Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
How the Context Matters Language & Interaction in Dialogues
YUN-NUNG (VIVIAN) CHEN
 Introduction
 Word-Level Contexts in Sentences
 Learning from Prior Knowledge –
Knowledge-Guided Structural Attention ...
3
 Dialogue systems are intelligent agents that are able to help users finish tasks more efficiently
via conversational int...
LU and DM significantly benefit from contexts in sentences and in dialogues
Speech
Recognition
Language Understanding (LU)...
 Word-level context
 Prior knowledge such as linguistic syntax
 Collocated words
 Sentence-level context
6
Smartphone ...
7
Knowledge-Guided Structural Attention Network (K-SAN)
Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L....
 Syntax (Dependency Tree)  Semantics (AMR Graph)
8
show
me
the
flights
from
seattle
to
san
francisco
ROOT
1.
3.
4.
2.
Se...
 Prior knowledge as a teacher
9
knowledge-guided structure {xi}
Knowledge
Encoding
Sentence
Encoding
Inner
Product
u
mi
K...
 Syntax (Dependency Tree)  Semantics (AMR Graph)
10
show
me
the
flights
from
seattle
to
san
francisco
ROOT
1.
3.
4.
2.
1...
knowledge-guided structure {xi}
Knowledge
Encoding
Sentence
Encoding
Inner
Product
u
mi
Knowledge Attention Distribution
p...
12
ATIS Dataset
(F1 slot filling)
Small
(1/40)
Medium
(1/10)
Large
Tagger (GRU) 73.83 85.55 93.11
Encoder-Tagger (GRU) 72....
13
ATIS Dataset
(F1 slot filling)
Small
(1/40)
Medium
(1/10)
Large
Tagger (GRU) 73.83 85.55 93.11
Encoder-Tagger (GRU) 72....
14
ATIS Dataset
(F1 slot filling)
Small
(1/40)
Medium
(1/10)
Large
Tagger (GRU) 73.83 85.55 93.11
Encoder-Tagger (GRU) 72....
 Joint Intent Prediction and Slot Filling
15
knowledge-guided structure {xi}
Knowledge
Encoding
Sentence
Encoding
Inner
P...
16
ATIS Dataset
(train: 4478/dev:
500/test: 893)
Small (1/40) Medium (1/10) Large
Slot (Indep) Slot (Joint) Frame Slot (In...
17
When data is scare, K-SAN with joint parsing significantly improves the performance (slot & frame)
ATIS Dataset
(train:...
 Darker blocks and lines correspond to higher attention weights
18
Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J....
 Darker blocks and lines correspond to higher attention weights
Using less training data with K-SAN allows the model pay ...
20
Modularizing Unsupervised Sense Embeddings (MUSE)
G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embe...
 Word embeddings are trained on a corpus in an unsupervised manner
 Using the same embeddings for different senses for N...
Smartphone companies including blackberry, and sony will be invited.
 Input: unannotated text corpus
 Two key mechanisms...
 Efficient sense selection [Neelakantan et al.,
2014; Li and Jurafsky, 2015]
 Use word embeddings as input to update the...
G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.
 Sense selec...
 Learning algorithm
 Sense selection strategy
 Stochastic policy: selects the sense based on the probability distributi...
Approach MaxSimC AvgSimC
Huang et al., 2012 26.1 65.7
Neelakantan et al., 2014 60.1 69.3
Tian et al., 2014 63.6 65.4
Li & ...
 Dataset: SCWS for multi-sense embedding evaluation
27
Approach MaxSimC AvgSimC
Huang et al., 2012 26.1 65.7
Neelakantan ...
 Dataset: SCWS for multi-sense embedding evaluation
28
Approach MaxSimC AvgSimC
Huang et al., 2012 26.1 65.7
Neelakantan ...
 Dataset: SCWS for multi-sense embedding evaluation
29
Approach MaxSimC AvgSimC
Huang et al., 2012 26.1 65.7
Neelakantan ...
G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.
 Dataset: SC...
G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.
31
Approach E...
 KNN senses sorted by collocation likelihood
32
Context KNN Senses
… braves finish the season in tie with the los angeles...
33
Leveraging Behavior Patterns of Mobile Apps for
Personalized Spoken Language Understanding
Y.-N. Chen, M. Sun, A. I Rudnic...
 Task: user intent prediction
 Challenge: language ambiguity
 User preference
 Some people prefer “Message” to “Email”...
 Subjects’ app invocation is logged on a daily basis
 Subjects annotate their app activities with
 Task Structure: link...
37
1
Lexical Intended App
photo check CAMERA IMtell
take this photo
tell vivian this is me in the lab
CAMERA
IM
Train
chec...
 The decomposed matrices represent low-rank latent semantics for utterances and
words/histories/apps respectively
 The p...
 Model implicit feedback by completing the matrix
 not treat unobserved facts as negative samples (true or false)
 give...
40
1
Lexical Intended App
photo check CAMERA IMtell
take this photo
tell vivian this is me in the lab
CAMERA
IM
Train
chec...
 Dataset: 533 dialogues (1,607 utterances); 455 multi-turn dialogues
 Google recognized transcripts (word error rate = 2...
 Dataset: 533 dialogues (1,607 utterances); 455 multi-turn dialogues
 Google recognized transcripts (word error rate = 2...
 Dataset: 533 dialogues (1,607 utterances); 455 multi-turn dialogues
 Google recognized transcripts (word error rate = 2...
Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalize...
 App functionality modeling
 Learning app embeddings
45
App embeddings encoding functionality help user-independent unde...
End-to-End Memory Networks with Knowledge Carryover
for Multi-Turn Spoken Language Understanding
Y.-N. Chen, D. Hakkani-Tu...
4747
just sent email to bob about fishing this weekend
O O O O
B-contact_name
O
B-subject I-subject I-subject
U
S
I send_e...
