alignment for morphology induction
play

Alignment for Morphology Induction Tzvetan Tchoukalov Christian - PowerPoint PPT Presentation

Jun 13, 2023 •246 likes •539 views

Multiple Sequence Alignment for Morphology Induction Tzvetan Tchoukalov Christian Monson Brian Roark Multiple Sequence Alignment in Biology prokaryote16S rRNA Columns 1623-1703 out of 7683

  1. Multiple Sequence Alignment for Morphology Induction Tzvetan Tchoukalov Christian Monson Brian Roark

  2. Multiple Sequence Alignment in Biology prokaryote16S rRNA Columns 1623-1703 out of 7683 ---T--C---C-G--------------C----T-G---A-TA-G---AT---G-G-----G-CTC-GCG--T-CTG--A ------G---T-G--------------G----T-A---T-AA-G---AT---G-G-----A-CCC-GCG--T-TGG--A ------G---T-G--------------G----T-A---T-AG-G---AT---G-G-----A-CCC-GCG--T-CTG--A ------G--GC-G--------------G----T-G---A-AG-G---AT---G-A-----G-CCC-GCG--G-CCT--A ------C---C-G--------------G----T-A---G-AC-G---AT---G-G-----G-GAT-GCG--T-TCC--A ---T--C---C-G--------------C----T-T---T-GA-G---AT---G-G-----C-CTC-GCG--T-CCG--A

  3. Multiple Sequence Alignment in Biology prokaryote16S rRNA Columns 1623-1703 out of 7683 ---T--C---C-G--------------C----T-G---A-TA-G---AT---G-G-----G-CTC-GCG--T-CTG--A ------G---T-G--------------G----T-A---T-AA-G---AT---G-G-----A-CCC-GCG--T-TGG--A ------G---T-G--------------G----T-A---T-AG-G---AT---G-G-----A-CCC-GCG--T-CTG--A ------G--GC-G--------------G----T-G---A-AG-G---AT---G-A-----G-CCC-GCG--G-CCT--A ------C---C-G--------------G----T-A---G-AC-G---AT---G-G-----G-GAT-GCG--T-TCC--A ---T--C---C-G--------------C----T-T---T-GA-G---AT---G-G-----C-CTC-GCG--T-CCG--A Sequences of symbols Sequences are related e.g. serve same function in different organisms Why? To identify conserved regions To identify regions with similar physical structure

  4. Multiple Sequence Alignment for Morphology d – a n c – e s English Verbs d – a n c – e d d – a n c - e d – a n c i n g r – u n n i n g j – u m p i n g j – u m p – e d j – u m p - s j – u m p - - - l a u g h i n g Sequences of symbols Sequences are related e.g. serve same function in different words Why? To learn morphological structure

  5. Language Vs. Biology Differences # of Length of Symbol = Sequences Sequences Meaning to Align 10’s Language Millions No 10’s Biology Millions Yes Similarities Both involve sequences Size of Alphabet (less than 100)

  6. What We Did 1. Progressive alignment To build a profile 2. Leave-one-out realignment 3. Align words to the profile 4. Segment words Based on alignment

  7. Step 1) Progressive Alignment A Profile 1 2 3 4 5 6 7 8 d – a n c – e s d – a n c – e d d – a n c - e d – a n c i n g r – u n n i n g j – u m p i n g j – u m p – e d j – u m p - s j – u m p - - - l a u g h i n g

  8. Step 1) Progressive Alignment A Profile 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 d – a n c – e s a 1 2 5 1 1 1 5 1 Column d – a n c – e d c 1 1 1 1 5 1 1 1 d – a n c - e d 5 1 1 1 1 1 1 3 Distributions d – a n c i n g e 1 1 1 1 1 1 1 1 r – u n n i n g g 1 1 1 2 1 1 1 5 j – u m p i n g h 1 1 1 1 2 1 1 1 j – u m p – e d i 1 1 1 1 1 5 1 1 j – u m p - s j 5 1 1 1 1 1 1 1 j – u m p - - - l 2 1 1 1 1 1 1 1 l a u g h i n g m 1 1 1 5 1 1 1 1 n 1 1 1 6 2 1 5 1 p 1 1 1 1 5 1 1 1 r 2 1 1 1 1 1 1 1 s 1 1 1 1 1 1 2 2 u 1 1 7 1 1 1 1 1 gap 1 A 1 1 1 7 2 4

