1. Page 1 Learning and Global Inference for Information Access and Natural Language Understanding Dan Roth Department of Computer Science University of Illinois at Urbana-Champaign May 2007 DIMACS ONR Workshop

2. Page 2 Nice to Meet You

3. Page 3 Learning and Inference  Global decisions in which several local decisions play a role but there are mutual dependencies on their outcome.  (Learned) models/classifiers for different sub-problems  Incorporate classifiers’ information, along with constraints, in making coherent decisions – decisions that respect the local models as well as domain & context specific constraints.  Global inference for the best assignment to all variables of interest.

4. Page 4 Inference

5. Page 5 Comprehension 1. Christopher Robin was born in England. 2. Winnie the Pooh is a title of a book. 3. Christopher Robin’s dad was a magician. 4. Christopher Robin must be at least 65 now. A process that maintains and updates a collection of propositions about the state of affairs. (ENGLAND, June, 1989) - Christopher Robin is alive and well. He lives in England. He is the same person that you read about in the book, Winnie the Pooh. As a boy, Chris lived in a pretty home called Cotchfield Farm. When Chris was three years old, his father wrote a poem about him. The poem was printed in a magazine for others to read. Mr. Robin then wrote a book. He made up a fairy tale land where Chris lived. His friends were animals. There was a bear called Winnie the Pooh. There was also an owl and a young pig, called a piglet. All the animals were stuffed toys that Chris owned. Mr. Robin made them come to life with his words. The places in the story were all near Cotchfield Farm. Winnie the Pooh was written in 1925. Children still love to read about Christopher Robin and his animal friends. Most people don't know he is a real person who is grown now. He has written two books of his own. They tell what it is like to be famous.

6. Page 6 Illinois’ bored of education board...Nissan Car and truck plant is … …divide life into plant and animal kingdom (This Art) (can N) (will MD) (rust V) V,N,N The dog bit the kid. He was taken to a veterinarian a hospital What we Know: Stand Alone Ambiguity Resolution Learn a function f: X  Y that maps observations in a domain to one of several categories or < Broad Coverage

7. Page 7 Comprehension 1. Christopher Robin was born in England. 2. Winnie the Pooh is a title of a book. 3. Christopher Robin’s dad was a magician. 4. Christopher Robin must be at least 65 now. A process that maintains and updates a collection of propositions about the state of affairs. (ENGLAND, June, 1989) - Christopher Robin is alive and well. He lives in England. He is the same person that you read about in the book, Winnie the Pooh. As a boy, Chris lived in a pretty home called Cotchfield Farm. When Chris was three years old, his father wrote a poem about him. The poem was printed in a magazine for others to read. Mr. Robin then wrote a book. He made up a fairy tale land where Chris lived. His friends were animals. There was a bear called Winnie the Pooh. There was also an owl and a young pig, called a piglet. All the animals were stuffed toys that Chris owned. Mr. Robin made them come to life with his words. The places in the story were all near Cotchfield Farm. Winnie the Pooh was written in 1925. Children still love to read about Christopher Robin and his animal friends. Most people don't know he is a real person who is grown now. He has written two books of his own. They tell what it is like to be famous. This is an Inference Problem

8. Page 8 This Talk  Global Inference over Local Models/Classifiers + Expressive Constraints  Model  Generality of the framework  Training Paradigms  Global vs. Local training  Semi-Supervised Learning  Examples  Semantic Parsing  Information Extraction  Pipeline processes

9. Page 9 Random Variables Y: Conditional Distributions P (learned by models/classifiers) Constraints C– any Boolean function defined on partial assignments (possibly: + weights W ) Goal: Find the “best” assignment  The assignment that achieves the highest global accuracy. This is an Integer Programming Problem Problem Setting y4y4 y5y5 y6y6 y7y7 y8y8 y1y1 y2y2 y3y3 C(y 1,y 4 ) C(y 2,y 3,y 6,y 7,y 8 ) Y*=argmax Y P  Y subject to constraints C (+ W  C) Other, more general ways to incorporate soft constraints here [ACL’07]

