Difference between revisions of "Vulcan/MeetingNotes/Aug09 2013"
From Knowitall
(→Notes) |
(→Notes) |
||
Line 28: | Line 28: | ||
** Scalability issues with BLP/MLN | ** Scalability issues with BLP/MLN | ||
− | * Abductive reasoning using linguistically motivated templates. | + | * What happens when deductive inference fails? |
+ | ** We find the answer that is most plausible. | ||
+ | *** e.g., (x, helps, a fox find food), where x is one of {sense of smell, thick fur, long tail, pointed teeth}<br/> | ||
+ | |||
+ | "sense of smell helps * find food" returns 7 million hits on Google.<br/> | ||
+ | "fur helps * find food" returns no hits. | ||
+ | |||
+ | |||
+ | |||
+ | ** Abductive reasoning using linguistically motivated templates. | ||
** Will be handy when deductive inference fails. | ** Will be handy when deductive inference fails. | ||
** Provides a different form of evidence. | ** Provides a different form of evidence. |
Revision as of 22:18, 9 August 2013
Update
- Focusing on system development.
- 1. Inferencer stub using Jena. Stub takes in axioms and rules and outputs a derviation.
- 2. Tested stub with axioms and rules that would help us solve the iron nail example.
- 3. Built Proposition extractor stub that converts the answer assertions into propositions represented as Open IE 4.0 tuples
- 4. Exploring other triplestore/inference systems (OWLIM and Seasame). Jena API doesn't readily support multiple derivations. Ask Jena community to find out if this is possible.
- Resource collection
- 1. Gathered assertions from Peter/Phil. Each assertion corresponds to a single multiple choice answer.
- 2. Found RDF representations for WordNet and imported them into Jena.
- Analysis
- 1. Selected 10 propositions that are single Open IE tuples as starting targets.
- 2. Started to write down steps involved in verifying these propositions.
Notes
- Do we really needed an inference engine (Jena).
- Scalability issues with BLP/MLN
- What happens when deductive inference fails?
- We find the answer that is most plausible.
- e.g., (x, helps, a fox find food), where x is one of {sense of smell, thick fur, long tail, pointed teeth}
- e.g., (x, helps, a fox find food), where x is one of {sense of smell, thick fur, long tail, pointed teeth}
- We find the answer that is most plausible.
"sense of smell helps * find food" returns 7 million hits on Google.
"fur helps * find food" returns no hits.
- Abductive reasoning using linguistically motivated templates.
- Will be handy when deductive inference fails.
- Provides a different form of evidence.
Agenda
Things to discuss in the meeting.
- Axiom extraction and Rules generation -- What is the scope of rules and axioms? We are going to use the target propositions to guide the scope.
- Stephen had an idea on some form of self-supervision. For e.g.,
The rule (x, type of, metal) AND (metal, conductor of, electricity) -> (x, conductor of, electricity)
can be scored based on instances of x, where Open IE provides support for the consequent.
- Framework: BLP or MLN?
- Can we directly use these without Jena? We are using Jena for efficiency purposes. How large can the inference graph be for BLP or MLN?
To Do
- System building
- 1. Create a system architecture page with a figure and overview of the main components.
- 2. Continue system building.
- Create a derivation scorer stub. This will be replaced with a MLN or a BLP scorer.
- Test with iron nail example.
- 3. Jena API doesn't readily support multiple derivations.
- Ask Jena community to find out if this is possible.
- See if OWLIM can be used as a replacement.
- 4. Try out Tuffy MLN implemenatation.
- Use output of iron nail example
- If easy to use write wrappers around Tuffy to hook into our system.
- 5 Write evaluation code.
- Check with Peter.
- Resource Collection
- Experiments