48
u
Knowledge Attention Distributionpi
mi
Memory Representation
Weighted
Sum
h
∑ Wkg
o
Knowledge Encoding
Representation
...
Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Tu...
Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Tu...
 Dataset: Cortana communication session data
 GRU for all RNN
 adam optimizer
 embedding dim=150
 hidden unit=100
 d...
Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Tu...
Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Tu...
Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Tu...
55
Investigation of Language Understanding Impact for
Reinforcement Learning Based Dialogue Systems
X. Li, Y.-N. Chen, L. ...
 Dialogue management is framed as a reinforcement learning task
 Agent learns to select actions to maximize the expected...
 Dialogue management is framed as a reinforcement learning task
 Agent learns to select actions to maximize the expected...
 NLU and NLG are trained in a supervised manner
 DM is trained in a reinforcement learning framework (NLU and NLG can be...
 DM receives frame-level information
 No error model: perfect recognizer and LU
 Error model: simulate the possible err...
 User simulator sends natural language
 No recognition error
 Errors from NLG or LU
60
Natural Language
Generation (NLG...
 Frame-level semantics
61
 Natural language
The RL agent is able to learn how to interact with users to complete tasks m...
X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Le...
X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Le...
X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Le...
X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Le...
 Intent error rate  Slot error rate
66
The RL agent has better robustness to intent errors in terms of dialogue-level pe...
 Word-level contexts in sentences help understand word meanings
 Learning from Prior Knowledge –
K-SAN achieves better L...
Q & A
Upcoming SlideShare
Loading in …5
×

How the Context Matters Language and Interaction in Dialogues

903 views

Published on

We have envisioned that computers will understand natural language and predict what we need help in order to complete tasks via conversational interactions. This talk focuses on context-aware understanding in different levels: 1) word-level contexts in sentences 2) sentence-level contexts in dialogues. Word-level contexts contribute both semantic and syntactic relations, which benefit sense representation learning and knowledge-guided language understanding. Also, sentence-level contexts may significantly affect dialogue-level performance. This talk investigates how misunderstanding of a single-turn utterance degrades the success rate of an end-to-end reinforcement learning based dialogue system. Then we will highlight challenges and recent trends driven by deep learning and intelligent assistants.

Published in: Technology
  • DOWNLOAD FULL BOOKS, INTO AVAILABLE FORMAT ......................................................................................................................... ......................................................................................................................... 1.DOWNLOAD FULL. PDF EBOOK here { https://tinyurl.com/y3nhqquc } ......................................................................................................................... 1.DOWNLOAD FULL. EPUB Ebook here { https://tinyurl.com/y3nhqquc } ......................................................................................................................... 1.DOWNLOAD FULL. doc Ebook here { https://tinyurl.com/y3nhqquc } ......................................................................................................................... 1.DOWNLOAD FULL. PDF EBOOK here { https://tinyurl.com/y3nhqquc } ......................................................................................................................... 1.DOWNLOAD FULL. EPUB Ebook here { https://tinyurl.com/y3nhqquc } ......................................................................................................................... 1.DOWNLOAD FULL. doc Ebook here { https://tinyurl.com/y3nhqquc } ......................................................................................................................... ......................................................................................................................... ......................................................................................................................... .............. Browse by Genre Available eBooks ......................................................................................................................... Art, Biography, Business, Chick Lit, Children's, Christian, Classics, Comics, Contemporary, Cookbooks, Crime, Ebooks, Fantasy, Fiction, Graphic Novels, Historical Fiction, History, Horror, Humor And Comedy, Manga, Memoir, Music, Mystery, Non Fiction, Paranormal, Philosophy, Poetry, Psychology, Religion, Romance, Science, Science Fiction, Self Help, Suspense, Spirituality, Sports, Thriller, Travel, Young Adult,
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here

How the Context Matters Language and Interaction in Dialogues

  1. 1. How the Context Matters Language & Interaction in Dialogues YUN-NUNG (VIVIAN) CHEN
  2. 2.  Introduction  Word-Level Contexts in Sentences  Learning from Prior Knowledge – Knowledge-Guided Structural Attention Networks (K-SAN) [Chen et al., ‘16]  Learning from Observations – Modularizing Unsupervised Sense Embedding (MUSE) [Lee & Chen, ‘17]  Sentence-Level Contexts in Dialogues  Inference – Leveraging Behavioral Patterns for Personalized Understanding [Chen et al., ‘15] Contexts-Aware Spoken Language Understanding [Chen et al., ‘16]  Investigation of Understanding Impact – Reinforcement Learning Based Neural Dialogue System [Li et al., ‘17] Misunderstanding Impact [Li et al., ‘17]  Conclusion 2
  3. 3. 3
  4. 4.  Dialogue systems are intelligent agents that are able to help users finish tasks more efficiently via conversational interactions.  Dialogue systems are being incorporated into various devices (smart-phones, smart TVs, in- car navigating system, etc). 4 JARVIS – Iron Man’s Personal Assistant Baymax – Personal Healthcare Companion
  5. 5. LU and DM significantly benefit from contexts in sentences and in dialogues Speech Recognition Language Understanding (LU) • Domain Identification • User Intent Detection • Slot Filling Dialogue Management (DM) • Dialogue State Tracking (DST) • Dialogue Policy Natural Language Generation (NLG) Hypothesis are there any action movies to see this weekend Semantic Frame request_movie genre=action, date=this weekend System Action/Policy request_location Text response Where are you located? Text Input Are there any action movies to see this weekend? Speech Signal Backend Action / Knowledge Providers 5
  6. 6.  Word-level context  Prior knowledge such as linguistic syntax  Collocated words  Sentence-level context 6 Smartphone companies including apple, blackberry, and sony will be invited. show me the flights from seattle to san francisco (browsing movie reviews…) Good movies…. Find me a good action movie this weekend London Has Fallen is currently the number 1 action movie in America request_movie (genre=action, date=this weekend) How misunderstanding influences the dialogue system performance Contexts provide informative cues for better understanding How behavioral and sentence contexts influences the user intent
  7. 7. 7 Knowledge-Guided Structural Attention Network (K-SAN) Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016. Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Syntax or semantics? knowledge-guided joint semantic frame parsing,” in Proceedings of SLT, 2016.
