The task-centric revolution. Weaving information into workflows Dagobert Soergel College of Information Studies LACASIST 2009-02-05.
Post on 27-Mar-2015
217 Views
Preview:
Transcript
The task-centric revolution. Weaving information into
workflows
Dagobert Soergel
College of Information Studies
LACASIST 2009-02-05
Disclaimer
This talk pulls together many ideasmany old, some newmany other people’s, some minesome implemented here and there,some still awaiting implementation
The point is the total vision
2
Computer work must be organized
not around applications
but around tasks
Functional or vertical integration
3
Tasks are accomplished collaboratively
Collaboration
Cross-user or horizontal integration
4
Ontology S.
Search
History / PIM
Creating work product
History-aware personal information store. Planning
Ontology Support
Functional Integration
Collaboration
Search
Creating
History / PIM
Ontology S.
Creating
Search
Collaborator User in focus Collaborator
Sense-makingSense-makingSense making
Task oriented processing
Outline
• The digital library of the future ─ DELOS
• CLASS. A Collaborative Lesson-planning And Search System ─ Katy Lawley
• Sense-making ─ Pengyi Zhang
• History-aware personal information store ─ Anita Komlodi
• Relevance relationships ─ Xiaoli Huang
6
7
The digital library of the future: A look at the DELOS
vision
DELOS Network of Excellence in Digital Libraries
Now the DELOS association
Funded by the EU, >50 DL research groups as members
8
The digital library of the future
A broad system of interlinked information & services that
• Is person- and task-centric
• Provides rich seamlessly integrated functionality DL = information + tools to process information
• Supports process execution and workflowin business, government, and daily work
• Supports user-to user communication & collaboration
• Supports users as consumers and as contributorsSupports massive collaboration leveraging many small contributions to construct very large resources (e.g., Wikipedia, social tagging)
• Thus, supports new ways of intellectual work
9
Issues in research and teaching
• User studies to• develop a truly user-centric view of computer-supported
functionality• learn how users could and would collaborate if properly supported
• Methods for deriving rich user profiles with minimal active user involvement, including discovery of users’ conceptual structures
• Interfaces that use a well-structured ontology (a faceted classification, an entity-relationship model) to • help users analyze a problem they face or a search topic • help users over time to assimilate the ontology structure
• Usability, effectiveness, and impact of such interfaces
10
Issues in research and teaching• The core issue of ontologies / classifications /
tagging schemes• Capturing individual and shared understandings of the
concepts in a domain• Expressing these understandings in a structured way to
communicate a common understanding within a community• Supporting individual users in developing their own ontology• Collaborative development of ontologies • Automatic integration or harmonization of ontologies
• Integrated storage and search of documents and data in many formats and degrees of structure
• Annotation, communication, and collaboration functionality
• Seamless integration of multiple systems and tools
CLASS
A Collaborative Lesson-planning And Search System
Katy Lawley
11
12
CLASS functionality The design
of a lesson planning system
13
1 A knowledge organization infrastructure that fosters development of shared understandings and supports organization of materials.
2 Mediated access to digital libraries containing many types of materials with powerful, ontology-enhanced search
3 Intellectual property rights and access management
4 A collaborative template-based authoring system with annotation facility
5 Communication functionality
6 Integration with the teachers total work environment, including lesson plan evaluation
7 Interaction history. Customization and personalization, adaptive
8 An interface that provides easy access to all this functionality
9 Extensibility
14
CLASS walk-through
15
Figure 1. Opening screen for creating a lesson plan
16
Figure 2. Lesson plan outline
17
Figure 2. Lesson plan outline
18
Figure 3. Selection of applicable standards
19
Figure 4. Selection of applicable vocabulary words
20
Figure 5. Query formulation for a theme in the lesson planning module
21
Figure 6a. Browse the thesaurus
22
Figure 6b. Descriptor found in detailed information for soldiers
23
Figure 7. Display of results
24
Figure 8a. Segment Display. Detailed information about the segment
25
Figure 8a. Segment Display. Detailed information about the segment
26
Figure 8b. Segment Assessment. The full form
27
Figure 8b. Segment Assessment. The full form
28
Figure 8c. Segment Display and Segment Assessment sharing the screen
29
Figure 9. Interview segment added to lesson plan module
30
1 Knowledge organization infrastructure
Foster development of shared understandings among users and support organization of materials. This includes
