All Outputs (55)

A Snapshot of NLG Evaluation Practices 2005 - 2014
Presentation / Conference Contribution
Gkatzia, D., & Mahamood, S. (2015, September). A Snapshot of NLG Evaluation Practices 2005 - 2014. Presented at Proceedings of the 15th European Workshop on Natural Language Generation (ENLG), University of Brighton Brighton, UK

In this paper we present a snapshot of endto-end NLG system evaluations as presented in conference and journal papers1 over the last ten years in order to better understand the nature and type of evaluations that have been undertaken. We find that re... Read More about A Snapshot of NLG Evaluation Practices 2005 - 2014.

Generating and Evaluating Landmark-Based Navigation Instructions in Virtual Environments
Presentation / Conference Contribution
Cercas Curry, A., Gkatzia, D., & Rieser, V. (2015, September). Generating and Evaluating Landmark-Based Navigation Instructions in Virtual Environments. Presented at 15th European Workshop on Natural Language Generation (ENLG 2015), University of Brighton, Brighton, UK

Referring to landmarks has been identified to lead to improved navigation instructions. However, a previous corpus study suggests that human “wizards” also choose to refer to street names and generate user-centric instructions. In this paper, we cond... Read More about Generating and Evaluating Landmark-Based Navigation Instructions in Virtual Environments.

Comparing Multi-label Classification with Reinforcement Learning for Summarisation of Time-series Data
Presentation / Conference Contribution
Gkatzia, D., Hastie, H., & Lemon, O. (2014, June). Comparing Multi-label Classification with Reinforcement Learning for Summarisation of Time-series Data. Presented at The 52nd Annual Meeting of the Association for Computational Linguistics, Baltimore

We present a novel approach for automatic report generation from time-series data, in the context of student feedback generation. Our proposed methodology treats content selection as a multi-label (ML)
classification problem, which takes as input ti... Read More about Comparing Multi-label Classification with Reinforcement Learning for Summarisation of Time-series Data.

Finding middle ground? Multi-objective Natural Language Generation from time-series data
Presentation / Conference Contribution
Gkatzia, D., Hastie, H., & Lemon, O. (2014, April). Finding middle ground? Multi-objective Natural Language Generation from time-series data. Presented at Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, volume 2: Short Papers, Gothenburg, Sweden

A Natural Language Generation (NLG) system is able to generate text from nonlinguistic data, ideally personalising the content to a user’s specific needs. In some cases, however, there are multiple stakeholders with their own individual goals, needs... Read More about Finding middle ground? Multi-objective Natural Language Generation from time-series data.

It's Common Sense, isn't it? Demystifying Human Evaluations in Commonsense-enhanced NLG systems
Presentation / Conference Contribution
Mahamood, S., Clinciu, M., & Gkatzia, D. (2021, April). It's Common Sense, isn't it? Demystifying Human Evaluations in Commonsense-enhanced NLG systems. Presented at Workshop on Human Evaluation of NLP Systems (HumEval at EACL 2021), Kyiv, Ukraine (online)

Common sense is an integral part of human cognition which allows us to make sound decisions , communicate effectively with others and interpret situations and utterances. Endowing AI systems with commonsense knowledge capabilities will help us get cl... Read More about It's Common Sense, isn't it? Demystifying Human Evaluations in Commonsense-enhanced NLG systems.

The REAL Corpus: a crowd-sourced corpus of human generated and evaluated spatial references to real-world urban scenes
Presentation / Conference Contribution
Bartie, P., Mackaness, W., Gkatzia, D., & Rieser, V. (2016, May). The REAL Corpus: a crowd-sourced corpus of human generated and evaluated spatial references to real-world urban scenes. Presented at 10th International Conference on Language Resources and Evaluation (LREC)

We present a newly crowd-sourced data set of natural language references to objects anchored in complex urban scenes (In short: The REAL Corpus – Referring Expressions Anchored Language). The REAL corpus contains a collection of images of real-world... Read More about The REAL Corpus: a crowd-sourced corpus of human generated and evaluated spatial references to real-world urban scenes.

Commonsense-enhanced Natural Language Generation for Human-Robot Interaction
Presentation / Conference Contribution
Gkatzia, D. (2020, December). Commonsense-enhanced Natural Language Generation for Human-Robot Interaction. Presented at 2nd Workshop on Natural Language Generation for Human-Robot Interaction (HRI 2020), Online

Commonsense is vital for human communication, as it allows us to make inferences without explicitly mentioning the context. Equipping robots with commonsense knowledge would lead to better communication between humans and robots and will allow robots... Read More about Commonsense-enhanced Natural Language Generation for Human-Robot Interaction.

