WikiCorpus

So far so good on those hypotheses. The WikiCorpus alpha source code will be available (in a few weeks) here. I'm building it presently and testing it on Wikipedia articles. It will provide a user interface for generating corpora for pronoun and reference discernment, and semantic parsing algorithm design.

Wikipedia Example

Nations can have literatures, as can corporations, philosophical schools or historical periods. Popular belief commonly holds that the literature of a nation, for example, comprises the collection of texts which make it a whole nation. The Hebrew Bible, Persian Shahnama, the Indian Mahabharata, Ramayana and Thirukural, the Iliad and the Odyssey, Beowulf, and the Constitution of the United States, all fall within this definition of a kind of literature. [1]

As written, this sentence does not immediately parse as per the working hypothesis. The problem occurs around the "collection of texts which make it a whole nation."

[literature of a nation][comprises][collection of texts] which [make][it][a whole nation].

Labeling the chunks A - F, we can see
B(A,C)
D(A,...)

However, by inserting the word “into”, or “become” into the text:

[literature of a nation][comprises][collection of texts] which [make][it][into][a whole nation].

and labeling the chunks A-G, we can see
B(A,C)
D(A,F(E,G))

So there may be occasions where the machine reading algorithm has to add text to a document to get the algorithm the parse correctly, or use pattern-based rules, for example involving the predicate make(X,Y). These can be reflected in NLG as well to allow sentence candidates to be formed where, as in this example, the “into” or “become” is implicit.

So with that in mind,

Nations can have literatures, as can corporations, philosophical schools or historical periods. Popular belief commonly holds that the literature of a nation, for example, comprises the collection of texts which make it [into] a whole nation. The Hebrew Bible, Persian Shahnama, the Indian Mahabharata, Ramayana and Thirukural, the Iliad and the Odyssey, Beowulf, and the Constitution of the United States, all fall within this definition of a kind of literature. [1]

[Nations][can have][literatures], as [can][corporations], [philosophical schools] or [historical periods]. [Popular belief][commonly holds] that [the literature][of][a nation] ... [comprises] the [collection of texts] which [make][it][into][a whole nation]. [The Hebrew Bible], [Persian Shahnama], [the Indian Mahabharata], [Ramayana and Thirukural], [the Iliad and the Odyssey], [Beowulf], and [the Constitution of the United States], all [fall within] this [definition][of][a kind of literature].

Machine Reading
B(A,C)
B(E&F&G,C)
I(H,N(K(J,L),O))
P(O,R(Q,S))
AA(T&U&V&W&X&Y&Z,AC(AB,AD))

The ampersands indicate the processing of commas.

Start a new predicate with A as the left argument.
Name that predicate B.
Place C in the right argument of that predicate.
Start a new predicate with the label B.
Place C in the right argument of that predicate.
Place E in the left argument of that predicate.
In the context of a comma list,
Place F in the left argument of that predicate.
In the context of a comma list,
Place G in the left argument of that predicate.
Start a new predicate with H as the left argument.
Name that predicate I.
In the right argument of that predicate,
Start a new predicate with J as the left argument.
Name that predicate K.
Place L in the right argument of that predicate.
...
Wrap that predicate in a new predicate with the current content in the left argument.
Name that predicate N.
Place O in the right argument of that predicate.
Start a new predicate with O as the left argument.
Name that predicate P.
In the right argument of that predicate,
Start a new predicate with Q as the left argument.
Name that predicate R.
Start a new predicate with S as the right argument.
Start a new predicate with T as the left argument.
In the context of a comma list,
Place U in the left argument of that predicate.
In the context of a comma list,
Place V in the left argument of that predicate.
In the context of a comma list,
Place W in the left argument of that predicate.
In the context of a comma list,
Place X in the left argument of that predicate.
In the context of a comma list,
Place Y in the left argument of that predicate.
In the context of a comma list,
Place Z in the left argument of that predicate.
Name that predicate AA.
In the right argument of that predicate,
Start a new predicate with AB as the left argument.
Name that predicate AC.
Place AD in the right argument of that predicate.

We can use “symbolic logic” operations to turn:
B(E&F&G,C)
into
B(E,C)
B(F,C)
B(G,C)

resulting in

B(A,C)
B(E,C)
B(F,C)
B(G,C)
I(H,N(K(J,L),O))
P(O,R(Q,S))
AA(T,AC(AB,AD))
AA(U,AC(AB,AD))
AA(V,AC(AB,AD))
AA(W,AC(AB,AD))
AA(X,AC(AB,AD))
AA(Y,AC(AB,AD))
AA(Z,AC(AB,AD))

We can also observe that V and W can become V₁&V₂, and W₁&W₂.