  9. Step 1) Progressive Alignment A Profile 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 d – a n c – e s a 1 2 5 1 1 1 1 1 Laplace d – a n c – e d c 1 1 1 1 5 1 1 1 d – a n c - e d 5 1 1 1 1 1 1 3 Smoothing d – a n c i n g e 1 1 1 1 1 1 5 1 r – u n n i n g g 1 1 1 2 1 1 1 5 j – u m p i n g h 1 1 1 1 2 1 1 1 j – u m p – e d i 1 1 1 1 1 5 1 1 j – u m p - s j 5 1 1 1 1 1 1 1 j – u m p - - - l 2 1 1 1 1 1 1 1 l a u g h i n g m 1 1 1 5 1 1 1 1 n 1 1 1 6 2 1 5 1 p 1 1 1 1 5 1 1 1 r 2 1 1 1 1 1 1 1 s 1 1 1 1 1 1 2 2 u 1 1 7 1 1 1 1 1 gap 1 A 1 1 1 7 2 4

  10. Step 1) Progressive Alignment A Profile 1 2 3 4 5 6 7 8 1. Sort words by frequency d – a n c – e s d – a n c – e d 2. Using Levenshtein distance d – a n c – e - In first n=1000 words d – a n c i n g Find most similar pair of words, W 1 and W 2 r – u n n i n g 3. Align W 1 and W 2 (using Levenshtein) j – u m p i n g This is our Profile j – u m p – e d j – u m p – s - 4. For i =3 to M=5000, 10000, … j – u m p - - - Find word W i , most similar to W j , j<i l a u g h i n g Align W i to profile

  11. Step 1) Progressive Alignment A Profile 1 2 3 4 5 6 d a n c e s d a n c e d d a n c e - d a n c i n g New W i

  12. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 d a n c i The Goal n g

  13. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 d 4.4 a 5.9 n 7.4 c 8.8 i 10.4 n 11.9 g 13.4

  14. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 d 4.4 1.6 a 5.9 Match n 7.4 cost = -log P(character) c 8.8 i 10.4 n 11.9 g 13.4

  15. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 d 4.4 8.8 a 5.9 Insert gap into new word n 7.4 cost = -log P(gap) c 8.8 i 10.4 n 11.9 g 13.4

  16. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 d 4.4 8.8 a 5.9 Insert gap into n 7.4 alignment profile cost = -log P(unattested) c 8.8 i 10.4 n 11.9 g 13.4

  17. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 d 4.4 1.6 a 5.9 Match n 7.4 cost = -log P(character) c 8.8 i 10.4 n 11.9 g 13.4

  18. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 d 4.4 1.6 6.0 7.5 9.0 10.5 12.0 a 5.9 3.1 3.1 7.6 9.1 10.6 12.1 n 7.4 4.6 4.6 4.7 9.1 10.6 12.1 c 8.8 6.1 6.1 6.2 6.2 10.7 12.2 i 10.4 7.6 7.6 7.7 7.7 9.2 12.9 n 11.9 9.1 9.1 9.2 9.2 10.7 12.1 g 13.4 10.6 10.6 10.7 10.7 12.2 13.6

  19. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 d 4.4 1.6 6.0 7.5 9.0 10.5 12.0 a 5.9 3.1 3.1 7.6 9.1 10.6 12.1 n 7.4 4.6 4.6 4.7 9.1 10.6 12.1 c 8.8 6.1 6.1 6.2 6.2 10.7 12.2 i 10.4 7.6 7.6 7.7 7.7 9.2 12.9 n 11.9 9.1 9.1 9.2 9.2 10.7 12.1 g 13.4 10.6 10.6 10.7 10.7 12.2 13.6

  20. Align W i to Profile d a n c e d Dynamic d a n c e s Programming d a n c e - 0.0 4.4 5.9 7.4 8.9 10.4 11.9 1 2 3 4 5 6 7 d a n c – e s d 4.4 1.6 6.0 7.5 9.0 10.5 12.0 d a n c – e d d a n c – e - a 5.9 3.1 3.1 7.6 9.1 10.6 12.1 d a n c i n g n 7.4 4.6 4.6 4.7 9.1 10.6 12.1 c 8.8 6.1 6.1 6.2 6.2 10.7 12.2 i 10.4 7.6 7.6 7.7 7.7 9.2 12.9 n 11.9 9.1 9.1 9.2 9.2 10.7 12.1 g 13.4 10.6 10.6 10.7 10.7 12.2 13.6

  21. Steps 2 & 3 Step 2) Leave-one-out realignment Improves the greedy alignment Step 3) Align remaining words Profile is frozen Gaps inserted in word only

  22. Step 4) Segmentation 6 Hungarian words from a real alignment -----k----ö---z-----ö-------t-----------t------- -----k----ö---z-----ö-------t-----------t----i-- -----k----ö---z-----ö-------t-----------t----i-t -----k----ö---z-----ö-------t-----------t----e-- -----k----ö---z-----ö-------t-----------t----e-m -----k----ö---t-----ö-------t-----------t----e-m Where are the morpheme boundaries?

