Papers by Milica Gašić
Dynamic Dialogue Policy for Continual Reinforcement Learning (2022.coling-1)
Copied to clipboard
Christian Geishauser, Carel van Niekerk, Hsien-chin Lin, Nurul Lubis, Michael Heck, Shutong Feng, Milica Gašić
| Challenge: | Continual reinforcement learning of the dialogue policy has remained unaddressed . lack of a framework with training protocols, baseline models and suitable metrics has hindered research in this direction. |
| Approach: | They propose a continual learning algorithm, baseline architectures and metrics for assessing continual reinforcement learning models. |
| Outcome: | The proposed architecture can integrate new knowledge seamlessly and achieve significant zero-shot performance when exposed to unseen domains. |
Robust Dialogue State Tracking with Weak Supervision and Sparse Data (2022.tacl-1)
Copied to clipboard
Michael Heck, Nurul Lubis, Carel van Niekerk, Shutong Feng, Christian Geishauser, Hsien-Chin Lin, Milica Gašić
| Challenge: | Generalizing dialogue state tracking (DST) to new data and domains is especially challenging due to the strong reliance on abundant and fine-grained supervision during training. |
| Approach: | They propose a training strategy to build extractive DST models without the need for fine-grained manual span labels. |
| Outcome: | The proposed model improves robustness against sample sparsity, new concepts, and topics, leading to state-of-the-art performance on a range of benchmarks. |
Large-Scale Multi-Domain Belief Tracking with Knowledge Sharing (P18-2)
Copied to clipboard
| Challenge: | Existing approaches to scalability of dialogue belief tracking are dependent on the ontology of the dialogue . current approaches are not scalable to multi-domain dialogues because of the effort required to define a semantic dictionary for each domain. |
| Approach: | They propose a model that utilizes semantic similarity between dialogue utterances and ontology terms to allow information to be shared across domains. |
| Outcome: | The proposed model outperforms state-of-the-art models in multi-domain dialogue tracking tasks while maintaining high quality. |
MultiWOZ - A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modelling (D18-1)
Copied to clipboard
Paweł Budzianowski, Tsung-Hsien Wen, Bo-Hsiang Tseng, Iñigo Casanueva, Stefan Ultes, Osman Ramadan, Milica Gašić
| Challenge: | a dataset of 10k human-human written conversations is one order of magnitude larger than previous annotated task-oriented corpora. |
| Approach: | They propose to collect 10k human-human written conversations from a crowd-sourced dataset using crowd-sourcing. |
| Outcome: | The proposed dataset is one order of magnitude larger than previous annotated task-oriented corpora and shows the usability of the data and sets a baseline for future studies. |
A Confidence-based Acquisition Model for Self-supervised Active Learning and Label Correction (2025.tacl-1)
Copied to clipboard
Carel van Niekerk, Christian Geishauser, Michael Heck, Shutong Feng, Hsien-chin Lin, Nurul Lubis, Benjamin Ruppik, Renato Vukovic, Milica Gašić
| Challenge: | Existing approaches to training deep neural networks require large amounts of meticulously annotated data. |
| Approach: | They propose a pool-based active learning framework that requires expert annotators to label only a fraction of a sequence and facilitates self-supervision for the remainder of the sequence. |
| Outcome: | The proposed model outperforms baselines on dialogue belief tracking tasks. |
Feudal Reinforcement Learning for Dialogue Management in Large Domains (N18-2)
Copied to clipboard
Iñigo Casanueva, Paweł Budzianowski, Pei-Hao Su, Stefan Ultes, Lina M. Rojas-Barahona, Bo-Hsiang Tseng, Milica Gašić
| Challenge: | Reinforcement learning (RL) is a promising approach to model dialogue policy optimisation but fails to scale to large domains due to the curse of dimensionality. |
| Approach: | They propose a novel approach to dialogue policy optimisation using reinforcement learning . they propose to decompose the decision into two steps using a domain ontology . |
| Outcome: | The proposed architecture outperforms state-of-the-art in several dialogue domains without any additional reward signal. |