Finite-state automata 3 Morphology Day 14 LING 681.02 Computational Linguistics Harry Howard Tulane University Course organization http://www.tulane.edu/~ling/NLP/ NLTK is installed on the computers in this room! How.
Download ReportTranscript Finite-state automata 3 Morphology Day 14 LING 681.02 Computational Linguistics Harry Howard Tulane University Course organization http://www.tulane.edu/~ling/NLP/ NLTK is installed on the computers in this room! How.
Finite-state automata 3 Morphology Day 14 LING 681.02 Computational Linguistics Harry Howard Tulane University Course organization http://www.tulane.edu/~ling/NLP/ NLTK is installed on the computers in this room! How would you like to use the Provost's $150? 25-Sept-2009 LING 681.02, Prof. Howard, Tulane University 2 SLP §2.2 Finite-state automata 2.2.6 Recognition as search Non-deterministic recognition: Search In a non-deterministic FSA, there is at least one path through the machine for a string that is in the language defined by the machine. There is no path through the machine that leads to an accept state for a string not in the language. But not all paths directed through the machine for an accept string lead to an accept state. 25-Sept-2009 LING 681.02, Prof. Howard, Tulane University 4 Non-deterministic recognition So success in non-deterministic recognition occurs when a path is found through the machine that ends in an accept. Failure occurs when all of the possible paths for a given string lead to failure. 25-Sept-2009 LING 681.02, Prof. Howard, Tulane University 5 Back to the example q0 q1 b 25-Sept-2009 q2 a q2 a q3 a q4 ! LING 681.02, Prof. Howard, Tulane University $ 6 q 0 b a a a ! a a a ! a a ! Example q 1 1 b q 2 2 b 3 a q b a a 4 5 q 2 3 a ! b a a a ! q q 2 4 X b 25-Sept-2009 a a a ! b LING 681.02, Prof. Howard, Tulane University a a a 6 ! 7 Summary States in the search space are pairings of tape positions and states in the machine. By keeping track of as yet unexplored states, a recognizer can systematically explore all the paths through the machine given an input. 25-Sept-2009 LING 681.02, Prof. Howard, Tulane University 8 Keeping track But how do you keep track? Depth-first/last in first out (LIFO)/stack Unexplored states are added to the front of the agenda, and they are explored by going to the most recent. Breadth-first/first in first out (FIFO)/queue Unexplored states are added to the back of the agenda, and they are explored by going to the most recent. 25-Sept-2009 LING 681.02, Prof. Howard, Tulane University 9 Depth-first/LIFO/stack q2 q12 q27 q18 q12 q2 q27 q31 stack q41 25-Sept-2009 q50 LING 681.02, Prof. Howard, Tulane University 10 Breadth-first/FIFO/queue q2 q12 q27 q18 q31 q2 q12 q27 queue q41 25-Sept-2009 q50 LING 681.02, Prof. Howard, Tulane University 11 SLP §2.2 Finite-state automata 2.2.7 Comparison Equivalence Non-deterministic machines can be converted to deterministic ones with a fairly simple construction. That means that they have the same power: non-deterministic machines are not more powerful than deterministic ones in terms of the languages they can accept. 25-Sept-2009 LING 681.02, Prof. Howard, Tulane University 13 Why bother? Non-determinism doesn’t get us more formal power and it causes headaches, so why bother? More natural (understandable) solutions. 25-Sept-2009 LING 681.02, Prof. Howard, Tulane University 14 SLP §3 Words and transducers Intro Concepts and terminology study of spelling orthography study of word composition morphology to build a structured representation of a word or sentence input to this process a process that applies without limitations Can all forms be stored in advance? 25-Sept-2009 parsing surface or input form productive LING 681.02, Prof. Howard, Tulane University 16 Concepts and terminology the minimal meaning-bearing unit in a morpheme language the main unit additional units a unit that: stem affix prefix precedes the main one follows the main one surrounds the main one is inserted within the main one a language in which the main unit can have many additional units 25-Sept-2009 suffix circumfix infix agglutinative LING 681.02, Prof. Howard, Tulane University 17 Concepts and terminology Combining an affix to a stem inflection does not change the part of speech of the stem. Combining an affix to a stem DOES change the part of speech of the stem. Combining multiple stems. Combining a stem with a phonologically reduced stem. derivation 25-Sept-2009 compounding cliticization LING 681.02, Prof. Howard, Tulane University 18 SLP §3 Words and transducers §3.1 Survey of (mostly) English morphology Inflectional morphology stem -s -ing preterite past part. walk walks walking walked walked try tries trying tried tried map maps mapping mapped mapped eat eats eating eaten catch catches catching caught caught be is been 25-Sept-2009 being ate was LING 681.02, Prof. Howard, Tulane University 20 Next time P4 SLP §3.2ff