Textual Entailment. Arindam Bhattacharya. M.Tech, Computer Science Indian Institute of Technology, Bombay. November 9, 2011

Size: px
Start display at page:

Download "Textual Entailment. Arindam Bhattacharya. M.Tech, Computer Science Indian Institute of Technology, Bombay. November 9, 2011"

Transcription

1 Textual Entailment Arindam Bhattacharya M.Tech, Computer Science Indian Institute of Technology, Bombay November 9, 2011 Arindam (IITB) Textual Entailment November 9, / 59

2 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

3 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

4 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

5 Definition Classical Definition A text t entails a hypothesis h if h is true in every circumstance (possible world) in which t is true. Strict Entailment! Doesn t account for real world uncertainties. Example: T: Ram was born and brought up in Maharashtra. H: Ram can speak Marathi. Applied Definition t entails h (t h) if humans reading t will infer that h is most likely true. Arindam (IITB) Textual Entailment November 9, / 59

6 Definition Classical Definition A text t entails a hypothesis h if h is true in every circumstance (possible world) in which t is true. Strict Entailment! Doesn t account for real world uncertainties. Example: T: Ram was born and brought up in Maharashtra. H: Ram can speak Marathi. Applied Definition t entails h (t h) if humans reading t will infer that h is most likely true. Arindam (IITB) Textual Entailment November 9, / 59

7 Probabilistic Interpretation Applied definition sounds good. But doesn t sound concrete of mathematical. Probabilistic interpretation t probabilistically entails h if: P(h is true t) > P(h is true) P(h is true t) is called Entailment Confidence. Arindam (IITB) Textual Entailment November 9, / 59

8 Probabilistic Interpretation Applied definition sounds good. But doesn t sound concrete of mathematical. Probabilistic interpretation t probabilistically entails h if: P(h is true t) > P(h is true) P(h is true t) is called Entailment Confidence. Arindam (IITB) Textual Entailment November 9, / 59

9 Goal Figure: Textual Entailment Arindam (IITB) Textual Entailment November 9, / 59

10 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

11 Entailment Triggers Triggers are linguistic features that affect entailment [?]. Here are some examples to show how these various factors affect entailment. Synonymy: Very common form of entailment trigger, where a word is replaced by its synonym. T: World War I began in H: World War I started in Arindam (IITB) Textual Entailment November 9, / 59

12 Entailment Triggers Hypernymy/Hyponymy: Certain concept can be either generalized or specialized, leading to entailment. T: Reptiles have scale. H: Snakes have scale. (Specialization or Hyponymy) T: Beckham plays football. H: Beckham plays a game. (Generalization or Hypernymy) Arindam (IITB) Textual Entailment November 9, / 59

13 Entailment Triggers Co-reference: One of the main sources for text entailment. Especially with long text containing paragraphs! T: Barrack Obama came to India. The American President had a meeting with Manmahon Singh. H: Barrack Obama had a meeting with Manmahon Singh. Arindam (IITB) Textual Entailment November 9, / 59

14 Entailment Triggers Modality/Polarity/Factive: Plays critical role in entailment as they affect the degree of reliability on the remaining sentence. Especially troublesome for lexical approaches. Modality denotes possibility or necessity and sometimes may lead to wrong entailment. e.g. may, can, shall, must etc. are modality triggers. T: The government may approve the anti-corruption bill. H: The government approved the anti-corruption bill. Arindam (IITB) Textual Entailment November 9, / 59

15 Entailment Triggers Polarity determines whether the fact asserted or its negation is going to occur. e.g. not, never, deny etc. are polarity triggers. T: The watchman denied that he was sleeping. H: The watchman was sleeping. Factivity deals with presupposition. It states a fact assuming another has occurred. e.g. realize, regret etc. are factivity triggers. T: Martha regrets eating John s homemade cake. H: Martha ate John s homemade cake. Arindam (IITB) Textual Entailment November 9, / 59

16 Entailment Triggers Passivization: In some case one of the text or hypothesis was is in active while the other is in passive. Subject and object of the main verb gets reversed. Can only be handled by assigning semantic roles to each entity. T: Yahoo bought Overture. H: Overture was bought by Yahoo. Arindam (IITB) Textual Entailment November 9, / 59

17 Entailment Triggers Dropping or Inserting Adjunct: Adding or dropping adjuncts affect entailment based on which of T or H is modified, and the polarity. T: Bob was running quickly. H: Bob was running. T: Carl was eating. H: Carl was eating slowly. [Incorrect entailment] T: Alice was not driving. H: Alice was not driving fast. T: Derek was not writing properly. H: Derek was not writing. [Incorrect entailment] Arindam (IITB) Textual Entailment November 9, / 59

18 Entailment Triggers Protocols: Some common conventions such as mentioning birth-death year may trigger entailment. T: Charles de Gaulle, , French general and statesman, was the first president of the Fifth Republic. H: Charles de Gaulle died in Arindam (IITB) Textual Entailment November 9, / 59

19 Entailment Triggers Numerals: In some cases, certain level of numeric calculation affects entailment. T: 3 men and 2 women were found dead in the apartment. H: 5 people were found dead in an apparent. Arindam (IITB) Textual Entailment November 9, / 59

20 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

21 Role of Knowledge Background knowledge is crucial in entailment as in any AI application! Example T: President of Russia visited Paris. H: President of Russia visited France. B: Paris is situated in France. Background knowledge B alone should not entail the hypothesis H and text T must contain necessary information (may not be sufficient). (T B) = H but B H Arindam (IITB) Textual Entailment November 9, / 59

22 Role of Knowledge Background knowledge is crucial in entailment as in any AI application! Example T: President of Russia visited Paris. H: President of Russia visited France. B: Paris is situated in France. Background knowledge B alone should not entail the hypothesis H and text T must contain necessary information (may not be sufficient). (T B) = H but B H Arindam (IITB) Textual Entailment November 9, / 59

23 Role of Knowledge Background knowledge is crucial in entailment as in any AI application! Example T: President of Russia visited Paris. H: President of Russia visited France. B: Paris is situated in France. Background knowledge B alone should not entail the hypothesis H and text T must contain necessary information (may not be sufficient). (T B) = H but B H Arindam (IITB) Textual Entailment November 9, / 59

24 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

25 Recognizing Textual Entailment Challenges Goal The recognizing textual entailment is an attempt to promote an abstract generic task that captures major semantic inference needs across applications. Held every year starting RTE - 1,2 and 3 organized by PASCAL 1. Organized by Text Analysis Conference (TAC) since then. Shifted focus to real world applications since RTE-5 (2009) rather than T-H pair entailment recognition. 1 Pattern Analysis, Statistical Modeling and Computational Learning Arindam (IITB) Textual Entailment November 9, / 59

26 RTE Main Task: Given a corpus and a set of candidate sentences retrieved by Lucene from that corpus, RTE systems are required to identify all the sentences from among the candidate sentences that entail a given Hypothesis. each topic contains two sets of documents ( A and B ) Corpus is the set A and H is a sentence taken from B Arindam (IITB) Textual Entailment November 9, / 59

27 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

28 Resources used for Textual Entailment Resource Type Author Brief Description WordNet Verbnet Roget s Thesaurus Lexical DB Lexical DB Univer- Princeton sity University of Colorado, Boulder DB of nouns, verbs, adjectives and adverbs Lexicon for English verbs organized into classes extending Levin (1993) classes through refinement and addition of subclasses to achieve syntactic and semantic coherence among members of a class. Thesaurus Peter Mark Roget Roget s Thesaurus is a widely-used English thesaurus. The electronic edition (version 1.02) is made available by University of Chicago. Table: Knowledge Resources Arindam (IITB) Textual Entailment November 9, / 59

29 Resources used for Textual Entailment Resource Type Author Brief Description DIRT Paraphrase Collection Collection of paraphrases TEASE Collection Collection of Entailment Rules University of Alberta Bar-Ilan University Table: Knowledge Resources DIRT (Discovery of Inference Rules from Text) knowledge collection of paraphrases from over a 1GB set of newspaper text. Output of the TEASE algorithm. Collection of several entailment templates from web resources. Arindam (IITB) Textual Entailment November 9, / 59