1.1 A lesson plan template that lays out the components of a lesson plan.
1.2 A query template to assist in formulating queries for learning objects and other materials.
1.3 A material appraisal form using the same structure as the query template, same structure as 1.2.
1.5 A hierarchy of educational standards from several jurisdictions.
1.4 A thesaurus / classification of topics in each accessed database domain that is useful for searching and for giving teachers content ideas.
31
6 Integration with the teacher’s total work environment
6.1 Associate lesson plans with time slots for a given class.
6.2 Prepare requests for permission to use materials as needed.
6.3 Have material, such as quizzes, prepared in sufficient number or arrange for electronic administration.
6.4 Order equipment needed as specified in the lesson plan.
6.5 Notify the school library media specialist of assignments that require the use of the school library media center.
6.6 Have a function for recording grades and importing detailed standardized test scores (broken down by educational standard).
32
Experience with CLASS
• Teachers were able to search the archive.
• Teachers used the materials appraisal form for their own purposes (not enough critical mass for collaboration).
• Teachers were able to develop well-structured lesson plans in a short time.
• Teachers indicated that they do not collaborate much at present but would be inclined to collaborate more if they had a tool like CLASS.
Sense-making
Pengyi Zhang
33
Beyond Search
Beyond search: Users need support for the next step: making sense of and applying large quantities of information found.
34
Outline
• What is sense-making?
• Examples
• Theoretical framework and comprehensive sense-making model
• Findings from a pilot study
• Conclusions on the design of sense-making systems
35
Sense-making
Sense-making is
the process of creating an understanding of a problem or task
so that further actions may be taken in an informed manner (Stefik et al., 1999).
36
Sense-making
• Sense-making is a pre-requisite for many other tasks such as decision making and problem solving;
• Sense-making involves making clear the interrelated concepts and their relationships in a problem or task space.
37
Sense-making examples
38
Sample Sense-making Scenario 1
Task T1: al-Bashir (Abridged Version)
The US wants to take action to towards a resolution of the Darfur conflict . Al-Bashir, the Sudanese president, is one of the key players in the area who is believed to have significant responsibility for continuous conflicts in the region. The administration needs to know as much as possible about al-Bashir in order to better negotiate with the involved parties and strategize its efforts. Your task is to produce a report that identifies information to assess the influence of al-Bashir and makes recommendations for policy decisions and diplomatic actions.
Requested information includes:• key figures, organizations, and countries who have been associated with
al-Bashir;• his rise to power; and • groups who have resisted him and the level of success in their resistance.
39
40
Sample Sense-making Scenario 2
Task T2: Energy Security (Abridged Version)
At present, U.S. energy security depends on a range of countries across the globe, many of which could be characterized as politically unstable and afflicted with war, piracy and terrorism. Your task is to produce a report of the geopolitics of oil in the major suppliers of U.S., including Mexico, Saudi Arabia, Venezuela, Nigeria, Algeria, etc. Requested information includes:
•the political, economic, and military status of major oil suppliers;
•threats to U.S. oil supplies;
•transit chokepoints of world oil.41
Sample Think-aloud Protocol with Coding
Protocol (Energy Security Task) Loops Processes Conceptual Changes
Cognitive Mechanisms
Okay that was actually a very useful search. So let’s still take this query and look at Algeria, ‘cause obviously Algeria and Nigeria are very close…
Loop 4 Focused Search for data
Comparison
I understand some of the keywords in the article but I don’t understand what the article…
Sense-making Failure
Key item extraction
Okay this has to do with Algeria, southern Algeria. The minister of energy… OPEC meeting… so I am going to see what their connections are with OPEC.
