We went over properties of ZR languages and explained the ZR
inference algorithm.
30 Apr 2018
We went over the preliminaries section of the reversible
languages paper.
25 Apr 2018
Seouyoung presented an empirical comparison of MaxEnt methods
in phonology to MLE methods on Hungarian vowel harmony.
23 Apr 2018
We went over state-merging as an inference method.
18 Apr 2018
Class cancelled; instructor away.
Please read the following sections of Angluin
1982, for next Wednesday's (4/25) class.
sections 1, 2, 3
section 4: 4.1
section 5: 5.1, 5.2, 5.3
section 9
16 Apr 2018
Sophie presented her research on logically defining an
epenthetic transformation where the quality of the vowel is
determined by morphological, not phonological information.
11 Apr 2018
We discussed how to define logical transductions from one
relational structure to another.
09 Apr 2018
We continued the discussion of using propositional logic to
define sets of relational structures with k-factors. It was pointed
out that the without FO, special accomodations were needed to
distinguish the beginnings and ends of words.
04 Apr 2018
We provided a definition for k-factors of arbitrary relational
structures in addition to a propositional logic of which the
k-factors are the atomic propositions.
02 Apr 2018
We went over 1.3 of the Kearns and Vazirani chapter on
PAC-learning the class of concepts expressible with conjunctions of
literals.
Please read the draft of section 2.8 that I emailed for
Wednesday.
28 Mar 2018
Our guest
lecturer Rémi
Eyraud
presented slides
on distributional learning.
Ji Yea presented her research on optionality in Korean
etymological i-epenthesis and how it might be modeled as an OSL or
as an ISL function.
Hossep began talking about his research on reduplication with
2-way deterministic finite-state transducers.
26 Feb 2018
Nazila presented her research on how the tree languages
representing well-formed binding relations not be well described
with dominance and successor models of trees, but that it may be
more straightforward with c-command as a primitive relation (Frank
and Vijay-Shankar, 2001).
Lei also presented his work on binding relations from the
perspectibe Bruening's precede and command perspective.
21 Feb 2018
We
discussed this
handout on using relational learning to identify smallest
factors consistent with a most specific hypothesis which covers the
data.
19 Feb 2018
We discussed chapter 3 of De Raedt's book.
Jon presented his project on learning feature-based constraints in the context of this chapter.
14 Feb 2018
Hongchen presented her ideas on lifting OSL funnctions from strings to trees and we discussed what syntactic phenomena may be described in this way, if any.
We began discussion of Chapter 3 of De Raedt 2008 in anticipation of Jon's presentation.
We discussed the constraints definable with first order logic and precedence.
Reading for Wednesday Feb 14.
Chapter 3 of De Raedt's 2008 book Logical and Relational Learning
Front matter of De Raedt's 2008 book Logical and Relational Learning
05 Feb 2018
Class cancelled; instructor away.
31 Jan 2018
We discussed the successor model with features
Aniello presented the sections on MSO logic.
29 Jan 2018
Class now meets MW 4:00 to 5:20pm in the compling lab (SBS N250).
We discussed how to write constraints using the successor model and FO logic.
Reading for next class on successor model with features and MSO logic: chapter 2, DCP, sections 2.5, 2.6 and 2.7. It is a rough first draft! Please send typos etc and higher level comments to Hossep, who will compile and give to me.
25 Jan 2018
We finished going over the handout from last class.
Reading for next class on successor model and FO logic: chapter 2, DCP, up to section 2.5. It is a rough first draft! Please send typos etc and higher level comments to Hossep, who will compile and give to me.
23 Jan 2018
We reviewed the syllabus and introduced the course.
We discussed the time and location of class. It may change to MW 4-5:20pm, but next class will be 8:30-9:50am in the compling lab (SBS N250).
We began going over this handout on relational models. Do exercise 1 for next class.