30 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

31 Sub-tasks Recognizing textual entailment requires various sub-tasks such as: Phrasal Verb Recognition Named Entity Recognition Semantic Role Labeling An example illustrates the need for these tasks Arindam (IITB) Textual Entailment November 9, / 59

32 Example Arindam (IITB) Textual Entailment November 9, / 59

33 General Strategy A general two-step strategy involves: 1 Representation of the information into a form that can be used by the entailment algorithm 2 Entailment Recognition Algorithm that matches the text T along with knowledge B with hypothesis H Arindam (IITB) Textual Entailment November 9, / 59

34 Representation Logical Semantic Syntactic Lexical Raw Text T Re-representation φ(t ) Figure: Various Representations The complexity of representation increases as we go higher. Arindam (IITB) Textual Entailment November 9, / 59

35 Entailment Recognition Text φ(t ) φ(h) Hypothesis e? Y/N φ(b) Knowledge Base Figure: General Strategy e checks if the degree of subsumption of φ(h) with φ(t ) and φ(b) is over a certain threshold e Arindam (IITB) Textual Entailment November 9, / 59

36 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

37 Lexical Approaches Shallow Approaches: operate on surface level Carries out some basic preprocessing Does not compute elaborate representations Make the entailment decision solely based on the lexical evidences Arindam (IITB) Textual Entailment November 9, / 59

38 Preprocessing Surface preprocessing includes: tokenization stemming/lemmatization identifying the stop words Some systems does a bit deeper preprocessing such as: Phrasal Verb Recognition e.g. take off, put on Idiom processing e.g. A Picture Paints a Thousand Words Named Entity Recognition and Normalization Arindam (IITB) Textual Entailment November 9, / 59

39 Representation Lexical approaches use on of following representation Bag-of-words: Both T and H are represented as a set of words. n-grams: Sequence of n tokens are grouped together. Bag of words is an extreme case of n-gram, with n=1. Example: The fixed routine of a bedtime story before sleeping has a relaxing effect. Bag-of words: The, fixed, routine, of, a, bedtime, story, before, sleeping, has, relaxing, effect Bigram model (n-gram with n=2): The fixed, fixed routine, routine of, of a, a bedtime, bedtime story, story before, before sleeping, sleeping has, has a, a relaxing, relaxing effect Arindam (IITB) Textual Entailment November 9, / 59

40 Example: LLM Algorithm Local Lexical Matching (LLM) is a lexical approach for text entailment that uses bag of words representation INPUT: Text T and Hypothesis H. OUTPUT: The matching score. for all word in T and H do if word in stopwordlist then remove word; end if if no words left in T or H then return 0; end if end for numbermatched = 0; for all word W T in T do Lemma T = Lemmatize(W T ); for all word W H in H do Lemma H = Lemmatize(W H ); if LexicalCompare(Lemma H, Lemma T ) then numbermatched + +; end if end for end for Figure: LLM Algorithm Arindam (IITB) Textual Entailment November 9, / 59

41 LexicalCompare The LexicalCompare() procedure is checks similarity with help of WordNet. if Lemma H == Lemma T then return TRUE; end if if HypernymDistance(W H, W T ) d Hyp then return TRUE; end if if MeronymDistance(W H, W T ) d Mer then return TRUE; end if if MemberOfDistance(W H, W T ) d Mem then return TRUE; end if if SynonymOf(W H, W T ) then return TRUE; end if Figure: Lexical Compare Procedure Arindam (IITB) Textual Entailment November 9, / 59

42 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

43 Textual Entailment as a Classification Task Figure: Text Entailment as a Classification Task Arindam (IITB) Textual Entailment November 9, / 59

44 Feature Space What could be a possible feature space? Most important decision! Distance Features Features of some distance between T and H. Entailment Triggers Features that triggers entailment (or non-entailment) Syntactic Feature Syntax of T-H pair modeled to exploit rewrite rules. Arindam (IITB) Textual Entailment November 9, / 59

45 Distance Features Possible Features Number of words in common. Longest common subsequnce. Longest common syntactic subtree. Requires representation of T and H as Bag-of words or n-grams Syntactic representation Semantic Representation Arindam (IITB) Textual Entailment November 9, / 59

46 Distance Features Possible Features Number of words in common. Longest common subsequnce. Longest common syntactic subtree. Requires representation of T and H as Bag-of words or n-grams Syntactic representation Semantic Representation Arindam (IITB) Textual Entailment November 9, / 59

47 Distance Features For example: T: At the end of the year, all solid companies pay dividends. H: At the end of the year, all solid insurance companies pay dividends. The above example, possible feature, value pair could be WordsInCommon, 11 or LongestSubsequence, 8. Arindam (IITB) Textual Entailment November 9, / 59

48 Entailment Triggers [?] Capture presence of linguistic features that triggers entailment. Example T: The government may approve the anti-corruption bill. H: The government approved the anti-corruption bill. A feature, value pair could be modal, 1 Arindam (IITB) Textual Entailment November 9, / 59

49 Exploiting Re-write rules How the rewrite rules are exploited is illustrated by following example. Consider the the pair: T: Loki was killed by Thor. H: Loki died. Using the syntactic pair features we can learn rules such as: Figure: Exploiting rewrite rules Arindam (IITB) Textual Entailment November 9, / 59

50 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

51 Textual Entailment as Graph Matching Convert hypothesis and text into graphs. Either syntactic or semantic. Measure similarity. Similarity score gives entailment. Arindam (IITB) Textual Entailment November 9, / 59

52 Different from Classical Graph Matching! Scoring is not symmetric. Node similarity can not be reduced to label level (i.e token level). Consideration of linguistically motivated graph transformation (nominalization, passivization). Arindam (IITB) Textual Entailment November 9, / 59

53 Text to Graph Generation of dependency graph using a dependency parser Graph edges are labeled by hand made rules (e.g. subj, amod) Applying certain enhancements Arindam (IITB) Textual Entailment November 9, / 59

54 Enhancements to Dependency Graph Collapse Collocations and Named-Entities. Collocations: [blow] [off] [blow off] Named entities: [Micheal] [Jackson] [Micheal Jackson] Dependency Folding so that certain dependencies (such as modifying prepositions become labels). Skeleton -[in]-> cupboard. Arindam (IITB) Textual Entailment November 9, / 59

55 Enhancements to Dependency Graph Semantic Role Labeling. Arcs are labeled with Propbank style semantic roles. This helps to create links between words which share a deep semantic relation not evident in the surface syntax. e.g. Pakistan got independence in [1947] Temporal. Co-reference Links Using a co-reference resolution tagger, coref links are added throughout the graph. In the case of multiple sentence texts, it is our only link in the graph between entities in the two sentences. Arindam (IITB) Textual Entailment November 9, / 59

56 Entailment Model Entailment model determines the matching cost between graphs of T and H. The final cost is a linear combination of cost of matching vertices and edges. Cost = α VertexCost + (1 α) EdgeCost Arindam (IITB) Textual Entailment November 9, / 59

57 Additional Checks Certain additional checks can be applied to the system to improve its performance [?]. They are listed below. Negation Check: Check if there is a negation in a sentence. Example, T:Clinton s book is not a bestseller. H:Clinton s book is a bestseller. Factive Check: Non-factive verbs (claim, think, charged, etc.) in contrast to factive verbs (know, regret, etc.) have sentential complements which do not represent true propositions. T:Clonaid claims to have cloned 13 babies worldwide. H:Clonaid has cloned 13 babies. Arindam (IITB) Textual Entailment November 9, / 59

58 Additional Checks Superlative Check: invert the typical monotonicity of entailment. Example, T: The Osaka World Trade Center is the tallest building in Western Japan. H: The Osaka World Trade Center is the tallest building in Japan. Antonym Check: It is observed that the WordNet::Similarity measures gave high similarity to antonyms. Explicit check of whether a matching involved antonyms is done and unless one of the vertices had a negation modifier, its rejected. Arindam (IITB) Textual Entailment November 9, / 59