10. Page 10 Formal Model How to solve? This is an Integer Linear Program In many of our applications, large scale problems were solved efficiently using a commercial ILP package to yield exact solution. Search techniques are also possible (Soft) constraints component Weight Vector for “local” models Penalty for violating the constraint. How far away is y from a “legal” assignment Subject to constraints A collection of Classifiers; Log-linear models (HMM, CRF) or a combination

11. Page 11 A General Inference Setting Essentially all complex models studied today can be viewed as optimizing a linear objective function: HMMs/CRFs [Roth’99; Collins’02;Laffarty et. al 02] Linear objective functions can be derived from probabilistic perspective:  Markov Random Field  [standard; Kleinberg&Tardos] Optimization Problem (Metric Labeling)  [Chekuri et. al ’ 01] Linear Programming Problems  Inference as Constrained Optimization [Yih&Roth CoNLL ’ 04] [Punyakanok et. al COLING ’ 04] … The probabilistic perspective supports finding the most likely assignment  Not necessarily what we want Our Integer linear programming (ILP) formulation  Allows the incorporation of more general cost functions  General (non-sequential) constraint structure  Better exploitation (computationally) of hard constraints  Can find the optimal solution if desired

12. Page 12 Example 1: Semantic Role Labeling I left my pearls to my daughter in my will. [ I ] A0 left [ my pearls ] A1 [ to my daughter ] A2 [ in my will ] AM-LOC. A0Leaver A1Things left A2Benefactor AM-LOCLocation I left my pearls to my daughter in my will. Special Case (structure output problem): here, all the data is available at one time; in general, classifiers might be learned from different sources, at different times, at different contexts. Implications on training paradigms Overlapping arguments If A2 is present, A1 must also be present. Who did what to whom, when, where, why,…

13. Page 13 PropBank [Palmer et. al. 05] provides a large human-annotated corpus of semantic verb-argument relations.  It adds a layer of generic semantic labels to Penn Tree Bank II.  (Almost) all the labels are on the constituents of the parse trees. Core arguments: A0-A5 and AA  different semantics for each verb  specified in the PropBank Frame files 13 types of adjuncts labeled as AM- arg  where arg specifies the adjunct type Semantic Role Labeling (2/2)

14. Page 14 Algorithmic Approach Identify argument candidates  Pruning [Xue&Palmer, EMNLP’04]  Argument Identifier Binary classification (SNoW) Classify argument candidates  Argument Classifier Multi-class classification (SNoW) Inference  Use the estimated probability distribution given by the argument classifier  Use structural and linguistic constraints  Infer the optimal global output I left my nice pearls to her [ [ [ [ [ ] ] ] ] ] I left my nice pearls to her [ [ [ [ [ ] ] ] ] ] I left my nice pearls to her Identify Vocabulary Inference over (old and new) Vocabulary candidate arguments

15. Page 15 Argument Identification & Classification Both argument identifier and argument classifier are trained phrase-based classifiers. Features (some examples)  voice, phrase type, head word, path, chunk, chunk pattern, etc. [some make use of a full syntactic parse] Learning Algorithm – SNoW  Sparse network of linear functions weights learned by regularized Winnow multiplicative update rule  Probability conversion is done via softmax p i = exp{act i }/  j exp{act j } I left my nice pearls to her [ [ [ [ [ ] ] ] ] ] I left my nice pearls to her

16. Page 16 Inference I left my nice pearls to her The output of the argument classifier often violates some constraints, especially when the sentence is long. Finding the best legitimate output is formalized as an optimization problem and solved via Integer Linear Programming. [Punyakanok et. al 04, Roth & Yih 04] Input:  The probability estimation (by the argument classifier)  Structural and linguistic constraints Allows incorporating expressive (non-sequential) constraints on the variables (the arguments types).

17. Page 17 Integer Linear Programming Inference For each argument a i  Set up a Boolean variable: a i,t indicating whether a i is classified as t Goal is to maximize   i score(a i = t ) a i,t  Subject to the (linear) constraints Any Boolean constraints can be encoded as linear constraint(s). If score(a i = t ) = P(a i = t ), the objective is to find the assignment that maximizes the expected number of arguments that are correct and satisfies the constraints.