Building structures
Key item extraction
… with all the violence in Nigeria, I was expecting to find the same types of political outrage in Algeria…
Instantiating structures
Comparison and analogy
and I’m not seeing any notice of that at all.
Updating of knowledge
Re-structuring
42
Example Concepts and Relationships Concepts/Entity Relationships
in new informationRelationship in existing knowledge
Nigeria (entity)
Political violence (concept)
Nigeria <hasSituation> Political violence
Algeria Algeria <is very close to> Nigeria
Algeria <hasSituation> Political violence ?
Political stability Algeria <hasSituation> Political stability
Nigeria <hasSituation> Political violence
The minister of energy
OPEC meeting Nigeria <hasRelationshipTo> OPEC ?
… 43
Sample Sense-making Scenario 3
Task
Write a newspaper article on the role of energy policy in the presidential campaign.
44
45
Candidates positions on energy. Take 1
Candidates positions on energy. Take 2
46
Theoretical Framework
Sense-making models– Generic sense-making models
– Sense-making models of intelligence analysis
– Conducting research as sense-making
– Organizational sense-making
– Individual vs. collaborative sensemaking
47
Cognition– Types of conceptual changes
– Cognitive mechanisms
– Cognitive structures of knowledge
Learning– Schema theory
– Assimilation theory
– Generative Learning Theory
– Structural knowledge acquisition
Sense-making Elements
48
Processes
Sensing
Making sense
Activities Mechanisms
Accretion
Tuning
Restructuring
Inductive, data-driven•Key item extraction•Comparison•Schema induction•Generalization
Deductive, structure-driven•Definition•Specification•Elimination•Explanation •Inference
Outcomes
Identification of gaps
Search
Building structure
Instantiating structure
Consuming instantiated structure
Other•Metaphor•Classification•Semantic fit•Socratic dialogues
Exploratory searchFocused search
For dataFor structure
An Iterative Sense-making Model
49
Datagap
Structuregap
Search:exploratory / focused
OutcomesUpdated knowledge
Structure loop
Data loop
Identification of Gaps
Searchingfor data
Instantiating structure
Accretion: Instantiated structure
Tuning:Adapted structure
Re-structuring:New structure
Searching for structure
Building structure
Existing Knowledge
Structures and their
instantiations with data
The iterations proceed from exploratory to focused search and sense-making.
Task / Problem
Decision / Solution / Task completion
Findings – Search and Sense-making Loops
• The overall search and sense-making loops followed four stages:– Task analysis– Exploratory stage– Focused stage– Updates of knowledge representation.