59 An Example T: In 1994, Amazon.com was founded by Jeff Bezos. H: Bezos established a company. VC = ( )/3 = 0.13 EC = 0 (isomorphic edges) Cost = (0.55) (0.13) + (0.45) (0) = (let α = 0.55) Arindam (IITB) Textual Entailment November 9, / 59

60 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

61 What is UNL? tool for representing text in terms of semantic relation between different entities consist of Universal Words (UW), Relations and Attributes Arindam (IITB) Textual Entailment November 9, / 59

62 Example Google goes public. agt obj Google(icl>organization) public(aoj>thing, ant>private) agt(go(icl>do, equ>travel, Google(icl>organization)) obj(go(icl>do, equ>travel, public(aoj>thing, ant>private)) Arindam (IITB) Textual Entailment November 9, / 59

63 Outline 1 Introduction Definition Entailment Triggers Role of Knowledge RTE Challenges Resources 2 General Strategy 3 Lexical Approach 4 Machine Learning Approach 5 Graphical Approach 6 Deep Semantic Approach Text Entailment using UNL Arindam (IITB) Textual Entailment November 9, / 59

64 The Approach [CS : Lecture 29, Prasad Pradip Joshi] Represent both text and hypothesis in their UNL form and do analysis on the UNL expressions List of atomic facts (predicates) emerging from the UNL graph of the hypothesis statement must be a subset (either explicitly or implicitly) of the atomic facts emerging from the UNL graph of the text statement The algorithm has two main parts: Extending the set of atomic truths of the text graph based on those which are present. (referred to as growth-rules) Carrying out the matching of the atomic facts in the hypothesis and the text graph (referred to as matching-rules) Arindam (IITB) Textual Entailment November 9, / 59

65 Illustration [CS : Lecture 29, Prasad Pradip Joshi] Text: Manmohan Singh along with president George Bush signed a letter in Hypothesis: Bush signed a document. Text representation agt(sign@entry@past,manmohan Singh) cag(sign@entry@past,president) nam(president,george Bush) obj(sign@entry@past,letter@indef) tim(sign@entry@past,2006) aoj(president,george Bush) cag(sign@entry@past,george Bush) Hypothesis Representation agt(sign@entry@past,bush) obj(sign@entry@past,document@indef) tim(sign@entry@past,2006) Arindam (IITB) Textual Entailment November 9, / 59

66 Illustration [CS : Lecture 29, Prasad Pradip Joshi] Text: Manmohan Singh along with president George Bush signed a letter in Hypothesis: Bush signed a document. Text representation agt(sign@entry@past,manmohan Singh) cag(sign@entry@past,president) nam(president,george Bush) obj(sign@entry@past,letter@indef) tim(sign@entry@past,2006) aoj(president,george Bush) cag(sign@entry@past,george Bush) Hypothesis Representation agt(sign@entry@past,bush) obj(sign@entry@past,document@indef) tim(sign@entry@past,2006) Arindam (IITB) Textual Entailment November 9, / 59

67 Illustration [CS : Lecture 29, Prasad Pradip Joshi] Text: Manmohan Singh along with president George Bush signed a letter in Hypothesis: Bush signed a document. Text representation agt(sign@entry@past,manmohan Singh) cag(sign@entry@past,president) nam(president,george Bush) obj(sign@entry@past,letter@indef) tim(sign@entry@past,2006) aoj(president,george Bush) cag(sign@entry@past,george Bush) Hypothesis Representation agt(sign@entry@past,bush) obj(sign@entry@past,document@indef) tim(sign@entry@past,2006) Arindam (IITB) Textual Entailment November 9, / 59

68 Illustration [CS : Lecture 29, Prasad Pradip Joshi] Text: Manmohan Singh along with president George Bush signed a letter in Hypothesis: Bush signed a document. Text representation agt(sign@entry@past,manmohan Singh) cag(sign@entry@past,president) nam(president,george Bush) obj(sign@entry@past,letter@indef) tim(sign@entry@past,2006) aoj(president,george Bush) cag(sign@entry@past,george Bush) Hypothesis Representation agt(sign@entry@past,bush) obj(sign@entry@past,document@indef) tim(sign@entry@past,2006) Arindam (IITB) Textual Entailment November 9, / 59

69 Illustration [CS : Lecture 29, Prasad Pradip Joshi] Text: Manmohan Singh along with president George Bush signed a letter in Hypothesis: Bush signed a document. Text representation agt(sign@entry@past,manmohan Singh) cag(sign@entry@past,president) nam(president,george Bush) obj(sign@entry@past,letter@indef) tim(sign@entry@past,2006) aoj(president,george Bush) cag(sign@entry@past,george Bush) Hypothesis Representation agt(sign@entry@past,bush) obj(sign@entry@past,document@indef) tim(sign@entry@past,2006) Arindam (IITB) Textual Entailment November 9, / 59

70 Results On the training set, (200 pairs of gold standard UNL from RTE and FRACAS) the precision value stands at 96.55% and the recall stands at 95.72% Using UNL enconvertor (70.1% accurate), on phenomenon studied FRACAS (100 pairs), precision is 63.04% and recall is 60.1% On complete FRACAS dataset, precision 60.1% and recall 46% Arindam (IITB) Textual Entailment November 9, / 59

71 Growth Rule [CS : Lecture 29, Prasad Pradip Joshi] pos-mod rule: Presence of pos(a,b) add mod(a,b) Navy of India Indian Navy plc closure: Presence of plc(a,b) and plc(b,c) leads to the addition of plc(a,c) Paris is capital of France. France is in Europe. Paris is in Europe. Introduction of words based on UNL relations and attributes: finish or over Relations: plc located pos belongs to or owned by Arindam (IITB) Textual Entailment November 9, / 59

72 Growth Rules [Maheshwari, 2009] Figure: Growth Rules Arindam (IITB) Textual Entailment November 9, / 59

73 Matching Rules [CS : Lecture 29, Prasad Pradip Joshi] Two types Matching the UNL relations (predicate names) Look up whether a relation belongs to the same family as other e.g. agt(agent),cag(co-agent),aoj(attribute of object) Matching the argument part. A narrowing edit of thing pointed to by aoj A broadening edit of thing pointed to by obj Arindam (IITB) Textual Entailment November 9, / 59

74 Universal Words Representation UNL: <UW> := < integer > (<POS><WORDNETID>) Natural Language: <UW> := < root > [< suffix > ] Example Concept: a piece of furniture with tableware for a meal laid out on it UNL Representation: NL Representation : table(icl>furniture) Arindam (IITB) Textual Entailment November 9, / 59

75 Relations labeled arcs connecting a node to another node in a UNL graph correspond to two-place semantic predicates holding between two Universal Words used to describe semantic dependencies between syntactic constituents organized in a hierarchy where lower nodes subsume upper nodes Arindam (IITB) Textual Entailment November 9, / 59

76 Relations Example Bob slept = agt(slept,bob) Alice died = obj(died,alice) John believes in Mary = aoj(believes,john) John worked while Peter talked = coo(worked,talked) Arindam (IITB) Textual Entailment November 9, / 59

77 Attributes Syntax arcs linking a node to itself In opposition to relations, they correspond to one-place predicates used to represent information conveyed by natural language grammatical categories (such as tense, mood, aspect, number, etc) <attribute> <attribute-name> := <character>+ Arindam (IITB) Textual Entailment November 9, / 59

78 Pair Features Example T: At the end of the year, all solid companies pay dividends. H: At the end of the year, all solid insurance companies pay dividends. Possible feature pairs: Bag of words: Text Hypothesis end T year T end H year H solid T solid H We can learn: T implies H as when T contains end T does not imply H when H contains end Totally useless??? Arindam (IITB) Textual Entailment November 9, / 59

79 Pair Features Example T: At the end of the year, all solid companies pay dividends. H: At the end of the year, all solid insurance companies pay dividends. Possible feature pairs: Bag of words: Text Hypothesis end T year T end H year H solid T solid H We can learn: T implies H as when T contains end T does not imply H when H contains end Totally useless??? Arindam (IITB) Textual Entailment November 9, / 59