  8. 8.  Syntax (Dependency Tree)  Semantics (AMR Graph) 8 show me the flights from seattle to san francisco ROOT 1. 3. 4. 2. Sentence s show me the flights from seattle to san francisco show you flight I 1. 2. 4. city city Seattle San Francisco 3. . Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016. Prior knowledge about syntax or semantics may guide understanding
  9. 9.  Prior knowledge as a teacher 9 knowledge-guided structure {xi} Knowledge Encoding Sentence Encoding Inner Product u mi Knowledge Attention Distribution pi Encoded Knowledge Representation Weighted Sum ∑ h o Knowledge-Guided Representation slot tagging sequence s y show me the flights from seattle to sanfrancisco ROOT Input Sentence ht-1 ht+1ht W W W W wt-1 yt-1 U wt U wt+1 U V yt V yt+1 V RNN Tagger Knowledge Encoding Module CNNkg CNNin NNout MMM Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016.
  10. 10.  Syntax (Dependency Tree)  Semantics (AMR Graph) 10 show me the flights from seattle to san francisco ROOT 1. 3. 4. 2. 1. show me 2. show flights the 3. show flights from seattle 4. show flights to francisco san Sentence sshow me the flights from seattle to san francisco Knowledge-Guided Substructure xi (s / show :ARG0 (y / you) :ARG1 (f / flight :source (c / city :name (d / name :op1 Seattle)) :destination (c2 / city :name (s2 / name :op1 San :op2 Francisco))) :ARG2 (i / I) :mode imperative) Knowledge-Guided Substructure xi 1. show you 2. show flight seattle 3. show flight san francisco 4. show i show you flight I 1. 2. 4. city city Seattle San Francisco 3. . Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016.
  11. 11. knowledge-guided structure {xi} Knowledge Encoding Sentence Encoding Inner Product u mi Knowledge Attention Distribution pi Encoded Knowledge Representation Weighted Sum ∑ h o Knowledge-Guided Representation slot tagging sequence s y show me the flights from seattle to sanfrancisco ROOT Input Sentence ht-1 ht+1ht W W W W wt-1 yt-1 U Mwt U wt+1 U V yt V yt+1 V MM RNN Tagger Knowledge Encoding Module CNNkg CNNin NNout 11 The model will pay more attention to more important substructures that may be crucial for slot tagging. Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016.
  12. 12. 12 ATIS Dataset (F1 slot filling) Small (1/40) Medium (1/10) Large Tagger (GRU) 73.83 85.55 93.11 Encoder-Tagger (GRU) 72.79 88.26 94.75 Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016.
  13. 13. 13 ATIS Dataset (F1 slot filling) Small (1/40) Medium (1/10) Large Tagger (GRU) 73.83 85.55 93.11 Encoder-Tagger (GRU) 72.79 88.26 94.75 K-SAN (Stanford dep) 74.60+ 87.99 94.86+ K-SAN (Syntaxnet dep) 74.35+ 88.40+ 95.00+ Syntax provides richer knowledge and more general guidance when less training data. Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016.
  14. 14. 14 ATIS Dataset (F1 slot filling) Small (1/40) Medium (1/10) Large Tagger (GRU) 73.83 85.55 93.11 Encoder-Tagger (GRU) 72.79 88.26 94.75 K-SAN (Stanford dep) 74.60+ 87.99 94.86+ K-SAN (Syntaxnet dep) 74.35+ 88.40+ 95.00+ K-SAN (AMR) 74.32+ 88.14 94.85+ K-SAN (JAMR) 74.27+ 88.27+ 94.89+ Syntax provides richer knowledge and more general guidance when less training data. Semantics captures the most salient info so it achieves similar performance with much less substructures Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks,” arXiv: 1609.00777, 2016.
  15. 15.  Joint Intent Prediction and Slot Filling 15 knowledge-guided structure {xi} Knowledge Encoding Sentence Encoding Inner Product u mi Knowledge Attention Distribution pi Encoded Knowledge Representation Weighted Sum ∑ h o Knowledge-Guided Representation s show me the flights from seattle to sanfrancisco ROOT Input Sentence RNN Tagger Knowledge Encoding Module CNNkg CNNin NNout slot tagging sequence y ht -1 ht+ 1 h tW W W W wt- 1 yt-1 U Mwt U wt+1 U V yt V yt+1 V MM EOS U Intent V ht+1 Extend the K-SAN for joint semantic frame parsing by outputting the user intent at last timestamp (Hakkani-Tur et al.). Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Syntax or semantics? knowledge-guided joint semantic frame parsing,” in Proceedings of SLT, 2016.
  16. 16. 16 ATIS Dataset (train: 4478/dev: 500/test: 893) Small (1/40) Medium (1/10) Large Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Tagger 73.8 73.0 85.6 86.4 93.1 93.4 Encoder-Tagger 72.8 71.9 88.3 87.5 94.8 93.1 K-SAN (Syntax) 74.4+ 74.6+ 88.4+ 88.2+ 95.0+ 95.4+ K-SAN (Semantics) 74.3+ 73.4+ 88.3 88.1+ 94.9+ 95.1+ Communication (train: 10479/dev: 1000/test: 2300) Small (1/40) Medium (1/10) Large Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Tagger 45.5 50.3 69.0 69.8 80.4 79.8 Encoder-Tagger 45.5 47.7 69.4 73.1 85.7 86.0 K-SAN (Syntax) 45.0 55.1+ 69.5+ 75.3+ 85.0 84.5 K-SAN (Semantics) 45.1 55.0 69.1 74.3+ 85.3 85.2 Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Syntax or semantics? knowledge-guided joint semantic frame parsing,” in Proceedings of SLT, 2016.
