 |
|
| |
|
| Papers |
|
| |
|
Towards Building Digital
Library as an Institution of Knowledge |
|
| |
|
| |
|
NSF Post-DL Futures Workshop Short Paper
Hsinchun Chen, Professor, University of Arizona
Download: PDF Version WORD Version
- What are we trying to do?
What is the problem we’re
trying to solve?
Digital library research aims to develop
the engineering and science for generating, capturing, accessing,
and utilizing
data, information, and knowledge in various digital formats,
for a variety of applications, and in a global, collaborative
human and system network.
- How is it done today, and what
are the limitations of current practice?
Data and information
are captured and represented in various digital formats and
are proliferating rapidly. However, the
techniques for accessing data and information are rudimentary
and imprecise, mostly based on simple keyword indexes, relational
queries, and/or low-level image or audio features (i.e., research
results of the 70s and 80s).
Digital library should become an
institution for capturing and transferring
human knowledge, instead of simply for accessing
data and information.
- What is new in our approach/technology,
and why do we think it will be successful? What gives
evidence that it will
work?
A knowledge discovery approach, based on both top-down
knowledge creation (e.g., ontologies, subject headings, user
modeling)
and bottom-up automated knowledge extraction (e.g., data mining,
text mining, web mining), promises to help transfer digital
library from an institution of data and information to an institution
of knowledge.
The rapid proliferation of digital data and information,
the increased sophistication and robustness in machine learning
and data/text/web mining techniques, and the ever-increasing
power of modern networks and computers will help contribute
to the new science of digital library.
- Assuming we are successful,
what difference does it make?
Digital libraries will soon
become a ubiquitous, global knowledge resource for education,
training, and (international) collaboration.
Its impacts will be felt in all aspects of human activities,
from industries to governments, and from education to research.
- How
long will it take, how much will it cost, and what are
the (measurable) milestones, mid-term,
and final exams?
Many fundamental technical, social, and
policy issues remain to be developed and researched. It calls
for a long-term (ten-year),
systematic research agenda to develop digital libraries into
an institution of human knowledge. A long-term NSF research
program with an annual funding level of $10M-$30M is needed
to promote the field and develop the science.
In 3-5 years (mid-term),
we expect to see significant advancement in knowledge creation
for various digital library applications
and domains based on more advanced algorithmic techniques and
proper human-computer interaction principles. Instead of accessing
low-level, fragmented data and information pieces, high-level,
abstract and decision-relevant knowledge will be accessed in
a seamless manner.
In 8-10 years (final exams), we expect the
systematic science and theory for digital library knowledge
creation to be developed
and validated. We will be able to access human knowledge in
a multi-lingual, multi-media, mobile, and semantics-based digital
library knowledge network.
|
|
| |
|
|