80 Pair Features Example T: At the end of the year, all solid companies pay dividends. H: At the end of the year, all solid insurance companies pay dividends. Possible feature pairs: Bag of words: Text Hypothesis end T year T end H year H solid T solid H We can learn: T implies H as when T contains end T does not imply H when H contains end Totally useless??? Arindam (IITB) Textual Entailment November 9, / 59

81 Effectively using Pair Feature Space [?] Example T : At the end of the year, all solid companies pay dividends. H 1 : At the end of the year, all solid insurance companies pay dividends. H 2 : At the end of the year, all solid companies pay cash dividends. Distance feature will plot < T, H 1 > and < T, H 2 > to be same points. We need a space that considers the content and the structure of textual entailment examples. Arindam (IITB) Textual Entailment November 9, / 59

82 Effectively using Pair Feature Space [?] Example T : At the end of the year, all solid companies pay dividends. H 1 : At the end of the year, all solid insurance companies pay dividends. H 2 : At the end of the year, all solid companies pay cash dividends. Distance feature will plot < T, H 1 > and < T, H 2 > to be same points. We need a space that considers the content and the structure of textual entailment examples. Arindam (IITB) Textual Entailment November 9, / 59

83 Effectively using Pair Feature Space [?] Example T : At the end of the year, all solid companies pay dividends. H 1 : At the end of the year, all solid insurance companies pay dividends. H 2 : At the end of the year, all solid companies pay cash dividends. Distance feature will plot < T, H 1 > and < T, H 2 > to be same points. We need a space that considers the content and the structure of textual entailment examples. Arindam (IITB) Textual Entailment November 9, / 59

84 Effectively using Pair Feature Space [?] Example T : At the end of the year, all solid companies pay dividends. H 1 : At the end of the year, all solid insurance companies pay dividends. H 2 : At the end of the year, all solid companies pay cash dividends. Distance feature will plot < T, H 1 > and < T, H 2 > to be same points. We need a space that considers the content and the structure of textual entailment examples. Arindam (IITB) Textual Entailment November 9, / 59

85 Effectively using Pair Feature Space [?] Example T : At the end of the year, all solid companies pay dividends. H 1 : At the end of the year, all solid insurance companies pay dividends. H 2 : At the end of the year, all solid companies pay cash dividends. Distance feature will plot < T, H 1 > and < T, H 2 > to be same points. We need a space that considers the content and the structure of textual entailment examples. Arindam (IITB) Textual Entailment November 9, / 59

86 Kernel Trick Syntactic pair feature space. Cross Pair Similarity K(< T, H >, < T, H >) = K(< T, T >) + K(< H, H >) defining the distance K(P 1, P 2 ) instead of features as max (K T (t(h, c), t(h, i)) + K T (t(t, c), t(t, i))) c C Makes Pair Feature look useful. Arindam (IITB) Textual Entailment November 9, / 59

Semantic Structure of the Indian Sign Language

Semantic Structure of the Indian Sign Language Semantic Structure of the Indian Sign Language Purushottam Kar and Achla M. Raina Indian Institute of Technology Kanpur 6 January 2008 Overview Indian Sign Language An Introduction Sociolinguistic and

More information

. Semi-automatic WordNet Linking using Word Embeddings. Kevin Patel, Diptesh Kanojia and Pushpak Bhattacharyya Presented by: Ritesh Panjwani

. Semi-automatic WordNet Linking using Word Embeddings. Kevin Patel, Diptesh Kanojia and Pushpak Bhattacharyya Presented by: Ritesh Panjwani Semi-automatic WordNet Linking using Word Embeddings Kevin Patel, Diptesh Kanojia and Pushpak Bhattacharyya Presented by: Ritesh Panjwani January 11, 2018 Kevin Patel WordNet Linking via Embeddings 1/22

More information

Referring Expressions & Alternate Views of Summarization. Ling 573 Systems and Applications May 24, 2016

Referring Expressions & Alternate Views of Summarization. Ling 573 Systems and Applications May 24, 2016 Referring Expressions & Alternate Views of Summarization Ling 573 Systems and Applications May 24, 2016 Content realization: Referring expressions Roadmap Alternate views of summarization: Dimensions of

More information

Using a grammar implementation to teach writing skills

Using a grammar implementation to teach writing skills Using a grammar implementation to teach writing skills Dan Flickinger CSLI, Stanford University Workshop on Technology Enhanced Learning GWC 2018, Singapore 12 January 2018 Goals Automated error detection

More information

Chapter 2. Knowledge Representation: Reasoning, Issues, and Acquisition. Teaching Notes

Chapter 2. Knowledge Representation: Reasoning, Issues, and Acquisition. Teaching Notes Chapter 2 Knowledge Representation: Reasoning, Issues, and Acquisition Teaching Notes This chapter explains how knowledge is represented in artificial intelligence. The topic may be launched by introducing

More information

Thursday, July 14, Monotonicity

Thursday, July 14, Monotonicity Monotonicity inference: conserve truth from premises to conclusion find patterns that do this Monotonicity Upward: weaker (less specific) predicates can be substituted for stronger ones Downward: stronger

More information

Natural Logic Inference for Emotion Detection

Natural Logic Inference for Emotion Detection Natural Logic Inference for Emotion Detection Han Ren 1, Yafeng Ren 2, Xia Li 1, Wenhe Feng 1 and Maofu Liu 3 1 Laboratory of Language Engineering and Computing, Guangdong University of Foreign Studies,

More information

FOURTH EDITION. NorthStar ALIGNMENT WITH THE GLOBAL SCALE OF ENGLISH AND THE COMMON EUROPEAN FRAMEWORK OF REFERENCE

FOURTH EDITION. NorthStar ALIGNMENT WITH THE GLOBAL SCALE OF ENGLISH AND THE COMMON EUROPEAN FRAMEWORK OF REFERENCE 4 FOURTH EDITION NorthStar ALIGNMENT WITH THE GLOBAL SCALE OF ENGLISH AND THE COMMON EUROPEAN FRAMEWORK OF REFERENCE 1 NorthStar Reading & Writing 3, 4th Edition NorthStar FOURTH EDITION NorthStar, Fourth

More information

Foundations of Natural Language Processing Lecture 13 Heads, Dependency parsing

Foundations of Natural Language Processing Lecture 13 Heads, Dependency parsing Foundations of Natural Language Processing Lecture 13 Heads, Dependency parsing Alex Lascarides (slides from Alex Lascarides, Henry Thompson, Nathan Schneider and Sharon Goldwater) 6 March 2018 Alex Lascarides

More information

Stepwise Knowledge Acquisition in a Fuzzy Knowledge Representation Framework

Stepwise Knowledge Acquisition in a Fuzzy Knowledge Representation Framework Stepwise Knowledge Acquisition in a Fuzzy Knowledge Representation Framework Thomas E. Rothenfluh 1, Karl Bögl 2, and Klaus-Peter Adlassnig 2 1 Department of Psychology University of Zurich, Zürichbergstraße

More information

Today we will... Foundations of Natural Language Processing Lecture 13 Heads, Dependency parsing. Evaluating parse accuracy. Evaluating parse accuracy

Today we will... Foundations of Natural Language Processing Lecture 13 Heads, Dependency parsing. Evaluating parse accuracy. Evaluating parse accuracy Today we will... Foundations of Natural Language Processing Lecture 13 Heads, Dependency parsing Alex Lascarides (slides from Alex Lascarides, Henry Thompson, Nathan chneider and haron Goldwater) 6 March

More information

Do Future Work sections have a purpose?