B(A,C)
B(E,C)
B(F,C)
B(G,C)
I(H,N(K(J,L),O))
P(O,R(Q,S))
AA(T,AC(AB,AD))
AA(U,AC(AB,AD))
AA(V₁,AC(AB,AD))
AA(V₂,AC(AB,AD))
AA(W₁,AC(AB,AD))
AA(W₂,AC(AB,AD))
AA(X,AC(AB,AD))
AA(Y,AC(AB,AD))
AA(Z,AC(AB,AD))

which in quad form is

A B C 1
E B C 2
F B C 3
G B C 4
J K L 5
5 N O 6
H I 6 7
Q R S 8
O P 8 9
AB AC AD 10
T AA 10 11
U AA 10 12
V₁ AA 10 13
V₂ AA 10 14
W₁ AA 10 15
W₂ AA 10 16
X AA 10 17
Y AA 10 18
Z AA 10 19

[1] Literature, Wikipedia

Wikipedia, Semantic Web

After I make the web-deliverable interface that allows users to expand these semantic frames into all possible related entities, I refactor some SRL parsers to turn natural language into this semantic data and I point the system at Wikipedia... weeks later, what does this enormous OWL file (possibly stored numerically) have to do with the Semantic Web?

A quote from Cycorp:

The success of the Semantic Web hinges on solving two key problems: (1) enabling novice users to create semantic markup easily, and (2) developing tools that can harvest the semantically rich but ontologically inconsistent web that will result. To solve the first problem, it is important that any novice be able to author a web page effortlessly, with full semantic markup, using any ontology he understands. The Semantic Web must allow novices to construct their own individual or specialized-local ontologies, without imposing the need for them to learn about or integrate with an overarching, globally consistent, master ontology.

Allowing users to type in natural language is the easiest way to generate semantic markup. Because users prefer to use natural language, any ontology that software can roundtrip natural language with will likely be an overarching (prevalent) one. A possible problem with the approach I'm using is that the ontology of the dataset used to train SRL-based parsers would, instead of being handcrafted by an expert, be a collaborative effort of people, hopefully experts, visiting a site — wikiontology is a relatively new idea.

Only after I have the dataset will I be able to say if the ontology from the planned website is advantageous to machine reasoning tasks. It shouldn't be terribly difficult to make a benchmark for the consistency of the wiki-generated ontology — possibly using natural language (after the parser is completed). For example, paraphrase corpora and other instruments could be of use in both generating and refining.

Wikipedia is a proof of concept that people can come together to generate collective knowledge resources, so — if we get the post-NLP/pre-NLG ontology right (prevalent as argued above) — the Semantic Web may resemble a distributed wiki-knowledgebase. The gigabytes of Wikipedia data would be a launching point.

Ontology and File Compression

Thinking on this post-SRL/pre-NLG ontology made me realize there was no easy way to compare different models.

If we look at rule systems, ontology and taxonomy as interoperating towards efficiently storing knowledge, then there might be a metric. That is, if system A compresses the same knowledge set better than system B and is more computationally efficient (in decompression/utilization), then we can say that system A is superior to system B (on that set) without resorting to aesthetics or philosophy. We have SUMO, the CYC upper ontology, ISO 15926, and others designed around real-world data and it's difficult to rank them.

The metaphor of file compression to knowledgebases might allow competition between differing methods. As systems are envisioned that mechanically generate rules, ontological structure or taxonomy (optimizing generators that create a system for a given knowledgebase), these metrics may be of use in comparing the resulting generated systems. Personally, I think it would be interesting to have algorithms that compress knowledgebases like tar, zip and 7zip do to files. Unfortunately, this approach is storage and speed-based and doesn't consider interface considerations-- for example, sets of things that are categorized for navigation.

Here's a link to a paper describing a relationship between AI (my field of research) and file compression. Apparently, there's a prize for compressing Wikipedia.

Natural Language and Semantic Web

Been browsing on the Web for various tools in natural language processing (NLP) and natural language generation (NLG). Presently looking at cypher, heart of gold, gate, stanford parser, charniak parser, enju, opennlp, lkb, assert, halogen, kpml and a couple of others. Also looking at a couple of methods of storing structured knowledge from a parse including Berkeley's FrameNet ontology and some that came with halogen and kpml. Hoping to throw these tools at the Wikipedia dataset (9.74 gb) and post some results.