  23. Step 4) Segmentation 6 Hungarian words from a real alignment -----k----ö---z-----ö-------t-----------t------- -----k----ö---z-----ö-------t-----------t----i-- -----k----ö---z-----ö-------t-----------t----i-t -----k----ö---z-----ö-------t-----------t----e-- -----k----ö---z-----ö-------t-----------t----e-m -----k----ö---t-----ö-------t-----------t----e-m Where are the morpheme boundaries? Gaps do not correspond to morpheme boundaries Biologists don’t segment!!

  24. Step 4) Segmentation Mimic the ParaMor-Morfessor Union! Take ParaMor-Morfessor Union as THE TRUTH Greedy search For each column, c, in profile Segment all words at c Score against Union system Keep the best scoring segmentation column Repeat until no column improves score

  25. Turkish Linguistic Competition Results AUTHOR METHOD PREC. REC. F1 Monson et al. ParaMor-Morfessor Mimic 48.07% 60.39% 53.53% Monson et al. ParaMor-Morfessor Union 47.25% 60.01% 52.88% Monson et al. ParaMorMimic 49.54% 54.77% 52.02% Lavallée & Langlais RALI-COF 48.43% 44.54% 46.40% - Morfessor CatMAP 79.38% 31.88% 45.49% Spiegler et al. PROMODES 2 35.36% 58.70% 44.14% Spiegler et al. PROMODES 32.22% 66.42% 43.39% Bernhard MorphoNet 61.75% 30.90% 41.19% Can & Manandhar 2 41.39% 38.13% 39.70% Spiegler et al. PROMODES committee 55.30% 28.35% 37.48% Golénia et al. UNGRADE 46.67% 30.16% 36.64% Tchoukalov et al. MetaMorph 39.14% 29.45% 33.61% Virpioja & Kohonen Allomorfessor 85.89% 19.53% 31.82% - Morfessor Baseline 89.68% 17.78% 29.67% Lavallée & Langlais RALI-ANA 69.52% 12.85% 21.69% - letters 8.66% 99.13% 15.93% Can & Manandhar 1 73.03% 8.89% 15.86%

  26. Performance Before Profile is Frozen F 1 10 20 30 40 50 60 0 Multiple Sequence Alignment 5K ParaMor-Morfessor Union Multiple Sequence Alignment 10K M ParaMor-Morfessor Union Multiple Sequence Alignment 20K ParaMor-Morfessor Union

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

Improving Morphology Induction with Spelling Rules Jason Naradowsky University of Massachusetts

Improving Morphology Induction with Spelling Rules Jason Naradowsky University of Massachusetts

Improving Morphology Induction with Spelling Rules Jason Naradowsky University of Massachusetts Amherst narad@cs.umass.edu Joint Work with Sharon Goldwater Wednesday, July 15, 2009 Outline Morphology Induction Our Model

411 views • 39 slides
Induction of Multilingual Morphology with only Minimal Supervision Richard Wicentowski Computer

Induction of Multilingual Morphology with only Minimal Supervision Richard Wicentowski Computer

Introduction Task Definition Contextual Similarity Model Combination Induction of Multilingual Morphology with only Minimal Supervision Richard Wicentowski Computer Science Department Swarthmore College November 15, 2006 Introduction Task

816 views • 59 slides
Induction and Its Applications Example for Regular Induction: Correctness of a Decimal-

Induction and Its Applications Example for Regular Induction: Correctness of a Decimal-

10/3/2016 Lecture Outline Review of Induction and Strong Induction Example: A theorem about the first n natural numbers. CSE373: Data Structures and Algorithms Example for Strong Induction: Making Postage Induction and Its

498 views • 4 slides
Induction Stepwise induction (for T PA , T cons ) Complete induction (for T PA , T cons )

Induction Stepwise induction (for T PA , T cons ) Complete induction (for T PA , T cons )

Induction Stepwise induction (for T PA , T cons ) Complete induction (for T PA , T cons ) 4. Induction Theoretically equivalent in power to stepwise induction, but sometimes produces more concise proof Well-founded induction

382 views • 7 slides
Beyond Inductive Definitions Induction-Recursion, Induction-Induction, Coalgebras Anton

Beyond Inductive Definitions Induction-Recursion, Induction-Induction, Coalgebras Anton