• Reasons for starting a new loop of search and sense-making:– Success of previous sense-making– Failure of previous sense-making– New lead– Failure of search
50
Conclusions on the designof sense-making systems
51
Integration of tools
• Search tool
• Annotation / indexing tool
• Structure-building and visualization tools, manual and computer-assisted
• Writing tool
52
Functions in the integrated environment
53
Functions in the integrated environment
• Build structures: concept maps, templates, outlines
• Select in search results (within one document, a single document, or multiple documents), drag and drop on existing structure node or link (accretion), or drop in empty space to create a new node (structure modification)
• Have system find the node or link where a text passage (or image) should be attached
• Assist in extracting assertions – information extraction
• Find other sources for an existing assertion (automated accretion)
• Always preserve the source (as in MS OneNote) 54
Functions in the integrated environment
• Start a search from a structure node, using as query the node label or a query learned automatically from the documents already at the node
• Start a search from any selected text passage, for example, a text passage in the draft report
• KOS-supported search(KOS = Knowledge Organization System)
– Query expansion
– Browse KOS structure to clarify search topic, find search terms
55
Functions in the integrated environment
• Switch between structure formats, for example, from concept map to outline or template
• Assist in creating structure, for example
– Insert relationships extracted from text into a concept map or construct a concept map from scratch from extracted relationships
– Clustering
– Find existing structures – for example, search for structures on the Web
56
Functions in the integrated environment
• Create a draft report from a concept map
– Create outline and insert texts associated with each node
– Express relationships through text (text generation)
57
Functions in the integrated environment
• Screen real estate is important – a very large screen or two monitors
• Easy input mode (so as not to distract from thinking)
– Voice input
– OCR pen input from print material
58
History-aware personal or group information store
Anita Komlodi
59
The total information store
• Store everything
• Connect everything
• Search and find everything
• Keep detailed provenance & use history
• History is both past and future ─ what was done and what is to be done
60
Store everything
• Documents of all kinds
• Tasks
• Actions, events
• People and organizations
• Concepts, issues, frameworks (ontology)
61
Search and find everything
• Flexible search − any item as starting point, any connection type.For example– Find all items connected to a task– Find all events and actions in a given time span– Find all items connected, directly or indirectly, to a
person• Present search results in different views, organize
into a meaningful structure
62
Tools
• Search tool
• Gather tool
• Compare tool
• Connection tool, including annotation function
• Structure creation and display tool
• Scratch pad
• Ontology generation tool
– Personal ontology
– Group ontology
– Task ontology
• Writing / action tool63
Relevance relationships
Xiaoli Huang
64
Information arranged by role in argument
65
Relationship of info to task
Matching / allowing inference on a topic
. Matching topic / direct relevance
. Allowing inference on the topic / indirect relevance
Context
Comparison
Cause and effect
Goal
Method / Solution
Evaluation
Use these relationships when linking information to tasks
66
Relevance relationships
Matching / allowing inference on a topic
. Matching topic / direct relevance
. . Matching topic at same level of detail
. . . Direct statement
. . . Reference
. . . Definition
. . . Restatement
. . . . paraphrase,
. . . . clarification
. . Summarization
. . . Abstraction
. . Elaboration
. . Interpretation
67
Relevance relationships
Context. Background
. Scope, broader
. Framework
. Assumption or expectation
. Preparation
. Environmental setting
. . Physical environment
. . Social, political, cultural background
. By time sequence
. Condition
. . Enabling or hindering condition
68
Relevance relationshipsComparison. By similar vs. different
. . Comparison, similar
. . Comparison, different
. By factor that is different
. . Difference in external factor
. . . Different time
. . . Different place
. . . Different type of situation
. . Difference in participant
. . Difference in act or experience
. . . Difference in act
. . . . Different purpose
. . . . Different method
69
Some conclusions
• The system needs to think with the user– Adapt to the user’s way of thinking
but also– Help the user to structure a problem, organize information,
and prepare a plan for a task• Dividing users’ work into applications is a distraction• The gradual evolution of systems needs to make way for a
task-centric revolution• This presents challenges
– Task-oriented organization and intelligent processing of information
– Technical implementation
70
Questions
dsoergel @ umd.edu
www.dsoergel.com
71
Leftover slides
72
Pilot Test
• Participants
• Data collection
• Data analysis
• Findings
73
Participants
• 6 information science students
• Tasked with synthesizing data from a variety of sources, assessing the credibility of information, and evaluating claims based on supporting evidence
• Trained in using Rosetta - a multilingual, multimedia news retrieval system
74
Data Collection
• 2-hour task sessions, 6 taskstotal of 17 sessions completed
• Participants were instructed to verbalize their thoughts as they work on the tasksThink-aloud protocols were transcribed
• Post-session interviews as supplemental data
75
76
Data Analysis – Coding SchemeA Processes (from the model)
Search Sense-making
Exploratory search
Exploratory search for data
Exploratory search for structure
Focused search
Focused search for data
Focused search for structure
Gap identification
Data gap
Structural gap
Building structures
Using automatically extracted results
Extracting relationships manually
Instantiating structures
Updating knowledge
B Conceptual Changes (from the model)
Sense-making success Sense-making failure
Accretion Unable to fit data into structure
Tuning
Re-structuring Unable to build new structure
77
Data Analysis – Coding SchemeC Cognitive Mechanisms (from the model)
Inductive mechanisms Deductive mechanisms Other
Key item extraction
Comparison
Similarity
Differentiation
Schema induction
Generalization
Definition
Specification
Explanation-based
Elimination
Inference
Analogy and metaphor
Classification
Socratic dialogues
Semantic fit
D Emerging Codes Added During Analysis
Reasons starting a new loop Resolution of conflicts
New lead Disregard conflicting evidence
Sense-making success Compromise
Sense-making failure Accept new evidence
Search failure Confusion78
Findings – Cognitive Mechanisms
• Our participants used a two-way approach:
data-driven (bottom-up) 80%
structure-driven (top-down)20%
• Key item extraction and comparison were used most often.
