Thanks

1
Thanks
from an
Accidental Computer Scientist professor
??? (Jane W. S. Liu)
2
(No Transcript)
3
Lessons Learned

• What looks like something may be nothing
• What looks like nothing may turn out to be
  something
• There is something in everything

4
Lesson 1

• What looks like something may be nothing

Challenging, innovative, high-profile research
no impact
5
Optical link to deep space probes
6
Classical communication
s(t)
Quantum mechanical model channel capacity
r(t) s(t) n(t)
7
Lesson 2

• What looks like nothing may turn out to be
  something

Common sense, simple ideas high payoffs and
impact
8
Late 80s at Illinois
Why finish! How about do it imprecisely
Overloaded! No time to finish the work!
Kwei-Jay Lin
9
Imprecise Computations

• Language constructs for implementation of
  milestone (anytime) and sieve methods
• Algorithms for scheduling
• To minimize average error, maximum error,
  cumulative error, mean flow time, and so on
• With 0/1 constraint
• Incremental database queries and updates
• Support for graceful degradation and overload
  handling
• End-to-end quality of service management

10
Relational Model by E. F. Codd of IBM

• Widely adopted standard models in 1970 are
• Hierarchical model
• Network model (CODASYL standard)
• IBM IMS/DB dominated the DBMS market

• According to relational model proposed by Codd,
  data are represented by n-nary relations

11
Lesson 3

• There is something in everything

Even seemingly unchallenging work innovation
opportunity
12
SISARL Thematic Project

• People
• Team of faculties and students from NTU, NTHU,
  NCTU and Academia Sinica
• Advisors from NTU Hospital
• User-centered approach
• Requirements derived from user scenarios
• Technology- and infrastructure-constrained
  designs of consumer devices for the elderly
• Emphases
• Component-based design, production, quality
  assurance and configuration technologies
• Methods and tools for trading off in usability,
  configurability and other figures of merits
• Homepage http//sisarl.org

13
Examples of Human Centric Devices

• Home and personal automation devices smart
  storage panty, medication dispensers, object
  locator, fetchers, housekeeping aids
• Assistive devices helping-eyes and ears
  accessibility and mobility assistants
• Social robots pets and exercise companions
• Automation equipment for care-providing
  institutions smart food and medication carts,
  tools for moving and bathing patients, etc.

14
Load Pantry
15
(No Transcript)
16
HCD must be

• Safe
• All harmful conditions and user actions
  discovered and prevented
• Non-critical errors anticipated and most of such
  failures recoverable
• Usable Shirley likes to use it now will
  still want to use it when she is 95
• Flexible
• It can be configured and customized to suit
  different users and environments
• It can adapt to changes to users needs

17
Opportunities Challenges

• Architecture, Design and Implementation

• Excellent embedded and real-time platform,
  software, and algorithm choices

• Promising choices from other fields

• Configuration, customization adaptation

• User interaction and interface personalization

• Division of labor between user and machine

• Verification, validation, certification, and
  monitoring checking

18
http//SISARL.org
19
The highest reward for a persons toil is not
what one gets for it, but what one becomes by
it. by John Ruskin