18. Page 18 No duplicate argument classes  a  P OT A RG x { a = A0 }  1 R-ARG  a2  P OT A RG,  a  P OT A RG x { a = A0 }  x { a2 = R-A0 } C-ARG  a2  P OT A RG,  (a  P OT A RG )  (a is before a2 ) x { a = A0 }  x { a2 = C-A0 } Many other possible constraints: Unique labels No overlapping or embedding Relations between number of arguments If verb is of type A, no argument of type B Any Boolean rule can be encoded as a linear constraint. If there is an R-ARG phrase, there is an ARG Phrase If there is an C-ARG phrase, there is an ARG before it Constraints Joint inference can be used also to combine different SRL Systems. Universally quantified rules

19. Page 19 This approach produces a very good semantic parser. F 1 ~<90% Easy and fast: ~7 Sentences/Second (using Xpress-MP) A lot of room for improvement (additional constraints) Demo available http://L2R.cs.uiuc.edu/~cogcomphttp://L2R.cs.uiuc.edu/~cogcomp Semantic Parsing: Summary I Top ranked system in CoNLL’05 shared task Key difference is the Inference

20. Page 20 Extracting Relations via Semantic Analysis Screen shot from a CCG demo http://L2R.cs.uiuc.edu/~cogcomp http://L2R.cs.uiuc.edu/~cogcomp  Semantic parsing reveals several relations in the sentence along with their arguments.

21. Page 21 This approach produces a very good semantic parser. F 1 ~<90% Easy and fast: ~7 Sentences/Second (using Xpress-MP) A lot of room for improvement (additional constraints) Demo available http://L2R.cs.uiuc.edu/~cogcomphttp://L2R.cs.uiuc.edu/~cogcomp So far, shown the use of only declarative (deterministic) constraints. In fact, this approach can be used both with statistical and declarative constraints. Semantic Parsing: Summary I Top ranked system in CoNLL’05 shared task Key difference is the Inference

22. Page 22 ILP as a Unified Algorithmic Scheme Consider a common model for sequential inference: HMM/CRF  Inference in this model is done via the Viterbi Algorithm. Viterbi is a special case of the Linear Programming based Inference.  Viterbi is a shortest path problem, which is a LP, with a canonical matrix that is totally unimodular. Therefore, you can get integrality constraints for free.  One can now incorporate non-sequential/expressive/declarative constraints by modifying this canonical matrix  The extension reduces to a polynomial scheme under some conditions (e.g., when constraints are sequential, when the solution space does not change, etc.)  Not necessarily increases complexity and very efficient in practice [Roth&Yih, ICML’05] y1y1 y2y2 y3y3 y4y4 y5y5 y x x1x1 x2x2 x3x3 x4x4 x5x5 s A B C A B C A B C A B C A B C t Learn a rather simple model; make decisions with a more expressive model

23. Page 23 An Inference method for the “best explanation”, used here to induce a semantic representation of a sentence.  A general Information Integration framework. Allows expressive constraints  Any Boolean rule can be represented by a set of linear (in)equalities Combining acquired (statistical) constraints with declarative constraints  Start with shortest path matrix and constraints  Add new constraints to the basic integer linear program. Solved using off-the-shelf packages  If the additional constraints don’t change the solution, LP is enough  Otherwise, the computational time depends on sparsity; fast in practice Demo available http://L2R.cs.uiuc.edu/~cogcomphttp://L2R.cs.uiuc.edu/~cogcomp Integer Linear Programming Inference - Summary

24. Page 24 Pipelining is a crude approximation; interactions occur across levels and down stream decisions often interact with previous decisions. Leads to propagation of errors Occasionally, later stage problems are easier but upstream mistakes will not be corrected. There are good reasons for pipelining decisions Global inference over the outcomes of different levels can be used to break away from this paradigm. [between pipeline & fully global] Allows a flexible way to incorporate linguistic and structural constraints. POS TaggingPhrasesSemantic EntitiesRelations Vocabulary is generated in phases Left to Right processing of sentences is also a pipeline process ParsingWSDSemantic Role Labeling Raw Data Example 2: Pipeline Objective Function can be modified to support pipelines; Deep Pipelines for Dependency Parsing [Chang, Do, Roth, ACL’06]

25. Page 25 This Talk  Global Inference over Local Models/Classifiers + Expressive Constraints  Model  Generality of the framework  Training Paradigms  Global vs. Local training  Semi-Supervised Learning  Examples  Semantic Parsing  Information Extraction  Pipeline processes

26. Page 26 Training Paradigms that Support Global Inference Incorporating general constraints (Algorithmic Approach)  Allow both statistical and expressive declarative constraints  Allow non-sequential constraints (generally difficult) Coupling vs. Decoupling Training and Inference.  Incorporating global constraints is important but  Should it be done only at evaluation time or also at training time?  Issues related to: modularity, efficiency and performance, availability of training data May not be relevant in some problems.

