-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathdissertation_withAck.out
More file actions
84 lines (84 loc) · 5.9 KB
/
dissertation_withAck.out
File metadata and controls
84 lines (84 loc) · 5.9 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
\BOOKMARK [0][-]{chapter*.6}{LIST OF FIGURES}{}% 1
\BOOKMARK [0][-]{chapter*.7}{LIST OF TABLES}{}% 2
\BOOKMARK [0][-]{chapter*.8}{ABSTRACT}{}% 3
\BOOKMARK [0][-]{chapter.1}{CHAPTER INTRODUCTION}{}% 4
\BOOKMARK [1][-]{section.1.1}{Robustness}{chapter.1}% 5
\BOOKMARK [2][-]{subsection.1.1.1}{Aligning questions and answers}{section.1.1}% 6
\BOOKMARK [2][-]{subsection.1.1.2}{Customizing approaches to a specific question type}{section.1.1}% 7
\BOOKMARK [1][-]{section.1.2}{Interpretability}{chapter.1}% 8
\BOOKMARK [2][-]{subsection.1.2.1}{Multiple choice science questions as a proving ground}{section.1.2}% 9
\BOOKMARK [2][-]{subsection.1.2.2}{Reranking justifications with weak supervision}{section.1.2}% 10
\BOOKMARK [1][-]{section.1.3}{Contributions}{chapter.1}% 11
\BOOKMARK [2][-]{subsection.1.3.1}{Contribution 1: Using discourse structures to generate artificially aligned pairs for training question answering models}{section.1.3}% 12
\BOOKMARK [2][-]{subsection.1.3.2}{Contribution 2: Generating customized, task-specific word embeddings based on the question type}{section.1.3}% 13
\BOOKMARK [2][-]{subsection.1.3.3}{Contribution 3: Creating and ranking justifications for interpretable question answering}{section.1.3}% 14
\BOOKMARK [2][-]{subsection.1.3.4}{Contribution 4: Using neural networks to rank justifications for interpretable question answering}{section.1.3}% 15
\BOOKMARK [1][-]{section.1.4}{Overview}{chapter.1}% 16
\BOOKMARK [0][-]{chapter.2}{CHAPTER RELATED WORK}{}% 17
\BOOKMARK [0][-]{chapter.3}{CHAPTER Using Free Text to Train Monolingual Alignment Models for Question Answering }{}% 18
\BOOKMARK [1][-]{section.3.1}{Chapter overview}{chapter.3}% 19
\BOOKMARK [1][-]{section.3.2}{Related Work}{chapter.3}% 20
\BOOKMARK [1][-]{section.3.3}{Approach}{chapter.3}% 21
\BOOKMARK [1][-]{section.3.4}{Models and Features}{chapter.3}% 22
\BOOKMARK [1][-]{section.3.5}{Experiments}{chapter.3}% 23
\BOOKMARK [1][-]{section.3.6}{Results and Discussion}{chapter.3}% 24
\BOOKMARK [1][-]{section.3.7}{Conclusion}{chapter.3}% 25
\BOOKMARK [0][-]{chapter.4}{CHAPTER Creating Causal Embeddings for Question Answering with Minimal Supervision }{}% 26
\BOOKMARK [1][-]{section.4.1}{Chapter overview}{chapter.4}% 27
\BOOKMARK [1][-]{section.4.2}{Related Work}{chapter.4}% 28
\BOOKMARK [1][-]{section.4.3}{Chapter overview}{chapter.4}% 29
\BOOKMARK [1][-]{section.4.4}{Extracting Cause-Effect Tuples}{chapter.4}% 30
\BOOKMARK [1][-]{section.4.5}{Models}{chapter.4}% 31
\BOOKMARK [1][-]{section.4.6}{Direct Evaluation: Ranking Word Pairs}{chapter.4}% 32
\BOOKMARK [2][-]{subsection.4.6.1}{Data}{section.4.6}% 33
\BOOKMARK [2][-]{subsection.4.6.2}{Baselines}{section.4.6}% 34
\BOOKMARK [2][-]{subsection.4.6.3}{Results}{section.4.6}% 35
\BOOKMARK [1][-]{section.4.7}{Indirect Evaluation: QA Task}{chapter.4}% 36
\BOOKMARK [2][-]{subsection.4.7.1}{Data}{section.4.7}% 37
