Incremental on-line adaptation of POMDP-based dialogue managers to extended domains

By Milica Gasic, Dongho Kim, Pirros Tsiakoulis, Catherine Breslin, Matthew Henderson, Martin Szummer, Blaise Thomson, Susannah Young. 2014

65 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Dialogue & Conversational AI

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

In Computer Vision and Pattern Recognition

In Dialogue & Conversational AI

Citations in the corpus

Deep Reinforcement Learning for Dialogue Generation (2016, background): “Efforts include statistical models such as Markov Decision Processes (MDPs) (Levin et al., 1997; Levin et al., 2000; Walker et al., 2003; Pieraccini et al., 2009), POMDP (Young et al., 2010; Young et al., 2013; Gašic et al., 2013a; Gašic et al., 2014) models, and models that statistically learn ge”
Neural Approaches to Conversational AI (2018, neutral): “This problem, referred to as domain extension (Gašic et al., 2014), makes exploration even more challenging: the agent needs to explicitly quantify the uncertainty in its parameters for intents/slots, so as to explore new ones more aggressively while avoiding exploring those that have already been l”
Semantically Conditioned LSTM-based Natural Language Generation for Spoken Dialogue Systems (2015, background): “Furthermore, the approach does not easily scale to large open domain systems(Young et al., 2013; Gaˇsi´c et al., 2014; Henderson et al., 2014).”
Learning End-to-End Goal-Oriented Dialog (2016, methodology): “Datasets are often based on interactions between users and existing systems (or ensemble of systems) like DSTC datasets, SFCore (Gašic et al., 2014) or ATIS (Dahl et al.”
Continual Lifelong Learning in Natural Language Processing: A Survey (2020, background): “Research in dialogue agents, which are able to continually learn, is a very active area (e.g. Gasic et al., 2014; Su et al., 2016).”
Incremental Learning from Scratch for Task-Oriented Dialogue Systems (2019, background): “Gaˇsic et al. (2014) explicitly deﬁned kernel functions between belief states from different domains to extend the domain of dialogue systems.”
MultiWOZ - A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modelling (2018, methodology): “To illustrate the contribution of the new corpus, we compare it on several important statistics with the DSTC2 corpus (Henderson et al., 2014a), the SFX corpus (Gaˇsi´c et al., 2014), the WOZ 2 .”
Multi-domain Dialog State Tracking using Recurrent Neural Networks (2015, methodology): “tels data was collected during the Parlance project (Gašić et al., 2014).”
Efﬁcient Exploration for Dialogue Policy Learning with BBQ Networks & Replay Buffer Spiking (2016, methodology): “Our domain-extension experiments take inspiration from Gašic et al. [2014] and our agenda-based user simulator is modeled on Schatzmann et al. [2007].”
Transfer Learning for User Adaptation in Spoken Dialogue Systems (2016, methodology): “Representation transfer [37] is an active research area: one can find studies on option based transfers [20, 30], which are dependant on hierarchical reinforcement learning [36], on Proto-value functions [29, 15], or on representation transfer for Gaussian Processes Reinforcement Learning [16, 17, 5”
Variational Cross-domain Natural Language Generation for Spoken Dialogue Systems (2018, background): “Recently, statistical approaches have been studied that allow natural, efﬁcient and more diverse interaction with users without depending on pre-deﬁned rules (Young et al., 2013; Gaˇsi´c et al., 2014; Henderson et al., 2014).”
A Teacher-Student Framework for Maintainable Dialog Manager (2018, background): “To extend the domain of dialog systems, Gaˇsic et al. (2014) explicitly deﬁned kernel functions between the belief states that come from different domains.”
Learning from real users: rating dialogue success with neural networks for reinforcement learning in spoken dialogue systems (2015, background): “In all previous works the training of an SDS has been done with either recruited subjects [4, 5] who are presented with a predefined task to complete, or via simulated users [6, 7, 8, 9, 10] who randomly sample a goal over the specific ontology.”
ChartDialogs: Plotting from Natural Language Instructions (2020, unknown): “In particular, we compare to: DSTC2 (Henderson et al., 2014), SFX (Gaˇsi´c et al., 2014), WOZ (Wen et al., 2017), FRAMES (Asri et al., 2017), KVRET (Eric and Manning, 2017), M2M (Shah et al., 2018) and Im-ageEdits (Manuvinakurike et al., 2018b,a).”