Do Future Work sections have a purpose? Do Future Work sections have a purpose? Citation links and entailment for global scientometric questions Simone Teufel University of Cambridge and Tokyo Institute of Technology August 11, 2017 1/30 Future

More information

Intelligent Machines That Act Rationally. Hang Li Bytedance AI Lab

Intelligent Machines That Act Rationally. Hang Li Bytedance AI Lab Intelligent Machines That Act Rationally Hang Li Bytedance AI Lab Four Definitions of Artificial Intelligence Building intelligent machines (i.e., intelligent computers) Thinking humanly Acting humanly

More information

Generalizing Dependency Features for Opinion Mining

Generalizing Dependency Features for Opinion Mining Generalizing Dependency Features for Mahesh Joshi 1 and Carolyn Rosé 1,2 1 Language Technologies Institute 2 Human-Computer Interaction Institute Carnegie Mellon University ACL-IJCNLP 2009 Short Papers

More information

How to Recognize and Reduce Challenges to Children s Comprehension of Books

How to Recognize and Reduce Challenges to Children s Comprehension of Books How to Recognize and Reduce Challenges to Children s Comprehension of Books Rebecca Landa, Ph.D., CCC-SLP Director, Center for Autism and Related Disorders April 3, 2013 Main points Have a look at children

More information

Intelligent Machines That Act Rationally. Hang Li Toutiao AI Lab

Intelligent Machines That Act Rationally. Hang Li Toutiao AI Lab Intelligent Machines That Act Rationally Hang Li Toutiao AI Lab Four Definitions of Artificial Intelligence Building intelligent machines (i.e., intelligent computers) Thinking humanly Acting humanly Thinking

More information

Building Evaluation Scales for NLP using Item Response Theory

Building Evaluation Scales for NLP using Item Response Theory Building Evaluation Scales for NLP using Item Response Theory John Lalor CICS, UMass Amherst Joint work with Hao Wu (BC) and Hong Yu (UMMS) Motivation Evaluation metrics for NLP have been mostly unchanged

More information

Chapter 3 Software Packages to Install How to Set Up Python Eclipse How to Set Up Eclipse... 42

Chapter 3 Software Packages to Install How to Set Up Python Eclipse How to Set Up Eclipse... 42 Table of Contents Preface..... 21 About the Authors... 23 Acknowledgments... 24 How This Book is Organized... 24 Who Should Buy This Book?... 24 Where to Find Answers to Review Questions and Exercises...

More information

Common Sense Assistant for Writing Stories that Teach Social Skills

Common Sense Assistant for Writing Stories that Teach Social Skills Common Sense Assistant for Writing Stories that Teach Social Skills Kyunghee Kim MIT Media Laboratory 20 Ames Street. E15-443D Cambridge, MA 02139 USA khkim@media.mit.edu Rosalind W. Picard MIT Media Laboratory

More information

Learning the Fine-Grained Information Status of Discourse Entities

Learning the Fine-Grained Information Status of Discourse Entities Learning the Fine-Grained Information Status of Discourse Entities Altaf Rahman and Vincent Ng Human Language Technology Research Institute The University of Texas at Dallas Plan for the talk What is Information

More information

English and Persian Apposition Markers in Written Discourse: A Case of Iranian EFL learners

English and Persian Apposition Markers in Written Discourse: A Case of Iranian EFL learners 7 English and Persian Apposition Markers in Written Discourse: A Case of Iranian EFL learners Samaneh Chamanaraeian M.A. Student in Islamic Azad University (Isfahan Branch) samanechaman@yahoo.com and (corresponding

More information

RUBRICS: CRITICAL ATTRIBUTES

RUBRICS: CRITICAL ATTRIBUTES 1 RUBRICS: CRITICAL ATTRIBUTES FOR DIFFERENT GENRE Anita L. Archer, Ph.D. archerteach@aol.com (503-295-7749) 2 SIX TRAIT RUBRIC IDEAS 1. Is the author s message clear? 2. Did the author have enough information?

More information

defying complexity (lessons learned)

defying complexity (lessons learned) defying complexity (lessons learned) Karën Fort & Bruno Guillaume karen.fort@paris-sorbonne.fr / bruno.guillaume@loria.fr October, 2016 1 / 31 1 Overview of the game 2 Motivating players 3 Behind the curtain

More information

SPICE: Semantic Propositional Image Caption Evaluation

SPICE: Semantic Propositional Image Caption Evaluation SPICE: Semantic Propositional Image Caption Evaluation Presented to the COCO Consortium, Sept 2016 Peter Anderson1, Basura Fernando1, Mark Johnson2 and Stephen Gould1 1 Australian National University 2

More information

Data Mining in Bioinformatics Day 4: Text Mining

Data Mining in Bioinformatics Day 4: Text Mining Data Mining in Bioinformatics Day 4: Text Mining Karsten Borgwardt February 25 to March 10 Bioinformatics Group MPIs Tübingen Karsten Borgwardt: Data Mining in Bioinformatics, Page 1 What is text mining?

More information

ADDITIONAL TOOLS: Pushing Pause From Heads Up. Suggested Answers for Lesson Discussion Questions headsup.scholastic.com/pushingpause/lesson

ADDITIONAL TOOLS: Pushing Pause From Heads Up. Suggested Answers for Lesson Discussion Questions headsup.scholastic.com/pushingpause/lesson ADDITIONAL TOOLS: Pushing Pause From Heads Up Dear Teacher, The following tools are designed to help support and enrich the Heads Up lesson plan and student article Pushing Pause : 1A) Suggested Answers

More information

Expert System Profile

Expert System Profile Expert System Profile GENERAL Domain: Medical Main General Function: Diagnosis System Name: INTERNIST-I/ CADUCEUS (or INTERNIST-II) Dates: 1970 s 1980 s Researchers: Ph.D. Harry Pople, M.D. Jack D. Myers

More information

MA 1 Notes. moving the hand may be needed.

MA 1 Notes. moving the hand may be needed. Name Period MA 1 Notes Fingerspelling Consider frngerspelling to be like your. Being clear is vital to being understood, be enough not to worry, whether each letter is exactly right, and be able to spell

More information

Where the Truth Lies: Explaining the Credibility of Emerging Claims on the Web and Social Media

Where the Truth Lies: Explaining the Credibility of Emerging Claims on the Web and Social Media Where the Truth Lies: Explaining the Credibility of Emerging Claims on the Web and Social Media Kashyap Popat, Subhabrata Mukherjee, Jannik Strötgen, Gerhard Weikum WWW 2017 MOTIVATION Rapid spread of

More information

Parsing Discourse Relations. Giuseppe Riccardi Signals and Interactive Systems Lab University of Trento, Italy

Parsing Discourse Relations. Giuseppe Riccardi Signals and Interactive Systems Lab University of Trento, Italy Parsing Discourse Relations Giuseppe Riccardi Signals and Interactive Systems Lab University of Trento, Italy Behavioral Analytics Parser Run on Genia Corpus Among 25 cases, 2 homozygous deletions and

More information

AS and A Level English Language and Literature EXEMPLAR RESPONSES

AS and A Level English Language and Literature EXEMPLAR RESPONSES AS and A Level English Language and Literature EXEMPLAR RESPONSES Contents About this exemplars pack... 2 Mark scheme... 3 EXEMPLAR RESPONSE A... 4 Society and the Individual... 4 Marker s comments...