  17. 17. 17 When data is scare, K-SAN with joint parsing significantly improves the performance (slot & frame) ATIS Dataset (train: 4478/dev: 500/test: 893) Small (1/40) Medium (1/10) Large Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Tagger 73.8 73.0 33.5 85.6 86.4 58.5 93.1 93.4 79.7 Encoder-Tagger 72.8 71.9 35.2 88.3 87.5 61.9 94.8 93.1 82.5 K-SAN (Syntax) 74.4+ 74.6+ 37.6+ 88.4+ 88.2+ 63.5+ 95.0+ 95.4+ 84.3+ K-SAN (Semantics) 74.3+ 73.4+ 37.1+ 88.3 88.1+ 63.6+ 94.9+ 95.1+ 83.8+ Communication (train: 10479/dev: 1000/test: 2300) Small (1/40) Medium (1/10) Large Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Slot (Indep) Slot (Joint) Frame Tagger 45.5 50.3 48.9 69.0 69.8 68.2 80.4 79.8 79.5 Encoder-Tagger 45.5 47.7 52.7 69.4 73.1 71.4 85.7 86.0 83.9 K-SAN (Syntax) 45.0 55.1+ 57.2+ 69.5+ 75.3+ 73.5+ 85.0 84.5 84.5 K-SAN (Semantics) 45.1 55.0 54.1+ 69.1 74.3+ 73.8+ 85.3 85.2 83.4 Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Syntax or semantics? knowledge-guided joint semantic frame parsing,” in Proceedings of SLT, 2016.
  18. 18.  Darker blocks and lines correspond to higher attention weights 18 Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Syntax or semantics? knowledge-guided joint semantic frame parsing,” in Proceedings of SLT, 2016.
  19. 19.  Darker blocks and lines correspond to higher attention weights Using less training data with K-SAN allows the model pay the similar attention to the salient substructures that are important for tagging. 19 Y.-N. Chen, D. Hakkani-Tur, G. Tur, A. Celikyilmaz, J. Gao, and L. Deng, “Syntax or semantics? knowledge-guided joint semantic frame parsing,” in Proceedings of SLT, 2016.
  20. 20. 20 Modularizing Unsupervised Sense Embeddings (MUSE) G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.
  21. 21.  Word embeddings are trained on a corpus in an unsupervised manner  Using the same embeddings for different senses for NLP tasks, e.g. NLU, POS tagging 21 Finally I chose Google instead of Apple. Can you buy me a bag of apples, oranges, and bananas? G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017. Words with different senses should correspond different embeddings
  22. 22. Smartphone companies including blackberry, and sony will be invited.  Input: unannotated text corpus  Two key mechanisms  Sense selection given a text context  Sense representation to embed statistical characteristics of sense identity 22 G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017. apple apple-1 apple-2 sense selection sense embedding
  23. 23.  Efficient sense selection [Neelakantan et al., 2014; Li and Jurafsky, 2015]  Use word embeddings as input to update the sense posterior given words  Introduce ambiguity  Purely sense-level embedding [Qiu et al., 2016]  Inefficient sense selection  exponential time complexity 23 G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017. The prior approaches have disadvantages about either ambiguity or inefficiency
  24. 24. G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.  Sense selection  Policy-based  Value-based 24 Corpus: { Smartphone companies including apple blackberry, and sony will be invited.} sense selection ← rewardsignal← senseselection→ sample collocation1 2 2 3 Sense selection for collocated word 𝐶 𝑡′ Sense Selection Module …𝐶𝑡′ = 𝑤𝑗𝐶𝑡′−1 𝑞(𝑧𝑗1|𝐶𝑡′) 𝑞(𝑧𝑗2|𝐶𝑡′) 𝑞(𝑧𝑗3|𝐶𝑡′) matrix 𝑄𝑗 matrix 𝑃 … 𝐶𝑡′+1 apple andincluding sonyblackberry 𝑧𝑖1 Sense Representation Module …𝑃(𝑧𝑗2|𝑧𝑖1) 𝑃(𝑧 𝑢𝑣|𝑧𝑖1) negative sampling matrix 𝑉 matrix 𝑈  Sense representation learning  Skip-gram approximation Sense Selection Module …𝐶𝑡 = 𝑤𝑖𝐶𝑡−1 𝑞(𝑧𝑖1| 𝐶𝑡) 𝑞(𝑧𝑖2| 𝐶𝑡) 𝑞(𝑧𝑖3| 𝐶𝑡) Sense selection for target word 𝐶𝑡 matrix 𝑄𝑖 matrix 𝑃 … 𝐶𝑡+1 including apple blackberrycompanies and Collocated likelihood serves as a reward signal to optimize the sense selection module.
  25. 25.  Learning algorithm  Sense selection strategy  Stochastic policy: selects the sense based on the probability distribution  Greedy: selects the sense with the largest Q-value (no exploration)  ε-Greedy: selects a random sense with ε probability, and adopts the greedy strategy  Boltzmann: samples the sense based on the Boltzmann distribution modeled by Q-value 25 𝑧𝑖1 Sense Representation Module …𝑃(𝑧𝑗2|𝑧𝑖1) 𝑃(𝑧 𝑢𝑣|𝑧𝑖1) matrix 𝑈 matrix 𝑉 Sense Selection Module …𝐶𝑡 = 𝑤𝑖𝐶𝑡−1 𝑞(𝑧𝑖1| 𝐶𝑡) 𝑞(𝑧𝑖2| 𝐶𝑡) 𝑞(𝑧𝑖3| 𝐶𝑡) Sense selection for target word 𝐶𝑡 matrix 𝑄𝑖 matrix 𝑃 … 𝐶𝑡+1 including apple blackberrycompanies and G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.
