LLM-powered data augmentation for enhanced cross-lingual performance
This paper explores the potential of leveraging Large Language Models (LLMs) for data
augmentation in multilingual commonsense reasoning datasets where the available training …
augmentation in multilingual commonsense reasoning datasets where the available training …
Beyond Static models and test sets: Benchmarking the potential of pre-trained models across tasks and languages
Although recent Massively Multilingual Language Models (MMLMs) like mBERT and XLMR
support around 100 languages, most existing multilingual NLP benchmarks provide …
support around 100 languages, most existing multilingual NLP benchmarks provide …
DeMuX: Data-efficient Multilingual Learning
We consider the task of optimally fine-tuning pre-trained multilingual models, given small
amounts of unlabelled target data and an annotation budget. In this paper, we introduce …
amounts of unlabelled target data and an annotation budget. In this paper, we introduce …
[PDF][PDF] To train or not to train: Predicting the performance of massively multilingual models
Evaluating the performance of Massively Multilingual Language Models (MMLMs) is difficult
due to the shortage of evaluation datasets in low-resource languages. Due to computational …
due to the shortage of evaluation datasets in low-resource languages. Due to computational …
Performance Prediction via Bayesian Matrix Factorisation for Multilingual Natural Language Processing Tasks
Abstract Performance prediction for Natural Language Processing (NLP) seeks to reduce
the experimental burden resulting from the myriad of different evaluation scenarios, eg, the …
the experimental burden resulting from the myriad of different evaluation scenarios, eg, the …
[PDF][PDF] The SUMEval 2022 shared task on performance prediction of multilingual pre-trained language models
Abstract The SUMEval Workshop's shared task involved predicting performance of
multilingual PLMs across multiple languages when these models are fine-tuned with varying …
multilingual PLMs across multiple languages when these models are fine-tuned with varying …
A Proposal for Scaling the Scaling Laws
W Schellaert, R Hamon… - Proceedings of the …, 2024 - aclanthology.org
Scaling laws are predictable relations between the performance of AI systems and various
scalable design choices such as model or dataset size. In order to keep predictions …
scalable design choices such as model or dataset size. In order to keep predictions …
[PDF][PDF] The GMU System Submission for the SUMEval 2022 Shared Task
SS Akter, A Anastasopoulos - … of the First Workshop on Scaling …, 2022 - aclanthology.org
This paper describes the submission of our multilingual NLP model performance evaluation
system for the SUMEval 2022 shared task, a system for predict the performance of a model …
system for the SUMEval 2022 shared task, a system for predict the performance of a model …
[PDF][PDF] Zero-shot Cross-lingual Transfer Learning with Multiple Source and Target Languages for Information Extraction: Language Selection and Adversarial Training
NT Ngo, TH Nguyen - cs.uoregon.edu
The majority of previous researches addressing multi-lingual IE are limited to zero-shot
cross-lingual single-transfer (one-to-one) setting, with high-resource languages …
cross-lingual single-transfer (one-to-one) setting, with high-resource languages …