Diverse Paraphrasing and its Effectiveness in Data Augmentation - - PowerPoint PPT Presentation

diverse paraphrasing and its effectiveness in data
SMART_READER_LITE
LIVE PREVIEW

Diverse Paraphrasing and its Effectiveness in Data Augmentation - - PowerPoint PPT Presentation

Diverse Paraphrasing and its Effectiveness in Data Augmentation Ashutosh Kumar*, Satwik Bhattamishra*, Manik Bhandari, Partha Talukdar Machine and Language Learning Lab (MALL) Indian Institute of Science, Bangalore *Equal Contributions


slide-1
SLIDE 1

Diverse Paraphrasing and its Effectiveness in Data Augmentation

Ashutosh Kumar*, Satwik Bhattamishra*, Manik Bhandari, Partha Talukdar Machine and Language Learning Lab (MALL) Indian Institute of Science, Bangalore

*Equal Contributions

slide-2
SLIDE 2

Conversational Agents

slide-3
SLIDE 3

Conversational Agents

slide-4
SLIDE 4

Conversational Agents

User: I want to book a flight from Minneapolis to New York

slide-5
SLIDE 5

Conversational Agents

User: I want to book a flight from Minneapolis to New York Bot: Sure. When are you planning to travel ?

slide-6
SLIDE 6

Conversational Agents

User: I want to book a flight from Minneapolis to New York Bot: Sure. When are you planning to travel ? User: Can you book plane to New York from Minneapolis

slide-7
SLIDE 7

Conversational Agents

User: I want to book a flight from Minneapolis to New York Bot: Sure. When are you planning to travel ? User: Can you book plane to New York from Minneapolis

Sorry, I don’t understand what you’re saying

Bot: Sorry, I don’t understand what you’re saying

slide-8
SLIDE 8

Conversational Agents

Data augmentation might help

User: I want to book a flight from Minneapolis to New York Bot: Sure. When are you planning to travel ? User: Can you book plane to New York from Minneapolis

Sorry, I don’t understand what you’re saying

Bot: Sorry, I don’t understand what you’re saying

slide-9
SLIDE 9

Paraphrase Generation

Rephrasing a given text in multiple ways

slide-10
SLIDE 10

Paraphrase Generation

Source how do i increase body height ?

Rephrasing a given text in multiple ways

slide-11
SLIDE 11

Paraphrase Generation

Source how do i increase body height ? Paraphrases

  • how could i increase my height ?
  • what should i do to increase body height ?
  • what are the ways to increase height ?
  • are there some ways to increase body height ?

Rephrasing a given text in multiple ways

slide-12
SLIDE 12

Paraphrase Generation

Source how do i increase body height ? Paraphrases

  • how could i increase my height ?
  • what should i do to increase body height ?
  • what are the ways to increase height ?
  • are there some ways to increase body height ?

Rephrasing a given text in multiple ways

Fidelity

(Meaning preserving)

slide-13
SLIDE 13

Paraphrase Generation

Source how do i increase body height ? Paraphrases

  • how could i increase my height ?
  • what should i do to increase body height ?
  • what are the ways to increase height ?
  • are there some ways to increase body height ?

Rephrasing a given text in multiple ways

Fidelity

(Meaning preserving)

Diversity

(Lexical & syntactical variety)

slide-14
SLIDE 14

Current State

slide-15
SLIDE 15

Current State

Synonym or phrase replacement

slide-16
SLIDE 16

Current State

Synonym or phrase replacement Source how do i increase body height ?

slide-17
SLIDE 17

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ?

slide-18
SLIDE 18

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ? Phrase how do i increase the body measurement vertically?

slide-19
SLIDE 19

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ? Phrase how do i increase the body measurement vertically?

Encoder Decoder

Sentence Paraphrase

slide-20
SLIDE 20

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ? Phrase how do i increase the body measurement vertically? Subsequence Selection - Beam Search (Top-k)

Encoder Decoder

Sentence Paraphrase

slide-21
SLIDE 21

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ? Phrase how do i increase the body measurement vertically? Subsequence Selection - Beam Search (Top-k) Source how do i increase body height ?

Encoder Decoder

Sentence Paraphrase

slide-22
SLIDE 22

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ? Phrase how do i increase the body measurement vertically? Subsequence Selection - Beam Search (Top-k) Source how do i increase body height ? Beam

  • how do i increase my height ?
  • how do i increase my body height ?
  • how do i increase the height ?
  • how would i increase my body height ?

