Generative and discriminative algorithms for spoken language understanding

By Christian Raymond, Giuseppe Riccardi. Interspeech, 2007

339 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/9a592e2dd099aa42b5595b243b71a2c6fc6257f8

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Attention-Based Recurrent Neural Network Models for Joint Intent Detection and Slot Filling (2016, background): “Popular approaches to solving sequence labeling problems include maximum entropy Markov models (MEMMs) [3], conditional random fields (CRFs) [4], and recurrent neural networks (RNNs) [5, 6, 7].”
A Novel Bi-directional Interrelated Model for Joint Intent Detection and Slot Filling (2019, background): “CRF Layer : From Table 2 it can be seen that the CRF layer has a positive effect on the general model performance.”
Multi-Domain Joint Semantic Frame Parsing Using Bi-Directional RNN-LSTM (2016, positive): “Similarly, slot filling is framed as a sequence classification problem and hidden Markov models [12] and conditional random fields [13, 14] have been employed.”
Recurrent Neural Networks with External Memory for Spoken Language Understanding (2015, methodology): “The widely used approaches include conditional random ﬁelds (CRFs) [19,13], support vector machine [12], and, more recently, RNNs [25,14].”
Joint semantic utterance classification and slot filling with recursive neural networks (2014, methodology): “For slot filling, conditional random field (CRF) [10] is a proven technique and has been used extensively [11], but recently recurrent neural networks (RNNs) have also shown excellent performance [12, 13, 14].”
Leveraging Sentence-level Information with Encoder LSTM for Semantic Slot Filling (2016, methodology): “Yao et al. (2013) used RNN and outperformed methods that did not use neural networks, such as SVM (Raymond and Riccardi, 2007) and CRF (Deng et al., 2012).”
Improving Spoken Dialogue Understanding Using Phonetic Mixture Models (2011, background): “However it comes at a cost of an extremely narrow bandwidth for communication between the components: often the only information that passes from the speech recognizer to the NLU is a string of words, and typical NLU components use only word-level features (Litman et al. 2009; Raymond and Riccardi 2”
Multi-domain spoken language understanding with transfer learning (2009, background): “Moreover, many recent studies have achieved state-ofthe-art results on SLU tasks using conditional probabilistic models (Jeong and Lee, 2006; Moschitti et al., 2007; Raymond and Riccardi, 2007; Wang et al., 2005).”
Spoken language understanding from unaligned data using discriminative classification models (2009, background): “Index Terms— semantic analysis, spoken dialogue systems, spoken language understanding”
What is left to be understood in ATIS? (2010, methodology): “7% using the representation used by [10], who reported 95.”
Application of Deep Belief Networks for Natural Language Understanding (2014, positive): “Ground-breaking advances in speech recognition technology from early 1980 s to early 1990 s opened the way for spoken language understanding.”
Sample Efficient Deep Reinforcement Learning for Dialogue Systems With Large Action Spaces (2018, background): “Support Vector Machines, Conditional Random Fields and Recurrent Neural Networks with Attention have been applied to this problem with varying levels of success [25].”
Random features for Kernel Deep Convex Network (2013, methodology): “The result is obtained using only lexical features, with default parameters of CRF++ toolkit, following [18].”
User Information Augmented Semantic Frame Parsing Using Progressive Neural Networks (2018, methodology): “As a sequence labeling problem, slot ﬁlling can be solved using traditional machine learning approaches including maximum entropy Markov model [3] and conditional random ﬁelds (CRF) [14], as well as recurrent neural network (RNN) based approaches which takes and tags each word in an utterance one by”
Neural Belief Tracker: Data-Driven Dialogue State Tracking (2016, methodology): “These are used as input to the ﬁnal decision-making module which decides whether the user expressed the intent represented by the current candidate slot-value pair.”
Spoken language understanding using long short-term memory neural networks (2014, background): “In this paper we focus on spoken language understanding (SLU), in particular, word labeling with semantic information [16] [17] [18] [19] [20] [21] [22].”
Multi-Domain Adversarial Learning for Slot Filling in Spoken Language Understanding (2017, methodology): “Popular methods for solving sequence tagging include maximum entropy Markov models (MEMMs) [3], conditional random ﬁelds (CRFs) [4], and recurrent neural networks (RNNs) [5, 6, 7].”