| Member Profile : Jana Diesner |
|
 | Contact Information | Address: -Map Me-
Jana Diesner
University of Illinois at Urbana-Champaign, The iSchool
GSLIS, UIUC
501 East Daniel Street
Champaign, IL, United States 61820
Phone : 217-244-3576
E-mail : jdiesner@illinois.edu
Website : http://people.lis.illinois.edu/~jdiesner/
|
| Bibliographic Information |
|
| | | Software & Data | | Active Calendar Listings |
| AutoMap | (Software) |
AutoMap is a text mining tool that enables the extraction of network data from texts. AutoMap can extract content analytic data (words and frequencies), semantic networks, and meta-networks from unstructured texts developed by CASOS at Carnegie Mellon. AutoMap exports data in DyNetML and can be used interoperably with *ORA. AutoMap subsumes classical Content Analysis by analyzing the existence, frequencies, and covariance of terms and themes. AutoMap can operate in both a front end with gui, and backend mode.
Main functionalities of AutoMap are:
Extract, analyze and compare mental models of individuals and groups.
Reveal structure of socio-technical systems from texts.
AutoMap also offers a variety of techniques for pre-processing Natural Language:
Named-Entity Recognition
Flexible ontology usage
Parts of Speech Tagging
Stemming (Porter, KStem)
Positive and negative filters
Collocation (Bigram) Detection
Proximity based network extraction |
| Construct | (Software) |
Construct, developed by CASOS, is a multi-agent model of network evolution.
Social, knowledge and belief networks co-evolve. Groups and organizations are treated as complex systems thus capturing the variability in human and organizational factors. In Construct individuals and groups interact communicate, learn, and make decisions in a continuous cycle. The program takes into account how agents learn through interaction conducted over different media and change their information, beliefs, and activities based on what they learn. This can be used for forecasting how a network can evolve and seeing if two groups that appear identical on one dimension actually evolve in the same way. The non-linearity of the model generates complex temporal behavior due to dynamic relationships among agents. These dynamic relationships are grounded in constructural theory, structuration theory and influence theory. Consequently, in Construct, the socio-cultural system is constructed and reconstructed through human interaction based on rules and resources. The changes in the social system are defined and analyzed through the lens of social network analysis. Construct can be run directly from *ORA and so take as input actual or hypothetical networks and output from Construct can be assessed with *ORA. |
| ORA software | (Software) |
ORA is a dynamic meta-network assessment and analysis tool developed by CASOS at Carnegie Mellon. It contains hundreds of social network, dynamic network metrics, trail metrics, procedures for grouping nodes, identifying local patterns, comparing and contrasting networks, groups, and individuals from a dynamic meta-network perspective. *ORA has been used to examine how networks change through space and time, contains procedures for moving back and forth between trail data (e.g. who was where when) and network data (who is connected to whom, who is connected to where
), and has a variety of geo-spatial network metrics, and change detection techniques. *ORA can handle multi-mode, multi-plex, multi-level networks. It can identify key players, groups and vulnerabilities, model network changes over time, and perform COA analysis. It has been tested with large networks (106 nodes per 5 entity classes). Distance based, algorithmic, and statistical procedures for comparing and contrasting networks are part of this toolkit.
Based on network theory, social psychology, operations research, and management theory a series of measures of criticality have been developed at CMU. Just as critical path algorithms can be used to locate those tasks that are critical from a project management perspective, the *ORA algorithms can find those people, types of skills or knowledge and tasks that are critical from a performance and information security perspective. |
|
|
|
| | | Network Graduate Programs | | Network Courses |
|
|
|
| | | Jobs Posted | | Sunbelt Submissions |
|
| Sunbelt XXIX - March 10 to March 15, 2009 - Bahia Hotel | | Workshop : Dynamic Network Analysis (DNA) and AutoMap |
A lecture and hands-on workshop in which attendees learn about Dynamic Network Analysis (DNA) and the DNA toolkit AutoMap for extracting networks from unstructured texts. The collection and storage of unstructured, natural language text data has become fast, cheap, and easy. Examples for potentially large-scale corpora are answers to open questions in questionnaires, emails, wikis, blogs, chatlogs, news, political debates, mission statements, and annual reports, among others. The challenge is to efficiently and systematically extract network data from these unstructured texts. Such networks might connect diverse entities such as people, organizations, and events and the relations among them. Relational data extracted from texts can help us in answering questions like: Who communicates what with whom? How do trends emerge, spread and vanish in blogs and chats? What groups promote or suppress what ideas, and how successful are they in that?