79
Findings – Dealing with Conflicts
• Disregard – “… I wanted that article to say something else. I have to disregard it.”
• Compromise – “…it is not my understanding that this has anything to do with oil. But okay, this is a different take.”
• Acceptance – “…I thought these countries had similar serious problems. But [in fact, they] seem to be relatively stable and put a lot of effort into establishing their economy.”
• Confusion – “…Obviously I have no idea what this is about...”
80
Findings – Role of Instantiated Structures
• Entities (represented as names) and key concepts (represented as keywords) were often the basis for relevance judgments.
• The relationships embedded in new information and between the new information and participants’ previous knowledge seemed to play an important role in structure building and data fitting.
• Both concepts and relationships seemed to be crucial for updating knowledge.
81
Pengyi Zhang
pengyi@umd.edu
terpconnect.umd.edu/~pengyi/sensemaking/
82
Extending Sense-Making Models with Ideas from Cognition and Learning TheoriesPengyi Zhang, Dagobert Soergel, Judith Klavans, and Douglas Oard pengyi@umd.edu, ASIST 2008, Sunday Oct. 26, Session on Sense Making. Online Proceedings, Paper 29, p. 1 – 23 [34 – 56]http://terpconnect.umd.edu/~pengyi/sensemaking/files/zhang_asist08_sensemaking_model.pdf
Conclusions
The model is a useful framework for understanding the users’ sensemaking process
•The empirical think-aloud data seem to be consistent with the iterative sense-making model
•Users used a combination of data-driven and structure-driven mechanisms
•Instantiated structure elements play an important role in sense-making
83
Implications and Future Work
• Theoretical– Better understanding of sense-making processes
by extending the existing models with ideas from cognition and learning
– Further testing and refinement.
• Empirical– Better system design based on findings from user
studies.– Making design suggestions to support sense-
making, not just search.
84
A Sample Sense-making Scenario
Task
An intelligence analyst is tasked to gather, analyze, and synthesize information related to Al-Bashir and, on that basis, make recommendations for action in the form of a report.
Background
The Darfur conflict is a crisis in the Darfur region of western Sudan. Al-Bashir, the Sudanese president, is one of the key players in the area who is believed to have significant responsibility for continuous conflicts in the region. As part of an effort to resolve these armed conflicts, the administration needs to know as much as possible about al-Bashir in order to better negotiate with the involved parties and strategize its efforts.
85
A Sample Task Scenario
Task T1: al-Bashir (Abridged Version)
Omar Hasan Ahmad al-Bashir is a Sudanese military leader, dictator, and current president of Sudan. Your task is to produce a report identifying information to assess the influence of al-Bashir as a basis for policy decisions and diplomatic actions. Requested information includes:
• key figures, organizations, and countries who have been associated with al-Bashir;
• his rise to power; and • groups who have resisted him and the level of success
in their resistance.
86
87
88
89
top related