More information

Question 1: The narrator compares herself to a wounded zebra in a National Geographic special when telling of how she felt on her first day of high sc

Question 1: The narrator compares herself to a wounded zebra in a National Geographic special when telling of how she felt on her first day of high sc Let s grade the together Question 1: The prompt asks students to choose one example of imagery from the box and explain its significance. Don t DEFINE imagery. Students must show how the effective use

More information

Phobia Factor STUDENT BOOK, Pages 61 64

Phobia Factor STUDENT BOOK, Pages 61 64 UNDERSTANDING READING STRATEGIES Summarizing Materials Student Book pages 61 64 BLMs 2, 3, 6 Audio CD ACCESSIBILITY EASY AVERAGE CHALLENGING Phobia names are challenging, but not crucial to comprehension

More information

MA 1 Notes. Deaf vs deaf p. 3 MA1 F 13

MA 1 Notes. Deaf vs deaf p. 3 MA1 F 13 Name Period MA 1 Notes Fingerspelling Consider frngerspelling to be like your handwriting. Being clear is vital to being understood, be confident enough not to worry, whether each letter is exactly right,

More information

Corpus Construction and Semantic Analysis of Indonesian Image Description

Corpus Construction and Semantic Analysis of Indonesian Image Description Corpus Construction and Semantic Analysis of Indonesian Image Description Khumaisa Nur aini 1,3, Johanes Effendi 1, Sakriani Sakti 1,2, Mirna Adriani 3, Sathosi Nakamura 1,2 1 Nara Institute of Science

More information

Tips for Writing a Research Paper in APA format:

Tips for Writing a Research Paper in APA format: Tips for Writing a Research Paper in APA format: Basics: A research paper (especially one that requires APA style) is different than a term paper, a creative writing paper, a composition-style paper, or

More information

City of Angels School Independent Study Los Angeles Unified School District Contemporary Composition Instructional Guide

City of Angels School Independent Study Los Angeles Unified School District Contemporary Composition Instructional Guide City of Angels School Independent Study Los Angeles Unified School District Contemporary Composition Instructional Guide This is the instructional guide for the course that covers one semester of the eleventh

More information

What Is A Knowledge Representation? Lecture 13

What Is A Knowledge Representation? Lecture 13 What Is A Knowledge Representation? 6.871 - Lecture 13 Outline What Is A Representation? Five Roles What Should A Representation Be? What Consequences Does This View Have For Research And Practice? One

More information

Meets Requirements Exemplars for English for Academic Purposes. Level 4

Meets Requirements Exemplars for English for Academic Purposes. Level 4 Exemplar for English for Academic Purposes for Unit Standard US22750 v2 Meets Requirements Exemplars for English for Academic Purposes Level 4 These exemplars support assessment against: Unit Standard

More information

Writing Reaction Papers Using the QuALMRI Framework

Writing Reaction Papers Using the QuALMRI Framework Writing Reaction Papers Using the QuALMRI Framework Modified from Organizing Scientific Thinking Using the QuALMRI Framework Written by Kevin Ochsner and modified by others. Based on a scheme devised by

More information

Perceptual Organization (II)

Perceptual Organization (II) (II) Introduction to Computational and Biological Vision CS 202-1-5261 Computer Science Department, BGU Ohad Ben-Shahar Why do things look they way they do? [Koffka 1935] External (Environment) vs. Internal

More information

Vagueness, Context Dependence and Interest Relativity

Vagueness, Context Dependence and Interest Relativity Chris Kennedy Seminar on Vagueness University of Chicago 2 May, 2006 Vagueness, Context Dependence and Interest Relativity 1 Questions about vagueness Graff (2000) summarizes the challenge for a theory

More information

rskills Progress Monitoring Test 2a

rskills Progress Monitoring Test 2a rskills Test 2a, page 1 NAME: DATE: rskills Progress Monitoring Test 2a DIRECTIONS: This is a reading test. Follow the directions for each part of the test, and choose the best answer to each question.

More information

VINAYAKA MISSIONS SIKKIM UNIVERSITY

VINAYAKA MISSIONS SIKKIM UNIVERSITY Programme: BA(English) Session: 2015-16 Full Marks: 10 Assignment No. 1 Last Date of Submission: 31 st March 2016 NOTE : All Sections in the Assignments are compulsory to be attempted as per Instructions.

More information

NEW YORK STATE TEACHER CERTIFICATION EXAMINATIONS

NEW YORK STATE TEACHER CERTIFICATION EXAMINATIONS NEW YORK STATE TEACHER CERTIFICATION EXAMINATIONS TEST DESIGN AND FRAMEWORK April 2016 Authorized for Distribution by the New York State Education Department This test design and framework document is

More information

Chapter 7: Descriptive Statistics

Chapter 7: Descriptive Statistics Chapter Overview Chapter 7 provides an introduction to basic strategies for describing groups statistically. Statistical concepts around normal distributions are discussed. The statistical procedures of

More information

Jazyková kompetence I Session II

Jazyková kompetence I Session II Jazyková kompetence I Session II Essay Writing: The Basics What does a good essay need? An academic essay aims to persuade readers of an idea based on evidence. An academic essay should answer a question

More information

9-10 Issue 181 VIBE ACTIVITIES. Healthy Vibe - I Quit Because... page 22. Issue 181 Page 1 Y E A R. Name:

9-10 Issue 181 VIBE ACTIVITIES. Healthy Vibe - I Quit Because... page 22. Issue 181 Page 1 Y E A R. Name: Name: VIBE ACTIVITIES Healthy Vibe - I Quit Because... page 22 Page 1 Did you know that over 4000 chemical compounds are created by burning just one cigarette? None of these are good for your body. The

More information

Memory-Augmented Active Deep Learning for Identifying Relations Between Distant Medical Concepts in Electroencephalography Reports

Memory-Augmented Active Deep Learning for Identifying Relations Between Distant Medical Concepts in Electroencephalography Reports Memory-Augmented Active Deep Learning for Identifying Relations Between Distant Medical Concepts in Electroencephalography Reports Ramon Maldonado, BS, Travis Goodwin, PhD Sanda M. Harabagiu, PhD The University

More information

Implicit Information in Directionality of Verbal Probability Expressions

Implicit Information in Directionality of Verbal Probability Expressions Implicit Information in Directionality of Verbal Probability Expressions Hidehito Honda (hito@ky.hum.titech.ac.jp) Kimihiko Yamagishi (kimihiko@ky.hum.titech.ac.jp) Graduate School of Decision Science

More information

M California Copyright by CTB/McGraw-Hill

M California Copyright by CTB/McGraw-Hill O R T#-# r e m R L W BM AM Sp Sc IRIS Track # and Lesson # Reading At Your Own Pace English At Your Own Pace Math At Your Own Pace Reading Accelerated Learning Lab Language Accelerated Learning Lab Writing

More information

Using Scripts to help in Biomedical Text Interpretation

Using Scripts to help in Biomedical Text Interpretation Using Scripts to help in Biomedical Text Interpretation Working Note 30 Peter Clark (peter.e.clark@boeing.com), Boeing Research and Technology, 2009 Introduction This short note speculates on the use of

More information

Requirements for Maintaining Web Access for Hearing-Impaired Individuals

Requirements for Maintaining Web Access for Hearing-Impaired Individuals Requirements for Maintaining Web Access for Hearing-Impaired Individuals Daniel M. Berry 2003 Daniel M. Berry WSE 2001 Access for HI Requirements for Maintaining Web Access for Hearing-Impaired Individuals

More information

Review Questions in Introductory Knowledge... 37

Review Questions in Introductory Knowledge... 37 Table of Contents Preface..... 17 About the Authors... 19 How This Book is Organized... 20 Who Should Buy This Book?... 20 Where to Find Answers to Review Questions and Exercises... 20 How to Report Errata...

More information

May All Your Wishes Come True: A Study of Wishes and How to Recognize Them

May All Your Wishes Come True: A Study of Wishes and How to Recognize Them May All Your Wishes Come True: A Study of Wishes and How to Recognize Them Andrew B. Goldberg, Nathanael Fillmore, David Andrzejewski, Zhiting Xu, Bryan Gibson & Xiaojin Zhu Computer Sciences Department

More information

APPROVAL SHEET. Uncertainty in Semantic Web. Doctor of Philosophy, 2005

APPROVAL SHEET. Uncertainty in Semantic Web. Doctor of Philosophy, 2005 APPROVAL SHEET Title of Dissertation: BayesOWL: A Probabilistic Framework for Uncertainty in Semantic Web Name of Candidate: Zhongli Ding Doctor of Philosophy, 2005 Dissertation and Abstract Approved:

More information

FSA Training Papers Grade 7 Exemplars. Rationales

FSA Training Papers Grade 7 Exemplars. Rationales FSA Training Papers Grade 7 Exemplars Rationales Rationales for Grade 7 Exemplars Reading Grade 7 Reading Exemplar #1: Score 3 Comprehension of the passages and task clearly evident Generally purposeful