Encoder Decoder

Sentence Paraphrase

slide-23
SLIDE 23

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ? Phrase how do i increase the body measurement vertically? Subsequence Selection - Beam Search (Top-k) Source how do i increase body height ? Beam

  • how do i increase my height ?
  • how do i increase my body height ?
  • how do i increase the height ?
  • how would i increase my body height ?

Fidelity

Encoder Decoder

Sentence Paraphrase

slide-24
SLIDE 24

Current State

Synonym or phrase replacement Source how do i increase body height ? Synonym how do i grow body height ? Phrase how do i increase the body measurement vertically? Subsequence Selection - Beam Search (Top-k) Source how do i increase body height ? Beam

  • how do i increase my height ?
  • how do i increase my body height ?
  • how do i increase the height ?
  • how would i increase my body height ?

Fidelity Diversity

Encoder Decoder

Sentence Paraphrase

slide-25
SLIDE 25

What can we do ?

slide-26
SLIDE 26

What can we do ?

Subsequence Selection - Beam Search (Diverse selection)

slide-27
SLIDE 27

What can we do ?

Subsequence Selection - Beam Search (Diverse selection) Source how do i increase body height ?

slide-28
SLIDE 28

What can we do ?

Subsequence Selection - Beam Search (Diverse selection) Source how do i increase body height ? Beam

  • how do i increase my height ?
  • how can i decrease my body weight ?
  • what do i do to increase the height ?
  • i am 17, what steps to take to decrease weight ?
slide-29
SLIDE 29

What can we do ?

Subsequence Selection - Beam Search (Diverse selection) Source how do i increase body height ? Beam

  • how do i increase my height ?
  • how can i decrease my body weight ?
  • what do i do to increase the height ?
  • i am 17, what steps to take to decrease weight ?

Diversity

slide-30
SLIDE 30

What can we do ?

Subsequence Selection - Beam Search (Diverse selection) Source how do i increase body height ? Beam

  • how do i increase my height ?
  • how can i decrease my body weight ?
  • what do i do to increase the height ?
  • i am 17, what steps to take to decrease weight ?

Diversity Fidelity

slide-31
SLIDE 31

What we need

slide-32
SLIDE 32

What we need

Fidelity Diversity

slide-33
SLIDE 33

What we need

Fidelity Diversity

DiPS

Find k diverse paraphrases with high fidelity Method based on subset 
 selection of candidate (sub)sequences

slide-34
SLIDE 34

Subset Selection

how do i increase my … how can i decrease the … how can i grow the … what ways exist to increase … how would I increase the … how do I decrease the … i am 17 , what … are there ways to increase …

Vt

slide-35
SLIDE 35

Subset Selection

how do i increase my … how can i decrease the … how can i grow the … what ways exist to increase … how would I increase the … how do I decrease the … i am 17 , what … are there ways to increase …

Vt

how do i increase my … how can i grow the … what ways exist to increase … are there ways to increase …

X

slide-36
SLIDE 36

Subset Selection

how do i increase my … how can i decrease the … how can i grow the … what ways exist to increase … how would I increase the … how do I decrease the … i am 17 , what … are there ways to increase …

Vt

how do i increase my … how can i grow the … what ways exist to increase … are there ways to increase …

X

argmaxX⊆Vt, |X|=kF(X)

k

slide-37
SLIDE 37

Subset Selection

how do i increase my … how can i decrease the … how can i grow the … what ways exist to increase … how would I increase the … how do I decrease the … i am 17 , what … are there ways to increase …

Vt

If is sub modular + monotone = Greedy algo. with 
 good bounds exists

F

how do i increase my … how can i grow the … what ways exist to increase … are there ways to increase …

X

argmaxX⊆Vt, |X|=kF(X)