  26. 26. Approach MaxSimC AvgSimC Huang et al., 2012 26.1 65.7 Neelakantan et al., 2014 60.1 69.3 Tian et al., 2014 63.6 65.4 Li & Jurafsky, 2015 66.6 66.8 Bartunov et al., 2016 53.8 61.2 Qiu et al., 2016 64.9 66.1  Dataset: SCWS for multi-sense embedding evaluation 26 He borrowed the money from banks. I live near to a river. correlation=? Baseline bank bank-1 bank-2 0.6 0.4 He borrowed the money from 0.6 x + 0.4 x MaxSimC AvgSimC G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.
  27. 27.  Dataset: SCWS for multi-sense embedding evaluation 27 Approach MaxSimC AvgSimC Huang et al., 2012 26.1 65.7 Neelakantan et al., 2014 60.1 69.3 Tian et al., 2014 63.6 65.4 Li & Jurafsky, 2015 66.6 66.8 Bartunov et al., 2016 53.8 61.2 Qiu et al., 2016 64.9 66.1 MUSE-Policy 66.1 67.4 G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017. He borrowed the money from banks. I live near to a river. correlation=?
  28. 28.  Dataset: SCWS for multi-sense embedding evaluation 28 Approach MaxSimC AvgSimC Huang et al., 2012 26.1 65.7 Neelakantan et al., 2014 60.1 69.3 Tian et al., 2014 63.6 65.4 Li & Jurafsky, 2015 66.6 66.8 Bartunov et al., 2016 53.8 61.2 Qiu et al., 2016 64.9 66.1 MUSE-Policy 66.1 67.4 MUSE-Greedy 66.3 68.3 G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017. He borrowed the money from banks. I live near to a river. correlation=?
  29. 29.  Dataset: SCWS for multi-sense embedding evaluation 29 Approach MaxSimC AvgSimC Huang et al., 2012 26.1 65.7 Neelakantan et al., 2014 60.1 69.3 Tian et al., 2014 63.6 65.4 Li & Jurafsky, 2015 66.6 66.8 Bartunov et al., 2016 53.8 61.2 Qiu et al., 2016 64.9 66.1 MUSE-Policy 66.1 67.4 MUSE-Greedy 66.3 68.3 MUSE-ε-Greedy 67.4+ 68.6 G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017. He borrowed the money from banks. I live near to a river. correlation=?
  30. 30. G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.  Dataset: SCWS for multi-sense embedding evaluation 30 Approach MaxSimC AvgSimC Huang et al., 2012 26.1 65.7 Neelakantan et al., 2014 60.1 69.3 Tian et al., 2014 63.6 65.4 Li & Jurafsky, 2015 66.6 66.8 Bartunov et al., 2016 53.8 61.2 Qiu et al., 2016 64.9 66.1 MUSE-Policy 66.1 67.4 MUSE-Greedy 66.3 68.3 MUSE-ε-Greedy 67.4+ 68.6 MUSE-Boltzmann 67.9+ 68.7 MUSE with exploration achieves the best sense embeddings in MaxSimC. He borrowed the money from banks. I live near to a river. correlation=?
  31. 31. G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017. 31 Approach ESL-50 RD-300 TOEFL-80 Global Context 47.73 45.07 60.87 SkipGram 52.08 55.66 66.67 IMS+SkipGram 41.67 53.77 66.67 EM 27.08 33.96 40.00 MSSG (Neelakantan et al., 2014) 57.14 58.93 78.26 CRP (Li & Jurafsky, 2015) 50.00 55.36 82.61 MUSE-Policy 52.38 51.79 79.71 MUSE-Greedy 57.14 58.93 79.71 MUSE-ε-Greedy 61.90+ 62.50+ 84.06+ MUSE-Boltzmann 64.29+ 66.07+ 88.41+ Retro-GlobalContext 63.64 66.20 71.01 Retro-SkipGram 56.25 65.09 73.33 Conventional Word Embedding Word Sense Disambiguation Unsupervised Sense Embedding Supervised Sense Embedding MUSE with exploration achieves the state-of-the-art results for synonym selection.
  32. 32.  KNN senses sorted by collocation likelihood 32 Context KNN Senses … braves finish the season in tie with the los angeles dodgers … scoreless otl shootout 6-6 hingis 3-3 7-7 0-0 … his later years proudly wore tie with the chinese characters for … pants trousers shirt juventus blazer socks anfield … of the mulberry or the blackberry and minos sent him to … cranberries maple vaccinium apricot apple … of the large number of blackberry users in the us federal … smartphones sap microsoft ipv6 smartphone … ladies wore extravagant head ornaments combs pearl necklaces face … venter thorax neck spear millimeters fusiform … appoint john pope republican as head of the new army of … multi-party appoints unicameral beria appointed MUSE learns sense embeddings in an unsupervised way and achieves the first purely sense-level representation learning system with linear-time sense selection G.-H. Lee and Y.-N. Chen, “MUSE: Modularizing Unsupervised Sense Embeddings,” in Proceedings of EMNLP, 2017.