Beyond Inductive Definitions Induction-Recursion, Induction-Induction, Coalgebras Anton Setzer Swansea University, Swansea UK 1 March 2012 1/ 26 A Proof Theoretic Programme Sets in Martin-L of Type Theory Induction-Recursion,

295 views • 26 slides
Update on morphology WP activities M. Huertas-Company (GAL-SWG - morphology) EUCLID France - 7

Update on morphology WP activities M. Huertas-Company (GAL-SWG - morphology) EUCLID France - 7

Update on morphology WP activities M. Huertas-Company (GAL-SWG - morphology) EUCLID France - 7 Janvier 2016 Morphology WP in a nutshell Legacy Galaxies WP Provide? Request? shape / morphology measurements for EUCLID galaxies France

411 views • 18 slides
MA THEMA TICAL INDUCTION Induction and Deduction Mathematical Induction (its

MA THEMA TICAL INDUCTION Induction and Deduction Mathematical Induction (its

MA THEMA TICAL INDUCTION Induction and Deduction Mathematical Induction (its strength) Examples { Sum of rst n naturual numb ers Prove b y induction that the sum of the rst n natural numb ers is

756 views • 34 slides
Sequence Alignment (chapter 6) The biological problem l Global alignment l Local alignment l

Sequence Alignment (chapter 6) The biological problem l Global alignment l Local alignment l

Sequence Alignment (chapter 6) The biological problem l Global alignment l Local alignment l Multiple alignment l Introduction to bioinformatics, Autumn 2006 22 Background: comparative genomics Basic question in biology: what properties

934 views • 44 slides
Lexical Phonology and Morphology February 4, 2016 Lexical Phonology and Morphology Paul

Lexical Phonology and Morphology February 4, 2016 Lexical Phonology and Morphology Paul

Lexical Phonology and Morphology Lexical Phonology and Morphology February 4, 2016 Lexical Phonology and Morphology Paul Kiparsky: early 1980s 1 Developing work by Dorothy Siegel, Steven Strauss, Mark Aronoff, David Pesetsky. 2 A theory of many

522 views • 22 slides
Discrete Morphology and Distances on graphs Jean Cousty Four-Day Course on Mathematical

Discrete Morphology and Distances on graphs Jean Cousty Four-Day Course on Mathematical

Discrete Morphology and Distances on graphs Jean Cousty Four-Day Course on Mathematical Morphology in image analysis Bangalore 19-22 October 2010 J. Serra, J. Cousty, B.S. Daya Sagar : Course on Math. Morphology 1/34 Mathematical Morphology

1.1k views • 95 slides
Mathematical Induction Lecture 10-11 Menu Mathematical Induction Strong Induction

Mathematical Induction Lecture 10-11 Menu Mathematical Induction Strong Induction

Mathematical Induction Lecture 10-11 Menu Mathematical Induction Strong Induction Recursive Definitions Structural Induction Climbing an Infinite Ladder Suppose we have an infinite ladder: 1. We can reach the first rung of the

448 views • 31 slides
Introduction to English Linguistics 3: Morphology and Word Formation Part I: Morphology Part II:

Introduction to English Linguistics 3: Morphology and Word Formation Part I: Morphology Part II:

Introduction to English Linguistics 3: Morphology and Word Formation Part I: Morphology Part II: Word Formation Morphology morpheme (meaning-carrying) allomorph (meaningless variant) morph (concrete form) Systems and Nomenclature

507 views • 38 slides
Introduction to English Linguistics 3: Morphology and Word Formation Part I: Morphology Part II:

Introduction to English Linguistics 3: Morphology and Word Formation Part I: Morphology Part II:

Introduction to English Linguistics 3: Morphology and Word Formation Part I: Morphology Part II: Word Formation Morphology morpheme (meaning-carrying) allomorph (meaningless variant) morph (concrete form) Systems and Nomenclature

714 views • 36 slides
What is text alignment? Text alignment is the comparison of two or more parallel texts It

What is text alignment? Text alignment is the comparison of two or more parallel texts It

A tool for syntax-based intra-language text alignment Tariq Yousef, Chiara Palladino University of Leipzig Berlin Digital Classicist Seminars, November 29, 2016 What is text alignment? Text alignment is the comparison of two or more

688 views • 53 slides
Morphology and Corpora: Introduction Marco Baroni University of Bologna Granada Morphology

Morphology and Corpora: Introduction Marco Baroni University of Bologna Granada Morphology

Morphology and Corpora: Introduction Marco Baroni University of Bologna Granada Morphology and Corpora Seminar Outline Corpora General overview Data sparseness and the need for larger corpora Morphology Derivational vs. inflectional

746 views • 63 slides
Foundations of Computer Science Lecture 5 Induction: Proving For All . . . Induction:

Foundations of Computer Science Lecture 5 Induction: Proving For All . . . Induction:

Foundations of Computer Science Lecture 5 Induction: Proving For All . . . Induction: What and Why? Induction: Good, Bad and Ugly Induction, Well-Ordering and the Smallest Counter-Example Last Time 1 Proving if . . . , then . . .