The workshop attendees acquire methodological expertise in network text analysis as well as hands-on experience in using AutoMap. Participants in this workshop gain experience with text analysis by learning and using AutoMap. They are introduced to several techniques for natural language processing and information extraction that are often applied in this process, such as identification of central topics and terms in single documents or corpora, filtering techniques, stemming (translate words into their morphemes), Parts of Speech Tagging (assign part of speech to every word), anaphora resolution (translate pronouns into the social entities that the pronoun refers to), Named Entity Extraction (identification of agents, organizations and places that are referred to by a name), and text coding according to user-defined ontologies or taxonomies. The participants are furthermore introduced to the network analysis of email data with AutoMap and the CEMAP sub-tool. We show how relational data extracted from texts can be loaded into *ORA (a network analysis package) in order to analyze, assess, and visualize the extracted data. Throughout all phases of the workshop we discuss several empirical examples and real-world applications for the techniques covered.
AutoMap is a software package, developed by CASOS, that has been applied by multiple groups and across different domains and languages to extract concepts, semantic networks, and meta-network relational data from texts. AutoMap output is in DyNetML and can be easily assessed using *ORA. Relatively unique features include user control over windowing, machine learning models for ontological classification, limited anaphora resolution, and bi-gram extraction. The software presented in this tutorial is Windows operating system based. Versions for vista, macs and linux also exist and participants with such needs should pre-load the AutoMap software from the CASOS website http://www.casos.cs.cmu.edu/projects/automap/. Participants should bring their own laptops to workshop. |
| Sunbelt XXX - June 29 to July 04, 2010 - Riva del Garda Fierecongressi | | Abstract : Attention Networks among Members of Congress |
The 2008 election campaign in the U.S. demonstrated to the public that political leaders have started to adopt a broad range of social networking services for communication, civic engagement, and mostly for self-marketing purposes. One type of these services is micro-blogging, which facilitates the real-time dissemination of short pieces of information to create public conversations. In this study we focus on the usage of micro-blogging by a particular group of people, namely the members of the U.S. Congress. By using a multi-method approach that combines social network analysis of the 144 Members of Congress (MoC) who engage in micro-blogging through the Twitter.com service, qualitative text analysis in a grounded theory fashion, and automated semantic analyses of the disseminated messages, we address the following questions: For what purposes are MoC primarily using micro-blogging? What key topics emerge as central themes among what groups of MoC?
Our preliminary findings indicate that from a usage pattern point of view, MoC utilize Twitter as a one-directional channel for announcing meetings, promoting their webpages, and referring to press releases in order to push current issues all of which function as ways to control individual impression management. Beyond that, our preliminary text analysis results suggest that MoC not only frame sensitive yet controversial topics such as the health insurance reform and the You Lie outburst by Representative Joe Wilson, but also started to use micro-blogging as a mechanism to socialize their messages by creating attention networks around issues they are passionate about. Attention networks aim to capture who people are referring to, but also who mentions them in their messages. |
| Abstract : Computational integration of network theory and topic modeling for investigating the relationship between socio-technical networks, funding, and innovation in the European Union |
| When text data pertaining to socio-technical networks are available, these texts are often either analyzed separately from the network data, or are reduced to the fact and frequency of the flow of data or objects between nodes. Examples for the joint availability of text data and network data include answers to open questions in classical network surveys, social media such as emails, blogs, and wikis, and the semantic web. Previous research on the relationship between language and networks suggests an impact of the position of individuals in the network on their motivation and ability to induce innovation and change in socio-technical networks. We present our findings from a study in which we empirically tested this relationship for the case of research proposal that were granted funding by the European Union under the Framework Programmes and a methodology that we developed in order to facilitate this type of studies. This methodology computationally integrates network theory and topic modeling, an unsupervised machine learning technique that reduces the dimensionality of text data to sets of semantically related words, such that network data are enriched through information from text data and vice versa. Our approach is based on prior work that assumes not only texts, but also authors and other types of entities and metadata to have probability distributions over topics (Mimno & McCallum 2008). We extend this notion by abstracting away from the level of individual authors and collaborators to the structural role level, where the actual role is defined by network theory. |
| Workshop : Relational Text Analysis and Network Analysis: From AutoMap to ORA |
A lecture and hands-on workshop that introduce attendees to relational text analysis, the AutoMap toolkit for extracting networks from unstructured texts, social and dynamic network analysis, and the ORA toolkit for analyzing networks. The collection and storage of unstructured, natural language text data has become fast, cheap, and easy. Examples of potentially large-scale corpora are surveys, emails, wikis, blogs, chatlogs, news, political debates, legal documents, mission statements, and annual reports. The challenge with these data is to efficiently, systematically and reliably extract network data from them. Such networks entail entities of one or more node types, such as people, organizations, and events and the relations among them. Going from texts to networks has helped people in answering questions like: Who communicates what with whom? How do trends emerge, spread and vanish in blogs and chats? What groups promote or suppress what ideas, and how successful are they in that?
