News
Senior Best Presentation Award at COSIT
Alex Klippel and Rui Li were received the senior best presentation award that the Conference on Spatial Information Theory COSIT. The presentation is based on the paper:
Klippel, A., & Li, R. (2009). The endpoint hypothesis: A topological-cognitive assessment of geographic scale movement patterns. In K. Stewart Hornsby, C. Claramunt, M. Denis, & G. Ligozat (Eds.), Spatial Information Theory, 9th International Conference, COSIT 2009, Aber Wrac'h, France, September 21-25, 2009 Proceedings (pp. 177–194). Berlin: Springer.
The work is part of Dr. Klippel's NSF project on Contextual influences on the category construction of geographic-scale movement patterns.
The paper can be downloaded here or directly from Springer. More information can be found on the project website.
Paper accepted at workshop: Studying Moving Objects in a three-dimensional world, 3rd Workshop on Behaviour Monitoring and Interpretation, BMI'09.
Title: A Chorematic Approach to Characterizing Movement Patterns
Abstract: We adopt a perspective of characterizing movement patterns on the
basis of conceptual primitives that we call movement choremes: MC. This theory is
an extension of our existing work on wayfinding choremes that specifically
addressed movement patterns important for wayfinding and route directions. Just
like in our previous work the goal is to develop a formal language that allows for
characterizing the movement of individual agents and entities from a cognitively
unifying perspective. By this we mean that while our main work concentrates on
the conceptual level of movement patterns, the framework is intended to
incorporate externalizations such as natural language and graphics (sketches) and
also formal theories of qualitative movement and spatial relation characterizations.
We discuss our approach in relation to existing frameworks such as RCC and the
9-intersection formalism to ground the potential of a formal-spatial language
approach.
NSF awared grant to Alex Klippel and Luke Zhang
Title: Spatial awareness through sapient interfaces
NSF awarded grant to Alex Klippel
Title: Contextual influences on the category construction of geographic-scale movement patterns
Official NSF abstract:
The world we live in is dynamic on all scales. On a small scale, think of walking across a room or driving to the grocery store; on a large scale, imagine a hurricane crossing a state. Research that addresses how humans understand relationships between space and time is, therefore, central not only to geography, but also to cognitive and information sciences. The objectives of this project are two-fold: to develop a research framework for examining how movement patterns at the geographic scale (MPGS) are understood, and to evaluate how formalisms used in geographic information science are able to capture how people understand MPGS. Linking cognitive and formal characterizations enables models that align with how people think about large scale spatial processes, thereby enhancing communication at the interface between humans and computers.
Dr. Alexander Klippel at the Pennsylvania State University will conduct a set of experiments that are based on a grouping paradigm that is used to elicit conceptual knowledge (i.e., categorization). Participants view animations of large-scale spatial phenomena ending at different spatio-temporal stages. For example, participants will look at 63 different animations of a hurricane approaching shore, making landfall, and finally moving far inland. Participants will then classify these animations based on their similarities. The question these studies are attempting to answer is whether formal, topologically equivalent characterizations across different domains (e.g., whether the moving entity is a hurricane or a glacier) are also equivalent cognitively. In other words, how does the semantics of dynamic features influence the cognitive conceptualizations thereof? Does semantics change the cognitive salience of individual topological relations? Additionally, the project will address the geographically critical question of scale effects and a contribution will be made to the formal underpinning of geographic event language by relating the conceptualizations of movement patterns to linguistic externalizations. To accomplish these goals, the PI has developed software to both collect and analyze behavioral data efficiently. The software to analyze the data will be developed within existing geovisual analytics software developed at the GeoVISTA Center in the Geography Department at the Pennsylvania State University.
This project will contribute to the understanding of how humans conceptualize movement patterns at the geographic scale, such as hurricanes moving across a peninsula, from the perspective of how these movement patterns are characterized formally. Connecting formal and cognitive characterizations of spatio-temporal information is essential to develop efficient human-computer interfaces and models of spatial cognition. This project will develop software solutions for the design of experiments and the efficient analysis of the collected data using visual analytics approaches. As cognitive conceptualizations are core to research on ontologies and categorization, and also to fields such as anthropology, we expect both the developed research framework and the software solutions to be beneficial to a large research community.
Alex Klippel gave the first talk in the coffee hour lecture series Fall 2009
Alex inaugurated the Fall Coffee Hour Lecture Series of the Department of Geography with a talk titled:
Toward a decade of the spatial mind: contributions and visions
The talk has been recorded and can be accessed via the coffee hour website (some technical problems occur in the first 5 minutes)
Submission accepted at WEBDB 2009
Title: Extracting route directions from webpages.
Zhang, Xiao; Mitra, Prasenjit; Jaiswal, Anuj; Xu, Sen; MacEachren, A. M.; Klippel, A.
Abstract: Linguists and geographers are more and more interested in route direction documents because they
contain interesting motion descriptions and language
patterns. A large number of such documents
can be easily found on the Internet. A challenging
task is to automatically extract meaningful route
parts, i.e. destinations, origins and instructions,
from route direction documents. However, no work
exists on this issue. In this paper, we introduce our
effort toward this goal. Based on our observation
that sentences are the basic units for route parts, we
propose a new approach which utilizes both HTML
tag information and natural language knowledge to
delimit sentences in HTML documents. Additionally,
we study the sentence classication problem in
route direction documents and its sequential nature.