\BOOKMARK [2][-]{subsection.4.7.2}{Models and Features}{section.4.7}% 38
\BOOKMARK [2][-]{subsection.4.7.3}{Results}{section.4.7}% 39
\BOOKMARK [2][-]{subsection.4.7.4}{Error Analysis}{section.4.7}% 40
\BOOKMARK [1][-]{section.4.8}{Conclusion}{chapter.4}% 41
\BOOKMARK [0][-]{chapter.5}{CHAPTER Interpretable Question Answering: Building and Ranking Intersentence Answer Justifications }{}% 42
\BOOKMARK [1][-]{section.5.1}{Chapter Outline}{chapter.5}% 43
\BOOKMARK [1][-]{section.5.2}{Related Work}{chapter.5}% 44
\BOOKMARK [1][-]{section.5.3}{Approach}{chapter.5}% 45
\BOOKMARK [1][-]{section.5.4}{Focus Word Extraction}{chapter.5}% 46
\BOOKMARK [2][-]{subsection.5.4.1}{Scores and weights}{section.5.4}% 47
\BOOKMARK [1][-]{section.5.5}{Text Aggregation Graphs}{chapter.5}% 48
\BOOKMARK [1][-]{section.5.6}{Text Aggregation Graph Features}{chapter.5}% 49
\BOOKMARK [2][-]{subsection.5.6.1}{Features}{section.5.6}% 50
\BOOKMARK [2][-]{subsection.5.6.2}{Modeling Different TAG Types Using Domain Adaptation}{section.5.6}% 51
\BOOKMARK [1][-]{section.5.7}{Learning Model}{chapter.5}% 52
\BOOKMARK [2][-]{subsection.5.7.1}{Learning Algorithm}{section.5.7}% 53
\BOOKMARK [1][-]{section.5.8}{Experiments}{chapter.5}% 54
\BOOKMARK [2][-]{subsection.5.8.1}{Data}{section.5.8}% 55
\BOOKMARK [2][-]{subsection.5.8.2}{Tuning}{section.5.8}% 56
\BOOKMARK [2][-]{subsection.5.8.3}{Baselines}{section.5.8}% 57
\BOOKMARK [2][-]{subsection.5.8.4}{Results}{section.5.8}% 58
\BOOKMARK [1][-]{section.5.9}{Discussion}{chapter.5}% 59
\BOOKMARK [1][-]{section.5.10}{Error Analysis}{chapter.5}% 60
\BOOKMARK [2][-]{subsection.5.10.1}{Broader Error Classes}{section.5.10}% 61
\BOOKMARK [2][-]{subsection.5.10.2}{Summary of Errors}{section.5.10}% 62
\BOOKMARK [1][-]{section.5.11}{Conclusion}{chapter.5}% 63
\BOOKMARK [0][-]{chapter.6}{CHAPTER Robust and interpretable: A neural approach }{}% 64
\BOOKMARK [1][-]{section.6.1}{Chapter Outline}{chapter.6}% 65
\BOOKMARK [1][-]{section.6.2}{Related Work}{chapter.6}% 66
\BOOKMARK [1][-]{section.6.3}{Approach}{chapter.6}% 67
\BOOKMARK [1][-]{section.6.4}{Model and Features}{chapter.6}% 68
\BOOKMARK [2][-]{subsection.6.4.1}{Candidate Justification Retrieval}{section.6.4}% 69
\BOOKMARK [2][-]{subsection.6.4.2}{Feature Extraction}{section.6.4}% 70
\BOOKMARK [2][-]{subsection.6.4.3}{Neural Network}{section.6.4}% 71
\BOOKMARK [1][-]{section.6.5}{Experiments}{chapter.6}% 72
\BOOKMARK [2][-]{subsection.6.5.1}{Data and Setup}{section.6.5}% 73
\BOOKMARK [2][-]{subsection.6.5.2}{Baselines}{section.6.5}% 74
\BOOKMARK [2][-]{subsection.6.5.3}{Corpora}{section.6.5}% 75
\BOOKMARK [2][-]{subsection.6.5.4}{Model Tuning}{section.6.5}% 76
\BOOKMARK [1][-]{section.6.6}{Results}{chapter.6}% 77
\BOOKMARK [2][-]{subsection.6.6.1}{QA Performance}{section.6.6}% 78
\BOOKMARK [2][-]{subsection.6.6.2}{Justification Performance}{section.6.6}% 79
\BOOKMARK [2][-]{subsection.6.6.3}{Error Analysis}{section.6.6}% 80
\BOOKMARK [1][-]{section.6.7}{Conclusion}{chapter.6}% 81
\BOOKMARK [0][-]{chapter.7}{CHAPTER CONCLUSION}{}% 82
\BOOKMARK [0][-]{chapter.8}{CHAPTER Causal Extraction Rules Appendix }{}% 83
\BOOKMARK [0][-]{chapter*.23}{REFERENCES}{}% 84