Papers by Maja Popović
Informative Manual Evaluation of Machine Translation Output (2020.coling-main)
Copied to clipboard
| Challenge: | a new method for manual evaluation of machine translation output is proposed . evaluators mark problematic parts of the translated text, not just overall scores . |
| Approach: | They propose a method for manual evaluation of machine translation output based on marking actual issues in the translated text. |
| Outcome: | The proposed method can be applied on any genre/domain and language pair . it can be guided by various types of quality criteria and can be used for other types of generated text. |
DiHuTra: a Parallel Corpus to Analyse Differences between Human Translations (2022.lrec-1)
Copied to clipboard
| Challenge: | a new corpus of human translations contains both professional and student translations of news and reviews texts. |
| Approach: | They propose to use the data to compare human and professional translations of news and reviews in a new corpus which contains both professional and student translations. |
| Outcome: | The proposed corpus contains professional and student translations of news and reviews and a subcorpus containing reviews into Finnish. |
On Context Span Needed for Machine Translation Evaluation (2020.lrec-1)
Copied to clipboard
| Challenge: | a number of common patterns can be observed for context-aware MT evaluation, authors say . document-level evaluations have largely been performed at the sentence level . the definition of what constitutes a "document level" evaluation is still unclear . |
| Approach: | They propose to use a series of surveys to identify the necessary context span . they find common patterns that can be used to draw general guidelines . |
| Outcome: | The proposed evaluations of machine translation systems show that some issues and spans depend on domain and target language. |
Leveraging Pre-trained Language Models for Gender Debiasing (2022.lrec-1)
Copied to clipboard
| Challenge: | Existing methods to reduce gender bias in natural language are costly and time-consuming. |
| Approach: | They propose a method to generate gender variants for a given text using pre-trained language models as the resource without any task-specific labelled data. |
| Outcome: | The proposed method can reduce gender bias in a language generation context without a task-specific labelled data. |
Exploring Variation of Results from Different Experimental Conditions (2023.findings-acl)
Copied to clipboard
| Challenge: | Recent research has shown that reproducibility of NLP experiments is not guaranteed by arbitrary factors like random seed and different data splits. |
| Approach: | They propose to run the training and testing pipeline for three neural text simplification models under varying experimental conditions using the same data and train/dev/test set splits. |
| Outcome: | The proposed pipeline yields a large number of results for three neural text simplification models under varying experimental conditions. |