giboin@andy.inria.fr (Alain Giboin) (06/12/91)
If you are interested, could you please complete the following questionnaire,
and return it to us? Thank you.
-------------------------------------------------------
| A Questionnaire |
| about the Explanation Capabilities of Expert Systems |
| and of Industrial Expert System Shells |
-------------------------------------------------------
-------------------------------------------------------------------------
* QUESTIONNAIRE ADDRESSEES
=> Expert system shells DESIGNERS
=> Expert system USERS
* QUESTIONNAIRE GOALS
=> To better characterize the explanation capabilities of experts
systems and of industrial expert system shells
=> To identify the explanation needs of users and the effective
use of explanations provided by expert systems
=> To determine if ideas proposed by researchers in the field of
explanations are effectively integrated in current expert systems.
And if not, why
-------------------------------------------------------------------------
* FEATURES OF YOUR EXPERT SYSTEM SHELL
<> Is your system devoted to a specific kind of task?
-> Analysis tasks?
- Classification []
- Diagnosis []
- Monitoring []
- Prediction []
- Others []
Specify:
-> Synthesis tasks?
- Design []
- Planning []
- Modeling []
- Others []
Specify:
-> Modification tasks?
- Repair []
- Control []
- Others []
Specify:
<> What kinds of explanations do system users need to perform the tasks above?
-> Explanations for analysis tasks?
- For classification:
- For diagnosis:
- For monitoring:
- For prediction:
- For other analysis tasks:
-> Explanations for synthesis tasks?
- For design:
- For planning:
- For modeling:
- For other synthesis tasks:
-> Explanations for synthesis tasks?
- For repair:
- For control:
- For other modification tasks:
<> What is the formalism used to represent knowledge?
-> Rules
- Propositional rules []
- First order rules []
- Fuzzy rules []
- Metarules []
- Others []
Specify:
-> Frames []
-> Semantic networks []
-> Hybrid representations []
<> What are the mechanisms of the inference engine?
-> Forward chaining []
-> Backward chaining []
-> Mixed chaining []
-> Bidirectional chaining []
-> Others []
Specify:
<> Is the system based on a complex architecture?
-> Blackboards []
-> Second generation architecture with:
- Deep models []
- Causal models []
- Qualitative models []
- Other models []
Specify:
-> Other architectures []
Specify:
<> Does the system use explicitly the notion of task?
Specify:
<> Does the system use hypermedia tools?
Specify:
* KINDS OF INTERACTIONS
<> What kind of user-system interaction is allowed?
-> Natural language dialogues []
-> Query language []
-> Menus []
-> Graphics []
-> Others []
Specify:
<> Who can guide the interaction?
-> The system []
-> The user []
-> Both of them []
<> Is it possible for the user to influence the system reasoning?
And how?
<> Is it possible for the user to modify the results provided by the system?
And how?
* USER'S CHARACTERISTICS
<> Has the end user been involved in the application development?
And how?
<> In which stage of the KBS life cycle explanations are intended to be used?
-> Development []
-> Validation []
-> End use []
-> Others []
Specify:
<> And who uses the explanations?
-> Knowledge engineer []
-> Expert []
-> End user []
-> Student []
-> Others []
Specify:
<> And for achieving what kind of goal?
-> Debugging []
-> Understanding []
-> Learning []
-> Others []
Specify:
<> If possible, specify the following characteristics of the user:
-> Needs:
-> Goals:
-> Level of expertise in the domain:
-> Level of computing knowledge:
<> If possible, give a short description of user task:
<> If possible, give a short description of the help expected by the user:
<> Is the user satisfied by the help provided by the system?
If not, why?
* EXPLANATION MODULE
+ Existence of the module
-----------------------
<> Is an explanation module available in the system?
If not, why?
+ Design of the module
--------------------
<> Is the explanation module:
-> General []
-> Specific to the application []
<> How were selected the kinds of provided explanations?
<> Was the user involved in the explanation module design?
If not, why?
+ Internal functioning of the module
----------------------------------
<> Are the explanations based on (explicitly stored, or implicitly
acquired) user models?
<> Were explanations acquired or inspired from the analysis
of explanation dialogues between humans?
<> Are explanations based on the reasoning traces?
<> Does the system use:
-> Discourse strategies []
-> Planning of explanation text []
+ Contents of explanations
------------------------
<> Does the system provide (resp. does the user need):
-> Shallow explanations []
-> Deep explanations []
-> Graphical explanations []
-> Follow-up explanations []
-> Multi-level explanations []
<> What kinds of questions are available?
-> "Why" questions []
-> "How" questions []
-> "Why-not" questions []
-> "When" questions []
-> Other questions []
Specify:
<> How explanations are presented?
-> Textually []
-> Graphically []
-> Other modes []
Specify:
+ Use of the module
-----------------
<> Do end users effectively use the explanation module?
If not, why?
<> Do end users meet some problems with the explanation module?
If yes, what kinds of problems?
-------------------------------------------------------------------------
* ANSWERS
Answers to the questionnaire may be sent by email either to:
Rose Dieng (INRIA): dieng@mirsa.inria.fr
or to:
Alain Giboin (INRIA): giboin@mirsa.inria.fr
Answers may also be sent by mail to the address given in the signature
below.
NB: Answers will be considered as anonymous.
Documentation
-------------
If some documentation about your system (especially documentation
about its explanation capabilities) is available, we'll be grateful
to you to send it to us.
-------------------------------------------------------------------------
* RESULTS
If you are all interested in knowing the results of the questionnaire,
we will published in this newsgroup.
-------------------------------------------------------------------------
THANK YOU VERY MUCH FOR RESPONDING TO THIS QUESTIONNAIRE !!
--
Alain Giboin, INRIA Sophia Antipolis -- France
giboin@andy.inria.fr Phone: (33) 93.65.78.14, Fax: (33) 93 65 77 66
INRIA - Sophia Antipolis, 2004 route des Lucioles, 06565 Valbonne Cedex, FRANCE