The workshop attendees are introduced to several fundamental natural language processing and information extraction techniques that are often applied in text mining processes, such as the identification of central terms and themes within and across documents, positive and negative filters, stemming (converting words into their morphemes), parts of speech tagging (assigning a grammatical category to every word), anaphora resolution (translating pronouns into the social entities that the pronoun refers to), named entity extraction (identifying agents, organizations and places that are referred to by a name), and text coding according to user-defined ontologies. We will cover various ways of linking words and concepts into edges. The participants are furthermore introduced to the network analysis of email data with AutoMap and the CEMAP sub-tool. We cover how to load the extracted relational data into ORA in order to visualize, analyze, and interpret them. Throughout all phases of the workshop we discuss several empirical examples and real-world applications for the covered techniques.
All tools used in this workshop are developed by CASOS: AutoMap has been applied by multiple groups and across different domains and languages to perform a variety of Natural Language Processing techniques, to extract concepts, semantic networks, and meta-networks, and to infer beliefs from texts. Unique features include full user control over semantic units, machine learning based models for various Natural Language Processing routines such as entity extraction, and sentiment inference techniques.
CEMAP is a software tool that enables the user to extract social networks from email meta-data.
ORA is a widely used network science package that facilitates the examination of networks and their spatio-temporal dynamics at multiple levels. Unique features include the ability to handle multi-mode, multi-plex, and multi-level networks as well as both small and very large networks, 2D and 3D visualization, attribute and network statistics, change detection, network simulation, and the display of networks on maps.
Who Should Attend?
Those who are interested in text mining and the extraction of relational data from texts should attend. The material and its delivery is suitable for researchers and practitioners, alike. This is designed to be a non-technical workshop, however, by its very nature, the material will involve some mathematics, although this will be minimized as the delivery is driven towards forming an understanding the practice of using AutoMap, CEMAP, and ORA.
Topics Include:
Text Mining and Natural Language Processing: Content analysis, Extraction of semantic networks (one-mode networks) and meta-networks (multi-mode networks) from texts
Social and Dynamic Network Analysis
AutoMap and ORA software
Computer Equipment:
The software presented in this tutorial is Windows operating system based. Mac users should user virtual PC. All participants, if possible, are asked to pre-load the AutoMap from the CASOS website http://www.casos.cs.cmu.edu/projects/automap/ and the *ORA software from the CASOS website http://www.casos.cs.cmu.edu/projects/ora/. Participants are invited to bring their own laptops. Nevertheless, participants not able to bring a Windows-based laptop computer to the sessions are welcome to participate, and will still fully benefit from the workshop. The software will be screen-projected to the group as a live walk-through demonstration. Participants will be provided with a data CD containing installers for the software, sample data, and manuals; and will be guided through software installation and subsequent hands-on usage. |
| Sunbelt XXXI - February 08 to February 13, 2011 - Trade Winds Beach Resort
http://www.tradewindsresort.com/
St. Pete Beach | | Abstract : Extraction and validation of socio-technical network data about the Sudan from text corpora |
| In 2005, the Government of Sudan and the Sudan People's Liberation Movement (SPLM) signed a Comprehensive Peace Agreement. In 2011, Southern Sudan will hold a referendum regarding its independence from Northern Sudan. Network data representing interactions in this socio-technical system during this six year time period can help us to understand the development of culture and conflicts in this region. Since such data is hard to collect through classic methods such as surveys, we use Relation Extraction methods to approximate network data from publically available news coverage on Sudan. We will report on how we utilize theoretically grounded, lexicalized features and feedback loops with subject matter experts to adjust our relation extraction technology to this domain. This technology uses a model that we trained via supervised machine learning, the classifier used for that are Conditional Random Fields. We will present our results from analyzing the retrieved socio-technical network that comprises tribes, issues and resources. |
|
|
| |
|