  33. 33. 33
  34. 34. Leveraging Behavior Patterns of Mobile Apps for Personalized Spoken Language Understanding Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  35. 35.  Task: user intent prediction  Challenge: language ambiguity  User preference  Some people prefer “Message” to “Email”  Some people prefer “Outlook” to “Gmail”  App-level contexts  “Message” is more likely to follow “Camera”  “Email” is more likely to follow “Excel” 35 send to vivian v.s. Email? Message? Communication Considering behavioral patterns in history to model SLU for intent prediction. Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  36. 36.  Subjects’ app invocation is logged on a daily basis  Subjects annotate their app activities with  Task Structure: link applications that serve a common goal  Task Description: briefly describe the goal or intention of the task  Subjects use a wizard system to perform the annotated task by speech 36 TASK59; 20150203; 1; Tuesday; 10:48 play music via bluetooth speaker com.android.settings  com.lge.music Meta Desc App : Ready. : Connect my phone to bluetooth speaker. : Connected to bluetooth speaker. : And play music. : What music would you like to play? : Shuffle playlist. : I will play the music for you. W1 U1 W2 U2 W3 U3 W4 Dialogue SETTINGS MUSIC MUSIC Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  37. 37. 37 1 Lexical Intended App photo check CAMERA IMtell take this photo tell vivian this is me in the lab CAMERA IM Train check my grades on website send an email to professor … CHROME EMAIL send Behavioral NULL CAMERA .85 take a photo of this send it to alice CAMERA IM … email 1 1 1 1 1 1 .70 CHROME 1 1 1 1 1 1 CHROME EMAIL 1 1 1 1 .95 .80 .55 User Utterance Intended App Test take a photo of this send it to alex … hidden semantics Issue: unobserved hidden semantics may benefit understanding Solution: use matrix factorization to complete a partially-missing matrix based on a low-rank latent semantics assumption. Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  38. 38.  The decomposed matrices represent low-rank latent semantics for utterances and words/histories/apps respectively  The product of two matrices fills the probability of hidden semantics 38 1 Lexical Intended App photo check CAMERA IMtell send Behavioral NULL CAMERA .85 email 1 1 1 1 1 1 .70 CHROME 1 1 1 1 1 1 CHROME EMAIL 1 1 1 1 .95 .80 .55 𝑼 𝑾 + 𝑯 + 𝑨 ≈ 𝑼 × 𝒅 𝒅 × 𝑾 + 𝑯 + 𝑨 Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  39. 39.  Model implicit feedback by completing the matrix  not treat unobserved facts as negative samples (true or false)  give observed facts higher scores than unobserved facts  Objective:  the model can be achieved by SGD updates with fact pairs 39 1 𝑓+ 𝑓− 𝑓− 𝑢 𝑥 The objective is to learn a set of well-ranked apps per utterance. Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  40. 40. 40 1 Lexical Intended App photo check CAMERA IMtell take this photo tell vivian this is me in the lab CAMERA IM Train check my grades on website send an email to professor … CHROME EMAIL send Behavioral NULL CAMERA .85 take a photo of this send it to alice CAMERA IM … email 1 1 1 1 1 1 .70 CHROME 1 1 1 1 1 1 CHROME EMAIL 1 1 1 1 .95 .80 .55 User Utterance Intended App Reasoning with Matrix Factorization for Implicit Intents Test take a photo of this send it to alex … Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  41. 41.  Dataset: 533 dialogues (1,607 utterances); 455 multi-turn dialogues  Google recognized transcripts (word error rate = 25%)  Evaluation metric: accuracy of user intent prediction (ACC) mean average precision of ranked intents (MAP)  Baseline: Maximum Likelihood Estimation (MLE) Multinomial Logistic Regression (MLR) 41 Approach Lexical Behavioral All (a) MLE User-Indep 13.5 / 19.6 (b) User-Dep 20.2 / 27.9 The user-dependent model is better than the user-independent model. Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  42. 42.  Dataset: 533 dialogues (1,607 utterances); 455 multi-turn dialogues  Google recognized transcripts (word error rate = 25%)  Evaluation metric: accuracy of user intent prediction (ACC) mean average precision of ranked intents (MAP)  Baseline: Maximum Likelihood Estimation (MLE); Multinomial Logistic Regression (MLR) 42 Approach Lexical Behavioral All (a) MLE User-Indep 13.5 / 19.6 (b) User-Dep 20.2 / 27.9 (c) MLR User-Indep 42.8 / 46.4 14.9 / 18.7 (d) User-Dep 48.2 / 52.1 19.3 / 25.2 Lexical features are useful to predict intended apps for both user-independent and user-dependent models. Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  43. 43.  Dataset: 533 dialogues (1,607 utterances); 455 multi-turn dialogues  Google recognized transcripts (word error rate = 25%)  Evaluation metric: accuracy of user intent prediction (ACC) mean average precision of ranked intents (MAP)  Baseline: Maximum Likelihood Estimation (MLE); Multinomial Logistic Regression (MLR) 43 Approach Lexical Behavioral All (a) MLE User-Indep 13.5 / 19.6 (b) User-Dep 20.2 / 27.9 (c) MLR User-Indep 42.8 / 46.4 14.9 / 18.7 46.2+ / 50.1+ (d) User-Dep 48.2 / 52.1 19.3 / 25.2 50.1+ / 53.9+ Combining lexical and behavioral features shows imporvement, which models explicit information from observations. Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  44. 44. Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.  Dataset: 533 dialogues (1,607 utterances); 455 multi-turn dialogues  Google recognized transcripts (word error rate = 25%)  Evaluation metric: accuracy of user intent prediction (ACC) mean average precision of ranked intents (MAP)  Baseline: Maximum Likelihood Estimation (MLE); Multinomial Logistic Regression (MLR) 44 Approach Lexical Behavioral All (a) MLE User-Indep 13.5 / 19.6 (b) User-Dep 20.2 / 27.9 (c) MLR User-Indep 42.8 / 46.4 14.9 / 18.7 46.2+ / 50.1+ (d) User-Dep 48.2 / 52.1 19.3 / 25.2 50.1+ / 53.9+ (e) (c) + Personalized MF 47.6 / 51.1 16.4 / 20.3 50.3+* / 54.2+* (f) (d) + Personalized MF 48.3 / 52.7 20.6 / 26.7 51.9+* / 55.7+* Personalized MF significantly improves MLR results by considering hidden semantics.