More information

Mechanical rationality, decision making and emotions

Mechanical rationality, decision making and emotions Mechanical rationality, decision making and emotions Olga Markič University of Ljubljana olga.markic@guest.arnes.si 1. Rationality - Instrumental rationality and negative view of emotion 2. The possibility

More information

An Interval-Based Representation of Temporal Knowledge

An Interval-Based Representation of Temporal Knowledge An Interval-Based Representation of Temporal Knowledge James F. Allen Department of Computer Science The University of Rochester Rochester, NY 14627 Abstract This paper describes a method for maintaining

More information

Information Extraction

Information Extraction Information Extraction Claire Cardie Cornell University Information Extraction Introduction Task definition Evaluation IE system architecture Acquiring extraction patterns Manually defined patterns Learning

More information

CHAPTER 3 DATA ANALYSIS: DESCRIBING DATA

CHAPTER 3 DATA ANALYSIS: DESCRIBING DATA Data Analysis: Describing Data CHAPTER 3 DATA ANALYSIS: DESCRIBING DATA In the analysis process, the researcher tries to evaluate the data collected both from written documents and from other sources such

More information

Inferencing in Artificial Intelligence and Computational Linguistics

Inferencing in Artificial Intelligence and Computational Linguistics Inferencing in Artificial Intelligence and Computational Linguistics (http://www.dfki.de/~horacek/infer-ai-cl.html) no classes on 28.5., 18.6., 25.6. 2-3 extra lectures will be scheduled Helmut Horacek

More information

NYU Refining Event Extraction (Old-Fashion Traditional IE) Through Cross-document Inference

NYU Refining Event Extraction (Old-Fashion Traditional IE) Through Cross-document Inference NYU Refining Event Extraction (Old-Fashion Traditional IE) Through Cross-document Inference Heng Ji and Ralph Grishman (hengji,grishman)@cs.nyu.edu Computer Science Department New York University June,

More information

Lecture 10: POS Tagging Review. LING 1330/2330: Introduction to Computational Linguistics Na-Rae Han

Lecture 10: POS Tagging Review. LING 1330/2330: Introduction to Computational Linguistics Na-Rae Han Lecture 10: POS Tagging Review LING 1330/2330: Introduction to Computational Linguistics Na-Rae Han Overview Part-of-speech tagging Language and Computers, Ch. 3.4 Tokenization, POS tagging NLTK Book Ch.5

More information

China Summer Institute 2015 Connie Steinman Connecting Chinese & American Cultures Through Sign Language & Religious Gestures

China Summer Institute 2015 Connie Steinman Connecting Chinese & American Cultures Through Sign Language & Religious Gestures 1 China Summer Institute 2015 Connie Steinman Connecting Chinese & American Cultures Through Sign Language & Religious Gestures Organizing Questions: How is sign language used to communicate in an area

More information

TRANSLITERATING: THE INTERPRETING NO ONE WANTS TO TALK ABOUT by Karen Malcolm

TRANSLITERATING: THE INTERPRETING NO ONE WANTS TO TALK ABOUT by Karen Malcolm 59 TRANSLITERATING: THE INTERPRETING NO ONE WANTS TO TALK ABOUT by Karen Malcolm Since the 1960s, as sign language interpreting has developed as a profession, distinctions have been made between interpreting

More information

The Scarlet Letter. Character Analysis Essay

The Scarlet Letter. Character Analysis Essay The Scarlet Letter Character Analysis Essay Due Dates: Thesis due 10/4 Body Paragraph Outline due 10/9 Rough Draft due 10/16 Final Draft due 10/23 Prompt: Choose a character from The Scarlet Letter and

More information

Model answers. Childhood memories (signed by Avril Langard-Tang) Introduction:

Model answers. Childhood memories (signed by Avril Langard-Tang) Introduction: Childhood memories (signed by Avril Langard-Tang) Model answers Introduction: Many people love to watch British Sign Language because they see it as expressive and engaging. What they don t always understand

More information

Using Your Brain -- for a CHANGE Summary. NLPcourses.com

Using Your Brain -- for a CHANGE Summary. NLPcourses.com Using Your Brain -- for a CHANGE Summary NLPcourses.com Table of Contents Using Your Brain -- for a CHANGE by Richard Bandler Summary... 6 Chapter 1 Who s Driving the Bus?... 6 Chapter 2 Running Your Own

More information

Director of Testing and Disability Services Phone: (706) Fax: (706) E Mail:

Director of Testing and Disability Services Phone: (706) Fax: (706) E Mail: Angie S. Baker Testing and Disability Services Director of Testing and Disability Services Phone: (706)737 1469 Fax: (706)729 2298 E Mail: tds@gru.edu Deafness is an invisible disability. It is easy for

More information

Committee-based Decision Making in Probabilistic Partial Parsing

Committee-based Decision Making in Probabilistic Partial Parsing Committee-based Decision Making in Probabilistic Partial Parsing * * INUI Takashi and INUI Kentaro * Kyushu Institute of Technology PRESTO,Japan Science and Technology Corporation Background Tree banks

More information

Predicting Breast Cancer Survivability Rates

Predicting Breast Cancer Survivability Rates Predicting Breast Cancer Survivability Rates For data collected from Saudi Arabia Registries Ghofran Othoum 1 and Wadee Al-Halabi 2 1 Computer Science, Effat University, Jeddah, Saudi Arabia 2 Computer

More information

Grounding Ontologies in the External World

Grounding Ontologies in the External World Grounding Ontologies in the External World Antonio CHELLA University of Palermo and ICAR-CNR, Palermo antonio.chella@unipa.it Abstract. The paper discusses a case study of grounding an ontology in the

More information

READTHEORY Passage and Questions

READTHEORY Passage and Questions READTHEORY Passage and Questions Reading Comprehension Assessment Directions: Read the passage. Then answer the questions below. Name Date Sports Drinks: Better Than Water? You are playing basketball with

More information

Age Mean Number of Questions Correct

Age Mean Number of Questions Correct 1 Practice test 1 and Ch 7 IQ & Ch 3.1 genes For eye color: Brown allele is autosomal dominant and blue is autosomal recessive. (an autosome is any chromosome other than a sex chromosome). See the short

More information

S URVEY ON EMOTION CLASSIFICATION USING SEMANTIC WEB

S URVEY ON EMOTION CLASSIFICATION USING SEMANTIC WEB This work by IJARBEST is licensed under Creative Commons Attribution 4.0 International License. Available at https://www.ijarbest.com S URVEY ON EMOTION CLASSIFICATION USING SEMANTIC WEB R.S.Reminaa, Department

More information

Revised 2016 GED Test Performance Level Descriptors: Level 1 (Below Passing: )

Revised 2016 GED Test Performance Level Descriptors: Level 1 (Below Passing: ) Revised 2016 GED Test Performance Level Descriptors: Level 1 (Below Passing: 100-144) Test-takers who score at the Below Passing level are typically able to comprehend and analyze simple passages similar

More information

Measuring Focused Attention Using Fixation Inner-Density

Measuring Focused Attention Using Fixation Inner-Density Measuring Focused Attention Using Fixation Inner-Density Wen Liu, Mina Shojaeizadeh, Soussan Djamasbi, Andrew C. Trapp User Experience & Decision Making Research Laboratory, Worcester Polytechnic Institute

More information

AIR QUESTION STEMS RL.6.1 RL.6.2 RL.6.3 RL.6.4 RL.6.5 6TH. COLUMBUS CITY SCHOOLS-SECONDARY ENGLISH

AIR QUESTION STEMS RL.6.1 RL.6.2 RL.6.3 RL.6.4 RL.6.5 6TH. COLUMBUS CITY SCHOOLS-SECONDARY ENGLISH 6TH RL.6.1 RL.6.2 RL.6.3 RL.6.4 RL.6.5 2 Select the sentence that shows [specific character action]. Select two sentences that support the idea that [an idea about a character]. 3 Part A: What inference