Several machine learning methods are compared
and analyzed. The impacts of different sets of features
are studied. Based on the obtained insights,
we propose to use sequence labelling models such
as CRFs and MEMMs and they yield a high accuracy
in route part extraction. The approach is evaluated
on over 10,000 hand-tagged sentences in 100
documents. The experimental results show the effectiveness
of our method. The above techniques
have been implemented and published as the rst
module of the GeoCAM1 system, which will also
be briey introduced in this paper.
Submission accepted at COSIT (Conference on Spatial Information Theory)
Title: The endpoint hypothesis: A topological-cognitive assessment of geographic scale movement patterns
Alexander Klippel & Rui Li
Abstract: Movement patterns of individual entities at the geographic scale are becoming a prominent research focus in spatial sciences. From the perspective of research at the interface of spatial and cognitive sciences the question arises how cognitive and formal characterizations of movement patterns relate. In other words, are (mostly qualitative) formal characterizations cognitively adequate? This article experimentally evaluates movement patterns that can be characterized as paths through a conceptual neighborhood graph, that is, two extended spatial entities are changing their topological relationship gradually. The central questions addressed are: (a) Do humans naturally use topology to create cognitive equivalent classes, that is, is topology the basis for categorizing movement patterns spatially? (b) Are ‘all’ topological relations equally salient, and (c) does language influence categorization. The first two questions are addressed using the endpoint hypothesis that has been formulated by Regier and Zhang (2007). To this end, movement patterns are distinguished by the topological relation they end in. The third question addresses the question whether language has an influence on the classification of movement patterns, that is, whether there is a difference between linguistic and non-linguistic category construction. In contrast to our previous research we were able to document the importance of topology for conceptualizing movement patterns but also reveal differences in the cognitive saliency of topological relations. The latter aspect calls for a weighted conceptual neighborhood graph to cognitively adequately model human conceptualization processes.
Submission accepted at SNRG 2009
Rui Li and Alex Klippel’s proposal “Stuck in the Stacks: Assessing and Improving Wayfinding in the Pattee and Paterno Libraries” has been accepted to present at the SNRG 2009 conference, where librarians from the Northeastern U.S. who are using the Integrated Library System (ILS) meet. The presentation will detail wayfinding solutions for the Penn State Libraries to audiences from all over the Northeastern United States, aiming to provide possible solutions to wayfinding problems which might exist in other libraries and indoor environment.
Submission accepted at Spatial Cognition and Computation
Title: Topologically Characterized Movement Patterns: A Cognitive Assessment
Abstract: In this paper, we discuss the role of topology as a predictor for the conceptualization of dynamically changing spatial configurations (referred to as movement patterns). We define meaningful units of movement patterns as paths through a topologically defined conceptual neighborhood graph. Topology plays a central role in formal approaches to human cognition and in predicting cognitive similarity ratings—although primarily for static spatial configurations. Formal specifications of the role of topology for characterizing movement patterns do exist, yet there is paucity of behavioral validation. To bridge this gap, we conducted an experiment based on the grouping paradigm to assess factors that underlie conceptualizations of movement patterns. The experiment was designed such that paths through the conceptual neighborhood graph were distinguished by topologically differentiated ending relations. We believe topology can make an important contribution in explaining movement conceptualizations. One recently formulated topology-based contribution is the endpoint hypothesis, asserting that a cognitive focus is placed on event ending relations. We discuss the results of our experiment in relation to previous experiments targeted toward a framework for modeling the cognitive conceptualization of dynamically changing spatial relations.
Submission accepted at the International Conference on Spatial Cognition (ICSC 2009)
Title: Linguistically encoded movement patterns – Sampling from the WWW to reveal regional differences
Sen Xu*, Alexander Klippel*, Alan M. MacEachren*, Prasenjit Mitra†, Ian Turton*, Xiao Zhang†, Anuj Jaiswal†
*GeoVISTA Center, Department of Geography, The Pennsylvania State University, PA, USA †Information Science and Technology, The Pennsylvania State University, PA, USA
Abstract: Language has a long tradition as a window to cognition and a particular focus has been placed on actions encoded linguistically (e.g., route directions). With the development of computational linguistics, the World Wide Web is becoming a resource that allows for systematically analyzing large corpora of text containing action descriptions that are spatially referenced. Our contribution advances evaluation possibilities for linguistic pattern existing within the same language describing movement patterns across different regions. We studied regional differences in the English language within human-created route directions extracted from the web. A corpus of about 10.000 documents that contain route directions has been created. These route directions are parsed into distinct parts such as origin, destination and meaningful route parts. Geo-references are identified in each document and matched against various databases (e.g., GeoNames) returning geo-coordinates. These coordinates allow us to regionalize the linguistic descriptions of movement patterns for further analysis. As a first focus, we are addressing the often raised question of how cardinal directions are used in route directions and whether regional differences in this usage pattern can be found. This analysis will allow us to map linguistic usage patterns on a regional scale such as the 50 states within the US. This approach can be extended to other countries and other languages. Regional linguistic differences can be analyzed in terms of political boundaries, terrain differences, and urban/rural areas. Other linguistic patterns such as preferences for multiword expressions or prepositions can be subsequently addressed.
Key words: Regional Linguistic difference, route description, spatial orientation, language pattern, cardinal/relative direction usage
Rui Li elected as Student director of EPBG
Rui Li, PhD Candidate in the Department of Geography at The Pennsylvania State University was elected as the Student Director of Environmental Perception and Behavioral Geography Specialty Group (EPBG) of American Association of Geographers (AAG) at the 2009 AAG Conference in Las Vegas. His role is to promote the research priorities of this specialty group among students and engagement of students' research within this group.