  45. 45.  App functionality modeling  Learning app embeddings 45 App embeddings encoding functionality help user-independent understanding Y.-N. Chen, M. Sun, A. I Rudnicky, and A. Gershman, "Leveraging Behavioral Patterns of Mobile Applications for Personalized Spoken Language Understanding," in Proc. of ICMI, pages 83-86, 2015. ACM.
  46. 46. End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA.
  47. 47. 4747 just sent email to bob about fishing this weekend O O O O B-contact_name O B-subject I-subject I-subject U S I send_emailD communication  send_email(contact_name=“bob”, subject=“fishing this weekend”) are we going to fish this weekend U1 S2  send_email(message=“are we going to fish this weekend”) send email to bob U2  send_email(contact_name=“bob”) B-message I-message I-message I-message I-message I-message I-message B-contact_nameS1 Domain Identification  Intent Prediction  Slot Filling Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA.
  48. 48. 48 u Knowledge Attention Distributionpi mi Memory Representation Weighted Sum h ∑ Wkg o Knowledge Encoding Representation history utterances {xi} current utterance c Inner Product Sentence Encoder RNNin x1 x2 xi… Contextual Sentence Encoder x1 x2 xi… RNNmem slot tagging sequence y ht-1 ht V V W W W wt-1 wt yt-1 yt U U RNN Tagger M M Idea: additionally incorporating contextual knowledge during slot tagging 1. Sentence Encoding 2. Knowledge Attention 3. Knowledge Encoding Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA.
  49. 49. Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA. 49 u Knowledge Attention Distributionpi mi Memory Representation Weighted Sum h ∑ Wkg o Knowledge Encoding Representation history utterances {xi} current utterance c Inner Product Sentence Encoder RNNin x1 x2 xi… Contextual Sentence Encoder x1 x2 xi… RNNmem slot tagging sequence y ht-1 ht V V W W W wt-1 wt yt-1 yt U U RNN Tagger M M Idea: additionally incorporating contextual knowledge during slot tagging 1. Sentence Encoding 2. Knowledge Attention 3. Knowledge Encoding CNN CNN The model can automatically focus on salient contexts for current understanding
  50. 50. Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA.  Dataset: Cortana communication session data  GRU for all RNN  adam optimizer  embedding dim=150  hidden unit=100  dropout=0.5 50 Model Training Set Knowledge Encoding Sentence Encoder First Turn Other Overall RNN Tagger single-turn x x 60.6 16.2 25.5 The model trained on single-turn data performs worse for non-first turns due to mismatched training data
  51. 51.  Dataset: Cortana communication session data  GRU for all RNN  adam optimizer  embedding dim=150  hidden unit=100  dropout=0.5 51 Model Training Set Knowledge Encoding Sentence Encoder First Turn Other Overall RNN Tagger single-turn x x 60.6 16.2 25.5 multi-turn x x 55.9 45.7 47.4 Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA. Treating multi-turn data as single-turn for training performs reasonable
  52. 52. Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA.  Dataset: Cortana communication session data  GRU for all RNN  adam optimizer  embedding dim=150  hidden unit=100  dropout=0.5 52 Model Training Set Knowledge Encoding Sentence Encoder First Turn Other Overall RNN Tagger single-turn x x 60.6 16.2 25.5 multi-turn x x 55.9 45.7 47.4 Encoder- Tagger multi-turn current utt (c) RNN 57.6 56.0 56.3 multi-turn history + current (x, c) RNN 69.9 60.8 62.5 Directly encoding contexts improves the performance but increases training time
  53. 53. Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA.  Dataset: Cortana communication session data  GRU for all RNN  adam optimizer  embedding dim=150  hidden unit=100  dropout=0.5 53 Model Training Set Knowledge Encoding Sentence Encoder First Turn Other Overall RNN Tagger single-turn x x 60.6 16.2 25.5 multi-turn x x 55.9 45.7 47.4 Encoder- Tagger multi-turn current utt (c) RNN 57.6 56.0 56.3 multi-turn history + current (x, c) RNN 69.9 60.8 62.5 Proposed multi-turn history + current (x, c) RNN 73.2 65.7 67.1 Memory networks significantly outperforms all approaches with much less training time
  54. 54. Y.-N. Chen, D. Hakkani-Tur, G. Tur, J. Gao, and L. Deng, “End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding," in Proc. of Interspeech, 2016. ISCA.  Dataset: Cortana communication session data  GRU for all RNN  adam optimizer  embedding dim=150  hidden unit=100  dropout=0.5 54 Model Training Set Knowledge Encoding Sentence Encoder First Turn Other Overall RNN Tagger single-turn x x 60.6 16.2 25.5 multi-turn x x 55.9 45.7 47.4 Encoder- Tagger multi-turn current utt (c) RNN 57.6 56.0 56.3 multi-turn history + current (x, c) RNN 69.9 60.8 62.5 Proposed multi-turn history + current (x, c) RNN 73.2 65.7 67.1 multi-turn history + current (x, c) CNN 73.8 66.5 68.0 CNN produces comparable results for sentence encoding with shorter training time
  55. 55. 55 Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems X. Li, Y.-N. Chen, L. Li, and J. Gao, “End-to-End Task-Completion Neural Dialogue Systems,” arXiv: 1703.01008, 2017. X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems,” arXiv: 1703.07055, 2017.
  56. 56.  Dialogue management is framed as a reinforcement learning task  Agent learns to select actions to maximize the expected reward 56 Environment Observation Action Reward If booking a right ticket, reward = +30 If failing, reward = -30 Otherwise, reward = -1 Agent X. Li, Y.-N. Chen, L. Li, and J. Gao, “End-to-End Task-Completion Neural Dialogue Systems,” arXiv: 1703.01008, 2017.