k

slide-38
SLIDE 38

Sub-modularity

F = # Unique Coloured items

slide-39
SLIDE 39

Sub-modularity

# Items = 4 F = 2

F = # Unique Coloured items

slide-40
SLIDE 40

Sub-modularity

# Items = 4 F = 2

F = # Unique Coloured items

slide-41
SLIDE 41

Sub-modularity

# Items = 4 F = 2 # Items = 4 + 1 F = 2 + 1

F = # Unique Coloured items

slide-42
SLIDE 42

Sub-modularity

# Items = 4 F = 2 # Items = 4 + 1 F = 2 + 1

F = # Unique Coloured items

slide-43
SLIDE 43

Sub-modularity

# Items = 4 F = 2 # Items = 4 + 1 F = 2 + 1 # Items = 5 + 1 F = 3 + 0

F = # Unique Coloured items

slide-44
SLIDE 44

Sub-modularity

# Items = 4 F = 2 # Items = 4 + 1 F = 2 + 1 # Items = 5 + 1 F = 3 + 0

Diminishing Returns

F = # Unique Coloured items

slide-45
SLIDE 45

Monotonicity

A B A

F F

slide-46
SLIDE 46

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-47
SLIDE 47

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

Encoder

slide-48
SLIDE 48

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

Encoder Decoder

slide-49
SLIDE 49

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

Diversity Encoder Decoder

slide-50
SLIDE 50

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

Diversity Encoder Decoder Fidelity

slide-51
SLIDE 51

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-52
SLIDE 52

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-53
SLIDE 53

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-54
SLIDE 54

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-55
SLIDE 55

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-56
SLIDE 56

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-57
SLIDE 57

Diversity Components

Where How can can I I that that picture picture get find

Diversity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I

slide-58
SLIDE 58

Diversity Components

Where How can can I I that that picture picture get find

Diversity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage

N

n=1

βn ⋃

x∈X

xn−gram

N-gram uniqueness

Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I

slide-59
SLIDE 59

Diversity Components

Where How can can I I that that picture picture get find

Diversity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage

N

n=1

βn ⋃

x∈X

xn−gram

N-gram uniqueness

xi∈V(t) ∑ xj∈X

ℛ(xi, xj)

ℛ(xi, xj) = 1 − EditDistance(xi, xj) |xi| + |xj|

Structural Coverage

Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I

slide-60
SLIDE 60

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-61
SLIDE 61

DiPS

Induce Diversity while not compromising on Fidelity

<sos>

Where can I get that movie?

can

Where can I get that film?

I <eos>

How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I Where How can can I I that that picture picture get find get movie? Where can I Where can I that

k- sequences

Synonym (similar embeddings)

Diversity Components Fidelity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Source Sentence

Where

ENCODER DECODER

slide-62
SLIDE 62

Fidelity Components

Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I get movie? Where can I Where can I that

Synonym (similar embeddings)

Fidelity Components

s rage Source Sentence

slide-63
SLIDE 63

Fidelity Components

Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I get movie? Where can I Where can I that

Synonym (similar embeddings)

Fidelity Components

s rage Source Sentence

x∈X N

n=1

βn |xn-gram ∩ sn-gram|

Lexical Similarity

slide-64
SLIDE 64

Fidelity Components

Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I get movie? Where can I Where can I that

Synonym (similar embeddings)

Fidelity Components

s rage Source Sentence

x∈X

𝒯(x, s)

𝒯(x, s) = 1 |x| ∑

wi∈x

maxwj∈s ψ(vwi, vwj)

Embedding based Similarity

x∈X N

n=1

βn |xn-gram ∩ sn-gram|

Lexical Similarity

slide-65
SLIDE 65

DiPS Objective

Where How can can I I that that picture picture get find

Diversity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I get movie? Where can I Where can I that

Synonym (similar embeddings)

Fidelity Components

s rage Source Sentence

slide-66
SLIDE 66

DiPS Objective

Where How can can I I that that picture picture get find

Diversity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I get movie? Where can I Where can I that

Synonym (similar embeddings)

Fidelity Components

s rage Source Sentence

argmaxX⊆V, X=|k|F(X) F(X) = λ(μ1D1(X) + μ2D2(X)) + (1 − λ)(ν1L1(X, s) + ν2L2(X, s))

slide-67
SLIDE 67

DiPS Objective

Where How can can I I that that picture picture get find

Diversity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I get movie? Where can I Where can I that

Synonym (similar embeddings)

Fidelity Components

s rage Source Sentence

argmaxX⊆V, X=|k|F(X) F(X) = λ(μ1D1(X) + μ2D2(X)) + (1 − λ)(ν1L1(X, s) + ν2L2(X, s))

slide-68
SLIDE 68

DiPS Objective

Where How can can I I that that picture picture get find

Diversity Components

where , can , film , I , How , find that , that picture , .. I get , can I , Where can I

Rewards unique n-grams Rewards Structural Coverage Where can I get that film? How can I get that picture? : 3k Candidate Subsequences find film? Where can I that Where can I get movie? Where can I Where can I that