27. Page 27 Training in the presence of Constraints General Training Paradigm:  First Term: Learning from data  Second Term: Lead the model by constraints  Can choose if constraints are included in training or only in evaluation

28. Page 28 L+I: Learning plus Inference IBT: Inference-based Training Training w/o Constraints Testing: Inference with Constraints x1x1 x6x6 x2x2 x5x5 x4x4 x3x3 x7x7 y1y1 y2y2 y5y5 y4y4 y3y3 f1(x)f1(x) f2(x)f2(x) f3(x)f3(x) f4(x)f4(x) f5(x)f5(x) X Y Learning the components together! Which one is better? When and Why? Cartoon: each model can be more complex and may have a view on a set of output variables.

29. Page 29 1111Y’ Local Predictions Perceptron-based Global Learning x1x1 x6x6 x2x2 x5x5 x4x4 x3x3 x7x7 f1(x)f1(x) f2(x)f2(x) f3(x)f3(x) f4(x)f4(x) f5(x)f5(x) X Y 11 Y True Global Labeling 111 Y’ Apply Constraints:

30. Page 30 Claims When the local classification problems are “ easy ”, L+I outperforms IBT.  In many applications, the components are identifiable and easy to learn (e.g., argument, open-close, PER). Only when the local problems become difficult to solve in isolation, IBT outperforms L+I, but needs a larger number of training examples.  When data is scarce, problems are not easy and constraints can be used, along with a “ weak ” model, to label unlabeled data and improve mode.  Will show experimental results and theoretical intuition to support our claims. L+I: cheaper computationally; modular IBT is better in the limit, and other extreme cases. Combinations: L+I, and then IBT are possible

31. Page 31  opt =0.2  opt =0.1  opt =0 Bound Prediction Local  ≤  opt + ( ( d log m + log 1/  ) / m ) 1/2 Global  ≤ 0 + ( ( cd log m + c 2 d + log 1/  ) / m ) 1/2 BoundsSimulated Data L+I vs. IBT: the more identifiable individual problems are the better overall performance is with L+I Indication for hardness of problem

32. Page 32 Relative Merits: SRL Difficulty of the learning problem (# features) L+I is better. When the problem is artificially made harder, the tradeoff is clearer. easyhard

33. Page 33 Experiment is done in the context of Information extraction Objective function: Semi-Supervised Learning with Constraints # of available labeled examples Learning w Constraints Constraints are used to Bootstrap a semi- supervised learner Poor model + constraints are used to annotate unlabeled data, which in turn is used to keep training the model. Learning w/o Constraints: 300 examples.

34. Page 34 Given: Q: Who acquired Overture? Determine: A: Eyeing the huge market potential, currently led by Google, Yahoo took over search company Overture Services Inc last year. Example 3: Textual Entailment Eyeing the huge market potential, currently led by Google, Yahoo took over search company Overture Services Inc. last year Yahoo acquired Overture Entails Subsumed by  Overture is a search company Google is a search company ………. Google owns Overture By “semantically entailed” we mean: most people would agree that one sentence implies the other. Phrasal verb paraphrasing [Connor&Roth’06] Entity matching [Li et. al, AAAI’04, NAACL’04] Semantic Role Labeling

35. Page 35 Discussed a general paradigm for learning and inference in the context of natural language understanding tasks A general Constraint Optimization approach for integration of learned models with additional (declarative or statistical) expressivity How to train?  Learn Locally and make use globally (via global inference) [Punyakanok et. al IJCAI’05]  Ability to make use of domain & constraints to drive supervision [Klementiev & Roth, ACL’06; Chang, Do, Roth, ACL’07] How to drive component identification from the global decisoins? Conclusions LBJ (Learning Based Java): A modeling language that supports programming along with building learned models and allows the incorporation of and inference with constraints