  57. 57.  Dialogue management is framed as a reinforcement learning task  Agent learns to select actions to maximize the expected reward 57 Environment Observation Action Agent Natural Language Generation User Agenda Modeling User Simulator Language Understanding Dialogue Management Neural Dialogue System Text Input: Are there any action movies to see this weekend? Dialogue Policy: request_location X. Li, Y.-N. Chen, L. Li, and J. Gao, “End-to-End Task-Completion Neural Dialogue Systems,” arXiv: 1703.01008, 2017.
  58. 58.  NLU and NLG are trained in a supervised manner  DM is trained in a reinforcement learning framework (NLU and NLG can be fine tuned) 58 wi B- type wi+ 1 wi+2 O O EO S <intent> wi B- type wi+ 1 wi+2 O O EO S <intent> Dialogue Policy request_location User Dialogue Action Inform(location=San Francisco) Time t-1 wi <slot> wi+ 1 wi+2 O O EO S <intent > Language Understanding Time t-2 Time t Dialogue Management w 0 w1 w2 Natural Language Generation EO S User Goal User Agenda Modeling User Simulator End-to-End Neural Dialogue System Text Input Are there any action movies to see this weekend? Semantic Frame request_movie genre=action, date=this weekend X. Li, Y.-N. Chen, L. Li, and J. Gao, “End-to-End Task-Completion Neural Dialogue Systems,” arXiv: 1703.01008, 2017.
  59. 59.  DM receives frame-level information  No error model: perfect recognizer and LU  Error model: simulate the possible errors 59 Error Model • Recognition error • LU error Dialogue State Tracking (DST) system dialogue acts Dialogue Policy Optimization Dialogue Management (DM) User Model User Simulation user dialogue acts (semantic frames) X. Li, Y.-N. Chen, L. Li, and J. Gao, “End-to-End Task-Completion Neural Dialogue Systems,” arXiv: 1703.01008, 2017.
  60. 60.  User simulator sends natural language  No recognition error  Errors from NLG or LU 60 Natural Language Generation (NLG) Dialogue State Tracking (DST) system dialogue acts Dialogue Policy Optimization Dialogue Management (DM) User Model User Simulation NL Language Understanding (LU) X. Li, Y.-N. Chen, L. Li, and J. Gao, “End-to-End Task-Completion Neural Dialogue Systems,” arXiv: 1703.01008, 2017.
  61. 61.  Frame-level semantics 61  Natural language The RL agent is able to learn how to interact with users to complete tasks more efficiently and effectively, and outperforms the rule-based agent. X. Li, Y.-N. Chen, L. Li, and J. Gao, “End-to-End Task-Completion Neural Dialogue Systems,” arXiv: 1703.01008, 2017.
  62. 62. X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems,” arXiv: 1703.07055, 2017. 62 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 12 24 36 48 60 72 84 96 108 120 132 144 156 168 180 192 204 216 228 240 252 264 276 288 300 312 324 336 348 360 372 384 396 408 420 432 444 456 468 480 492 SuccessRate Simulation Epoch Learning Curve of System Performance Upper Bound DQN - 0.00 Rule - 0.00 RL Agent w/o LU errors Rule-Based Agent w/o LU errors
  63. 63. X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems,” arXiv: 1703.07055, 2017. 63 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 12 24 36 48 60 72 84 96 108 120 132 144 156 168 180 192 204 216 228 240 252 264 276 288 300 312 324 336 348 360 372 384 396 408 420 432 444 456 468 480 492 SuccessRate Simulation Epoch Learning Curve of System Performance Upper Bound DQN - 0.00 DQN - 0.05 Rule - 0.00 Rule - 0.05 RL Agent w/o LU errors RL Agent w/ 5% LU errors Rule-Based Agent w/o LU errors Rule-Based Agent w/ 5% LU errors >5% performance drop The system performance is sensitive to LU errors (sentence-level contexts), for both rule-based and RL agents.
  64. 64. X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems,” arXiv: 1703.07055, 2017.  Intent error type  I0: random  I1: within group  I2: between group  Intent error rate  I3: 0.00  I4: 0.10  I5: 0.20 64 Intent errors slightly influence the RL system performance Group 1: greeting(), thanks(), etc Group 2: inform(xx) Group 3: request(xx) Between-group intent errors degrade the system performance more request_moviename(actor=Robert Downey Jr) request_year
  65. 65. X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems,” arXiv: 1703.07055, 2017.  Slot error type  I0: random  I1: slot deletion  I2: value substitution  I3: slot substitution  Slot error rate  S4: 0.00  S5: 0.10  S6: 0.20 65 Slot errors significantly degrade the RL system performance Value substitution has the largest impact on the system performance request_moviename (actor=Robert Downey Jr) director Robert Downey Sr
  66. 66.  Intent error rate  Slot error rate 66 The RL agent has better robustness to intent errors in terms of dialogue-level performance Slot filling is more important than intent detection in language understanding X. Li, Y.-N. Chen, L. Li, J. Gao, and A. Celikyilmaz, “Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems,” arXiv: 1703.07055, 2017.
  67. 67.  Word-level contexts in sentences help understand word meanings  Learning from Prior Knowledge – K-SAN achieves better LU via known knowledge [Chen et al., ‘16]  Learning from Observations – MUSE learns sense embeddings with efficient sense selection [Lee & Chen, ‘17]  Sentence-level contexts have different impacts on dialogue performance  Inference – App contexts improve personalized understanding via inference [Chen et al., ‘15] Contexts for knowledge carryover benefits multi-turn LU [Chen et al., ‘16]  Investigation of Understanding Impact – Slot errors degrade system performance more than intent errors [Li et al., ‘17]  Contexts from different levels provide cues for better understanding in supervised and unsupervised ways 67
  68. 68. Q & A

×