Download presentation
Presentation is loading. Please wait.
Published byTobias Clark Modified over 8 years ago
1
A Sentence Interaction Network for Modeling Dependence between Sentences Biao Liu, Minlie Huang Tsinghua University
2
Motivation In answer selection: The semantic relations between sentences is crucial for some NLP tasks.
3
Motivation We want to model the semantic relation between two sentences. What do cats look like? Cats have large eyes and furry bodies.
4
Motivation Convolutional Neural Network Architectures for Matching Natural Language Sentences
5
Motivation Dynamic Pooling and Unfolding Recursive Autoencoders for Paraphrase Detection
6
Background RNN
7
Background LSTM
8
Method Step 1 Use a LSTM to model the two sentences
9
Method Step 2: Introduce interactions
10
Method Step 2
11
Method Much complex semantic relations can be modeled with a vector. Through gates, different words can have different weights for classification.
12
Method SIN It’s powerful to model interactions between words, but not strong enough for phrase interactions. SIN-CONV Add a convolution layer to model phrases.
13
Experiments Answer Selection To select correct answers from a set of candidates for a given question.
14
Experiments Answer Selection Results
15
Experiments Dialogue Act Analysis To identify the dialogue act of a sentence in a dialogue.
16
Experiments Dialogue Act Analysis
17
Interaction Analysis Interaction Mechanism Analysis
18
Thanks
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.