneme sequences is investigated The results in terms of response time to single and multi-word queries are evaluated on ICSI meeting database.

Design of the Human-Robot Interaction for a Semi-Autonomous Service Robot to Assist Elderly People

Service robots could support elderly people’s activities of daily living and enable them to live in their own residences independently as long as possible. Current robot technology does not allow reliable fully autonomous operation of service robots with manipulation capabilities in the heterogeneous environments of private homes. We developed and evaluated a usage concept for semi-autonomous robot control as well as user interfaces for three user groups. Elderly people are provided with simple access to autonomous robot services through a handheld device. In case of problems with autonomous execution the robot contacts informal caregivers (e.g. relatives) who can support the robot using semi-autonomous teleoperation. To solve more complex problems, professional teleoperators are contacted who have extended remote access.

Integrating ontologies into learning management systems: A case of Czech

The paper deals with the role of ontologies in e-learning. We report our research on the application of ontologies in intelligent platforms for self-learning and computer-mediated education that are developed at the Faculty of Informatics, Masaryk University, Brno, Czech Republic. The use of ontologies in the organization of knowledge about the area of learning management systems is tackled first. The next part discusses a new direction of our research aiming at automatic generation of e-learning self-tests based on information extracted from ontologies. A very new topic deals with the preliminary experiments on an ontology-based automatic evaluation of student essays.

A Pattern-Based Framework for Uncertainty Representation in Ontologies

We present a novel approach to representing uncertain information in ontologies based on design patterns. We provide a brief description of our approach, present its use in connection with fuzzy information and probabilistic information, and describe the possibility to model multiple types of uncertainty in a single ontology. We also shortly present an appropriate fuzzy reasoning tool and define a complex ontology architecture for well-founded handling of uncertain information.

Semi-Autonomous Domestic Service Robots: Evaluation of a User Interface for Remote Manipulation and Navigation With Focus on Effects of Stereoscopic Display

In this article, we evaluate a novel type of user interface for remotely resolving challenging situations for service robots in domestic environments. Our focus is on potential advantages of stereoscopic display. The user interface is based on a control architecture that allows involvement of a remote human operator when the robot encounters a problem. It offers semi-autonomous remote manipulation and navigation with low-cost interaction devices, incorporates global 3D environment mapping, and follows an ecological visualization approach that integrates 2D laser data, 3D depth camera data, RGB data, a robot model, constantly updated global 2D and 3D environment maps, and indicators into a single 3D scene with user-adjustable viewpoints and optional viewpoint-based control. We carried out an experiment with 28 participants in a home-like environment investigating the utility of stereoscopic display for three types of task: defining the shape of an unknown or unrecognized object to be grasped, positioning the gripper for semi-autonomous reaching and grasping, and navigating the robot around obstacles. Participants were able to successfully complete all tasks and highly approved the user interface in both monoscopic and stereoscopic display modes. They were significantly faster under stereoscopic display in positioning the gripper. For the other two task types, there was a tendency for faster task completion in stereo mode that would need to be verified in further studies. We did not find significant differences in perceived workload between display types for any type of task. We conclude that stereoscopic display seems to be a useful optional display mode for this type of user interface but that its utility may vary depending on the task.

Best analysis selection in inflectional languages

Ambiguity is the fundamental property of natural language. Perhaps, the most burdensome case of ambiguity manifests itself on the syntactic level of analysis. In order to face up to the high number of obtained derivation trees, this paper describes several techniques for evaluation of the figures of merit, which define a sort order on parsing trees. The presented methods are based on language specific features of synthetical languages and they improve the results of simple stochastic approaches.

Annotating images with suggestions: user study of a tagging system

This paper explores the concept of image-wise tagging. It introduces a web-based user interface for image annotation, and a novel method for modeling dependencies of tags using Restricted Boltzmann Machines which is able to suggest probable tags for an image based on previously assigned tags. According to our user study, our tag suggestion methods improve both user experience and annotation speed. Our results demonstrate that large datasets with semantic labels (such as in TRECVID Semantic Indexing) can be annotated much more efficiently with the proposed approach than with current class-domain-wise methods, and produce higher quality data.

Ontology acquisition for automatic building of scientific portals

Ontologies are commonly considered as one of the essential parts of the Semantic Web vision, providing a theoretical basis and implementation framework for conceptual integration and information sharing among various domains. In this paper, we present the main principles of a new ontology acquisition framework applied for semi-automatic generation of scientific portals. Extracted ontological relations play a crucial role in the structuring of the information at the portal pages, automatic classification of the presented documents as well as for personalisation at the presentation level.

Integrating natural language processing into e-learning: a case of Czech

The paper deals with the application of NLP technology in e-learning. We report our research on intelligent platforms for computer-mediated education. Some of the methods described in the paper have already taken part in the end-user applications that are in everyday use, others still wait for their implementation in the form of software products. The main message of the paper is that the language technology, even in the imperfect form of the current state of the art, can significantly enhance todays computer-mediated teaching and learning activities. It is true especially for languages different from English, where the adopted learning management systems often do not support even the basic functionality of a language-oriented search and retrieval of learning objects. As a case study, this paper demonstrates the application of the given ideas for e-learning materials in Czech.

Text Corpus with Errors

This paper presents a description of a Czech text corpus (Chyby) containing various kinds of errors such as spelling, typographical, grammatical, style, lexical. We explain how Chyby has been built, how the errors in it have been discovered, marked and annotated. The classification of the errors is presented and the statistics concerning the types of errors is given. The tools for annotating the errors are also described. To the best of our knowledge, this is first text corpus of this sort prepared for Czech.