391 views • 18 slides
MOL2NET, 2017 , 3, doi:10.3390/mol2net-03-xxxx 2 collected water samples for DNA extraction.

MOL2NET, 2017 , 3, doi:10.3390/mol2net-03-xxxx 2 collected water samples for DNA extraction.

MOL2NET, 2017 , 3, doi:10.3390/mol2net-03-xxxx 1 MDPI MOL2NET, International Conference Series on Multidisciplinary Sciences http://sciforum.net/conference/mol2net-03 Analysis of Microbial Communities Reflect Diel Vertical Migration in the Gulf

260 views • 3 slides
RNA Structure modeling GDR MASIM Paris, 16-17 th November 2017 Bruno Sargueil, CNRS UMR 8015

RNA Structure modeling GDR MASIM Paris, 16-17 th November 2017 Bruno Sargueil, CNRS UMR 8015

RNA Structure modeling GDR MASIM Paris, 16-17 th November 2017 Bruno Sargueil, CNRS UMR 8015 Facult de pharmacie PARIS Bruno.sargueil@parisdescartes.fr RNA structure Primary structure 5

401 views • 24 slides
Parallel Adaptations to High Temperatures in the Archean Eon Samuel Blanquart a 1 Bastien Boussau

Parallel Adaptations to High Temperatures in the Archean Eon Samuel Blanquart a 1 Bastien Boussau

Parallel Adaptations to High Temperatures in the Archean Eon Samuel Blanquart a 1 Bastien Boussau b 1 Anamaria sulea b Nicolas Lartillot a Manolo Gouy b Nec June 9, 2008 a LIRMM, CNRS. b BBE, CNRS, Universit e de Lyon I. 1 These authors

682 views • 39 slides
Two added structures in sparse recovery: nonnegativity and disjointedness Simon Foucart

Two added structures in sparse recovery: nonnegativity and disjointedness Simon Foucart

Two added structures in sparse recovery: nonnegativity and disjointedness Simon Foucart University of Georgia Semester Program on High-Dimensional Approximation ICERM 7 October 2014 Part I: Nonnegative Sparse Recovery (joint work with

957 views • 69 slides
The RISC-V Processor Hakim Weatherspoon CS 3410 Computer Science Cornell University

The RISC-V Processor Hakim Weatherspoon CS 3410 Computer Science Cornell University

The RISC-V Processor Hakim Weatherspoon CS 3410 Computer Science Cornell University [Weatherspoon, Bala, Bracy, and Sirer] Announcements Check online syllabus/schedule http://www.cs.cornell.edu/Courses/CS3410/2019sp/schedule Slides

1.81k views • 95 slides
Supporting 64 bit pointers in RISCV 32 bit LLVM backend Reshabh Sharma Background: Prof.

Supporting 64 bit pointers in RISCV 32 bit LLVM backend Reshabh Sharma Background: Prof.

Supporting 64 bit pointers in RISCV 32 bit LLVM backend Reshabh Sharma Background: Prof. Taylors Bespoke Silicon Group is developing a GP-GPU based on RISC-V 32 bit ISA (RV32) Why 32 bit ISA? It is good for very high energy efficiency and

557 views • 13 slides
This Unit: Single-Cycle Datapath App App App Datapath storage elements System software

This Unit: Single-Cycle Datapath App App App Datapath storage elements System software

This Unit: Single-Cycle Datapath App App App Datapath storage elements System software MIPS Datapath CIS 371 MIPS Control Mem CPU I/O Computer Organization and Design Unit 4: Single-Cycle Datapath Based on slides by Prof. Amir

287 views • 14 slides
CMPSC 497: Software Fault Isolation Trent Jaeger Systems and Internet Infrastructure

CMPSC 497: Software Fault Isolation Trent Jaeger Systems and Internet Infrastructure

CMPSC 497: Software Fault Isolation Trent Jaeger Systems and Internet Infrastructure Security (SIIS) Lab Computer Science and Engineering Department Pennsylvania State University Systems and Internet Infrastructure Security Laboratory

705 views • 34 slides

More recommend