More information

Application of Tree Structures of Fuzzy Classifier to Diabetes Disease Diagnosis

Application of Tree Structures of Fuzzy Classifier to Diabetes Disease Diagnosis , pp.143-147 http://dx.doi.org/10.14257/astl.2017.143.30 Application of Tree Structures of Fuzzy Classifier to Diabetes Disease Diagnosis Chang-Wook Han Department of Electrical Engineering, Dong-Eui University,

More information

Evaluation & Systems. Ling573 Systems & Applications April 9, 2015

Evaluation & Systems. Ling573 Systems & Applications April 9, 2015 Evaluation & Systems Ling573 Systems & Applications April 9, 2015 Evaluation: Pyramid scoring Scoring without models Roadmap Systems: MEAD CLASSY Deliverable #2 Ideally informative summary Does not include

More information

Hoare Logic and Model Checking. LTL and CTL: a perspective. Learning outcomes. Model Checking Lecture 12: Loose ends

Hoare Logic and Model Checking. LTL and CTL: a perspective. Learning outcomes. Model Checking Lecture 12: Loose ends Learning outcomes Hoare Logic and Model Checking Model Checking Lecture 12: Loose ends Dominic Mulligan Based on previous slides by Alan Mycroft and Mike Gordon Programming, Logic, and Semantics Group

More information

A frame-based analysis of synaesthetic metaphors

A frame-based analysis of synaesthetic metaphors A frame-based analysis of synaesthetic metaphors Wiebke Petersen, Jens Fleischhauer, Hakan Beseoglu, Peter Bücker Institute of Language and Information Heinrich-Heine-Universität Düsseldorf Abstract The

More information

Presupposition. forweb. Existence Presuppositions. Factive Presuppositions. Connotative Presuppositions. Blame vs. Criticize

Presupposition. forweb. Existence Presuppositions. Factive Presuppositions. Connotative Presuppositions. Blame vs. Criticize Presupposition forweb Propositions whose truth is taken for granted in the utterance of a linguistic expression It s too bad Nader lost the election. Existence Presuppositions The movie on Cinemax is rated

More information

Brooke DePoorter M.Cl.Sc. (SLP) Candidate University of Western Ontario: School of Communication Sciences and Disorders

Brooke DePoorter M.Cl.Sc. (SLP) Candidate University of Western Ontario: School of Communication Sciences and Disorders Critical Review: In school-aged children with Autism Spectrum Disorder (ASD), what oral narrative elements differ from their typically developing peers? Brooke DePoorter M.Cl.Sc. (SLP) Candidate University

More information

Clinical Coreference Annotation Guidelines (with excerpts from ODIE guidelines and modified for SHARP) Arrick Lanfranchi and Kevin Crooks

Clinical Coreference Annotation Guidelines (with excerpts from ODIE guidelines and modified for SHARP) Arrick Lanfranchi and Kevin Crooks Clinical Coreference Annotation Guidelines (with excerpts from ODIE guidelines and modified for SHARP) Arrick Lanfranchi and Kevin Crooks The following is a proposal/summary of the ODIE guidelines with

More information

Sentiment Analysis of Reviews: Should we analyze writer intentions or reader perceptions?

Sentiment Analysis of Reviews: Should we analyze writer intentions or reader perceptions? Sentiment Analysis of Reviews: Should we analyze writer intentions or reader perceptions? Isa Maks and Piek Vossen Vu University, Faculty of Arts De Boelelaan 1105, 1081 HV Amsterdam e.maks@vu.nl, p.vossen@vu.nl

More information

two of a kind Two of a kind page 19 Issue 175 Yung Warriors go back to their roots for national tour Issue Years 9-10

two of a kind Two of a kind page 19 Issue 175 Yung Warriors go back to their roots for national tour Issue Years 9-10 Two of a kind page 19 THE YUNG WARRIORS ARE ABOUT TO DROP THEIR NEW ALBUM, AND THEY LL ALSO BE HITTING THE ROAD THIS MONTH WITH COLOURED STONE FOR A SPECIAL KIND OF HOMECOMING. T he Yung Warriors new album

More information

Embedded Implicatures

Embedded Implicatures 1 1. The Symmetry problem - Summary The Gricean system (simple version): (1) s(peaker) John has 3 children (=: ϕ) H(earer) reasons Embedded Implicatures Basic Inf: There is something else that s could

More information

Chapter 12 Conclusions and Outlook

Chapter 12 Conclusions and Outlook Chapter 12 Conclusions and Outlook In this book research in clinical text mining from the early days in 1970 up to now (2017) has been compiled. This book provided information on paper based patient record

More information

Word2Vec and Doc2Vec in Unsupervised Sentiment Analysis of Clinical Discharge Summaries

Word2Vec and Doc2Vec in Unsupervised Sentiment Analysis of Clinical Discharge Summaries Word2Vec and Doc2Vec in Unsupervised Sentiment Analysis of Clinical Discharge Summaries Qufei Chen University of Ottawa qchen037@uottawa.ca Marina Sokolova IBDA@Dalhousie University and University of Ottawa

More information

Houghton Mifflin Harcourt Avancemos!, Level correlated to the

Houghton Mifflin Harcourt Avancemos!, Level correlated to the Houghton Mifflin Harcourt Avancemos!, Level 4 2018 correlated to the READING 1. Read closely to determine what the text says explicitly and to make logical inferences from it; cite specific textual evidence

More information

FOURTH EDITION. NorthStar ALIGNMENT WITH THE GLOBAL SCALE OF ENGLISH AND THE COMMON EUROPEAN FRAMEWORK OF REFERENCE

FOURTH EDITION. NorthStar ALIGNMENT WITH THE GLOBAL SCALE OF ENGLISH AND THE COMMON EUROPEAN FRAMEWORK OF REFERENCE 4 FOURTH EDITION NorthStar ALIGNMENT WITH THE GLOBAL SCALE OF ENGLISH AND THE COMMON EUROPEAN FRAMEWORK OF REFERENCE 1 NorthStar Listening & Speaking 4, 4th Edition NorthStar FOURTH EDITION NorthStar,

More information

Perspective of Deafness-Exam 1

Perspective of Deafness-Exam 1 Perspective of Deafness-Exam 1 20/04/2015 3:46 PM Deaf People and Society Single Most striking feature/ Verbal communication barriors See better because you get better at eye sight because you can t rely

More information

Relationships Between the High Impact Indicators and Other Indicators

Relationships Between the High Impact Indicators and Other Indicators Relationships Between the High Impact Indicators and Other Indicators The High Impact Indicators are a list of key skills assessed on the GED test that, if emphasized in instruction, can help instructors

More information

Cognitive Processes PSY 334. Chapter 5 Abstraction of Information into Memory

Cognitive Processes PSY 334. Chapter 5 Abstraction of Information into Memory Cognitive Processes PSY 334 Chapter 5 Abstraction of Information into Memory Features of a Penny 1. Does the Lincoln on the penny face right or left? 2. Is anything above his head? What? 3. Is anything

More information

Item Writing Guide for the National Board for Certification of Hospice and Palliative Nurses

Item Writing Guide for the National Board for Certification of Hospice and Palliative Nurses Item Writing Guide for the National Board for Certification of Hospice and Palliative Nurses Presented by Applied Measurement Professionals, Inc. Copyright 2011 by Applied Measurement Professionals, Inc.

More information

Automatic Context-Aware Image Captioning

Automatic Context-Aware Image Captioning Technical Disclosure Commons Defensive Publications Series May 23, 2017 Automatic Context-Aware Image Captioning Sandro Feuz Sebastian Millius Follow this and additional works at: http://www.tdcommons.org/dpubs_series

More information

1. 1. When you exercise, your body uses the fuel to keep you going strong. a) deep b) general c) extra d) hard

1. 1. When you exercise, your body uses the fuel to keep you going strong. a) deep b) general c) extra d) hard Pre- university book Lesson one --------------------------------------------------------------------------------------------------------------- I. choose the correct answer : 1. 1. When you exercise, your

More information

Lesson 11 Correlations

Lesson 11 Correlations Lesson 11 Correlations Lesson Objectives All students will define key terms and explain the difference between correlations and experiments. All students should be able to analyse scattergrams using knowledge

More information