Synonym (similar embeddings)

Fidelity Components

s rage Source Sentence

argmaxX⊆V, X=|k|F(X) F(X) = λ(μ1D1(X) + μ2D2(X)) + (1 − λ)(ν1L1(X, s) + ν2L2(X, s))

slide-69
SLIDE 69

Fidelity & Diversity

(Quora Dataset)

slide-70
SLIDE 70

Fidelity & Diversity

(Quora Dataset)

27 30 33 36 SBS DBS VAE-SVG DPP SSR DiPS (Ours)

BLEU (Fidelity)

slide-71
SLIDE 71

Fidelity & Diversity

(Quora Dataset)

27 30 33 36 SBS DBS VAE-SVG DPP SSR DiPS (Ours)

BLEU (Fidelity)

30 39 48 57 66

Models

SBS DBS VAE-SVG DPP SSR DiPS (Ours)

4-Distinct (Diversity)

slide-72
SLIDE 72

Fidelity & Diversity

(Quora Dataset)

27 30 33 36 SBS DBS VAE-SVG DPP SSR DiPS (Ours)

BLEU (Fidelity)

30 39 48 57 66

Models

SBS DBS VAE-SVG DPP SSR DiPS (Ours)

4-Distinct (Diversity)

DiPS induces diversity without compromising on fidelity

slide-73
SLIDE 73

Quora Dataset

Accuracy

67 70 73 76

Models

No Aug SBS DPP SSR DBS DiPS (Ours)

LogReg SiameseLSTM

Data Augmentation Paraphrase Detection

slide-74
SLIDE 74

Quora Dataset

Accuracy

67 70 73 76

Models

No Aug SBS DPP SSR DBS DiPS (Ours)

LogReg SiameseLSTM

DiPS data augmentation helps in paraphrase detection

Data Augmentation Paraphrase Detection

slide-75
SLIDE 75

Data Augmentation for Intent Classification

Dataset : SNIPS

Accuracy

93 94 95 96 97 98

Models

  • No. Aug

SBS DBS

  • Syn. Rep
  • Cont. Aug

DiPS (Ours)

LogReg LSTM

slide-76
SLIDE 76

Data Augmentation for Intent Classification

Dataset : SNIPS

Accuracy

93 94 95 96 97 98

Models

  • No. Aug

SBS DBS

  • Syn. Rep
  • Cont. Aug

DiPS (Ours)

LogReg LSTM

Data augmentation using DiPS improves intent classification

slide-77
SLIDE 77

Dataset : Yahoo-L31

Accuracy

62 63 64 65 66 67

Models

No Aug. SBS DBS Syn.Rep

  • Cont. Aug.

DiPS (Ours)

LogReg LSTM

Data Augmentation for Intent Classification

slide-78
SLIDE 78

Dataset : Yahoo-L31

Accuracy

62 63 64 65 66 67

Models

No Aug. SBS DBS Syn.Rep

  • Cont. Aug.

DiPS (Ours)

LogReg LSTM

Data Augmentation for Intent Classification

Data augmentation using DiPS improves intent classification

slide-79
SLIDE 79

Conclusion

slide-80
SLIDE 80

Conclusion

Problem

Diversity in Paraphrases Without compromising

  • n fidelity
slide-81
SLIDE 81

Conclusion

Method

DiPS

Sub-modular

  • ptimisation

Problem

Diversity in Paraphrases Without compromising

  • n fidelity
slide-82
SLIDE 82

Conclusion

Method

DiPS

Sub-modular

  • ptimisation

Data Augmentation Using Paraphrasing

Take-Aways

Seq2Seq + Diversity

Problem

Diversity in Paraphrases Without compromising

  • n fidelity
slide-83
SLIDE 83

Code

https://github.com/malllabiisc/DiPS

slide-84
SLIDE 84

Code

https://github.com/malllabiisc/DiPS

Acknowledgement

slide-85
SLIDE 85

Code

https://github.com/malllabiisc/DiPS

Acknowledgement

slide-86
SLIDE 86

Code

https://github.com/malllabiisc/DiPS

Acknowledgement

ashutosh@iisc.ac.in, satwik55@gmail.com

Find us on

slide-87
SLIDE 87

Code

https://github.com/malllabiisc/DiPS

Thank you

Acknowledgement

ashutosh@iisc.ac.in, satwik55@gmail.com

Find us on