Inflection Generation for Spanish Verbs using Supervised Learning
Presentation / Conference Contribution
Barros, C., Gkatzia, D., & Lloret, E. (2017, September). Inflection Generation for Spanish Verbs using Supervised Learning. Presented at First Workshop on Subword and Character Level Models in NLP, Copenhagen, Denmark

We present a novel supervised approach to inflection generation for verbs in Spanish. Our system takes as input the verb’s lemma form and the desired features such as person, number, tense, and is able to predict the appropriate grammatical conjugati... Read More about Inflection Generation for Spanish Verbs using Supervised Learning.

Improving the Naturalness and Expressivity of Language Generation for Spanish
Presentation / Conference Contribution
Barros, C., Gkatzia, D., & Lloret, E. (2017, September). Improving the Naturalness and Expressivity of Language Generation for Spanish. Presented at 10th International Conference on Natural Language Generation, Santiago de Compostela, Spain

We present a flexible Natural Language Generation approach for Spanish, focused on the surface realisation stage, which integrates an inflection module in order to improve the naturalness and expressivity of the generated language. This inflection mo... Read More about Improving the Naturalness and Expressivity of Language Generation for Spanish.

Second Workshop on Natural Language Generation for Human-Robot Interaction
Presentation / Conference Contribution
Buschmeier, H., Ellen Foster, M., & Gkatzia, D. (2020, March). Second Workshop on Natural Language Generation for Human-Robot Interaction. Presented at HRI '20: ACM/IEEE International Conference on Human-Robot Interaction, Cambridge

This workshop is the second in a series bringing together the Natural Language Generation and Human-Robot Interaction communities to discuss topics of mutual interest with the goal of developing an HRI-inspired NLG shared task. The workshop website i... Read More about Second Workshop on Natural Language Generation for Human-Robot Interaction.

Proceedings of the Workshop on NLG for Human–Robot Interaction
Presentation / Conference Contribution
(2018, November). Proceedings of the Workshop on NLG for Human–Robot Interaction. Presented at Workshop on NLG for Human–Robot Interaction, Tilburg, The Netherlands

Ellen Foster, M., H. Buschmeier, & D. Gkatzia (Eds.) (2018). Proceedings of the Workshop on NLG for Human–Robot Interaction.

Generating student feedback from time-series data using Reinforcement Learning
Presentation / Conference Contribution
Gkatzia, D., Hastie, H., Janarthanam, S., & Lemon, O. (2013, August). Generating student feedback from time-series data using Reinforcement Learning. Presented at 14th European Workshop On Natural Language Generation, Sofia, Bulgaria

We describe a statistical Natural LanguageGeneration (NLG) method for summarisa-tion of time-series data in the context offeedback generation for students. In thispaper, we initially present a method forcollecting time-series data f... Read More about Generating student feedback from time-series data using Reinforcement Learning.

Learning from limited datasets: Implications for Natural Language Generation and Human-Robot Interaction
Presentation / Conference Contribution
Belakova, J., & Gkatzia, D. (2018, November). Learning from limited datasets: Implications for Natural Language Generation and Human-Robot Interaction. Presented at Workshop on NLG for Human–Robot Interaction, Tilburg, The Netherlands

One of the most natural ways for human robot communication is through spoken language. Training human-robot interaction systems require access to large datasets which are expensive to obtain and labour intensive. In this paper, we des... Read More about Learning from limited datasets: Implications for Natural Language Generation and Human-Robot Interaction.

Underreporting of errors in NLG output, and what to do about it
Presentation / Conference Contribution
van Miltenburg, E., Clinciu, M.-A., Dušek, O., Gkatzia, D., Inglis, S., Leppänen, L., Mahamood, S., Manning, E., Schoch, S., Thomson, C., & Wen, L. (2021, September). Underreporting of errors in NLG output, and what to do about it. Presented at 14th International Conference on Natural Language Generation, Aberdeen, UK

We observe a severe under-reporting of the different kinds of errors that Natural Language Generation systems make. This is a problem, because mistakes are an important indicator of where systems should still be improved. If authors only report overa... Read More about Underreporting of errors in NLG output, and what to do about it.

Chefbot: A Novel Framework for the Generation of Commonsense-enhanced Responses for Task-based Dialogue Systems
Presentation / Conference Contribution
Strathearn, C., & Gkatzia, D. (2021, August). Chefbot: A Novel Framework for the Generation of Commonsense-enhanced Responses for Task-based Dialogue Systems. Presented at 14th International Conference on Natural Language Generation, Aberdeen

Conversational systems aim to generate responses that are accurate, relevant and engaging, either through utilising neural end-to-end models or through slot filling. Human-to-human conversations are enhanced by not only the latest utterance of the in... Read More about Chefbot: A Novel Framework for the Generation of Commonsense-enhanced Responses for Task-based Dialogue Systems.