Surangika Ranathunga

En-Shiun Annie Lee, Sarubi Thillainathan, Shravan Nayak et al.

What can pre-trained multilingual sequence-to-sequence models like mBART contribute to translating low-resource languages? We conduct a thorough empirical experiment in 10 languages to ascertain this, considering five factors: (1) the amount of fine-tuning data, (2) the noise in the fine-tuning data, (3) the amount of pre-training data in the model, (4) the impact of domain mismatch, and (5) language typology. In addition to yielding several heuristics, the experiments form a framework for evaluating the data sensitivities of machine translation systems. While mBART is robust to domain differences, its translations for unseen and typologically distant languages remain below 3.0 BLEU. In answer to our title's question, mBART is not a low-resource panacea; we therefore encourage shifting the emphasis from new models to new data.

Shravan Nayak, Surangika Ranathunga, Sarubi Thillainathan et al. · utoronto

NMT systems trained on Pre-trained Multilingual Sequence-Sequence (PMSS) models flounder when sufficient amounts of parallel data is not available for fine-tuning. This specifically holds for languages missing/under-represented in these models. The problem gets aggravated when the data comes from different domains. In this paper, we show that intermediate-task fine-tuning (ITFT) of PMSS models is extremely beneficial for domain-specific NMT, especially when target domain data is limited/unavailable and the considered languages are missing or under-represented in the PMSS model. We quantify the domain-specific results variations using a domain-divergence test, and show that ITFT can mitigate the impact of domain divergence to some extent.

Surangika Ranathunga, Nisansa de Silva

Linguistic disparity in the NLP world is a problem that has been widely acknowledged recently. However, different facets of this problem, or the reasons behind this disparity are seldom discussed within the NLP community. This paper provides a comprehensive analysis of the disparity that exists within the languages of the world. We show that simply categorising languages considering data availability may not be always correct. Using an existing language categorisation based on speaker population and vitality, we analyse the distribution of language data resources, amount of NLP/CL research, inclusion in multilingual web-based platforms and the inclusion in pre-trained multilingual models. We show that many languages do not get covered in these resources or platforms, and even within the languages belonging to the same language group, there is wide disparity. We analyse the impact of family, geographical location, GDP and the speaker population of languages and provide possible reasons for this disparity, along with some suggestions to overcome the same.

Vinura Dhananjaya, Piyumal Demotte, Surangika Ranathunga et al.

This research provides the first comprehensive analysis of the performance of pre-trained language models for Sinhala text classification. We test on a set of different Sinhala text classification tasks and our analysis shows that out of the pre-trained multilingual models that include Sinhala (XLM-R, LaBSE, and LASER), XLM-R is the best model by far for Sinhala text classification. We also pre-train two RoBERTa-based monolingual Sinhala models, which are far superior to the existing pre-trained language models for Sinhala. We show that when fine-tuned, these pre-trained language models set a very strong baseline for Sinhala text classification and are robust in situations where labeled data is insufficient for fine-tuning. We further provide a set of recommendations for using pre-trained models for Sinhala text classification. We also introduce new annotated datasets useful for future research in Sinhala text classification and publicly release our pre-trained models.

Aloka Fernando, Surangika Ranathunga

Out-of-Vocabulary (OOV) is a problem for Neural Machine Translation (NMT). OOV refers to words with a low occurrence in the training data, or to those that are absent from the training data. To alleviate this, word or phrase-based Data Augmentation (DA) techniques have been used. However, existing DA techniques have addressed only one of these OOV types and limit to considering either syntactic constraints or semantic constraints. We present a word and phrase replacement-based DA technique that consider both types of OOV, by augmenting (1) rare words in the existing parallel corpus, and (2) new words from a bilingual dictionary. During augmentation, we consider both syntactic and semantic properties of the words to guarantee fluency in the synthetic sentences. This technique was experimented with low resource Sinhala-English language pair. We observe with only semantic constraints in the DA, the results are comparable with the scores obtained considering syntactic constraints, and is favourable for low-resourced languages that lacks linguistic tool support. Additionally, results can be further improved by considering both syntactic and semantic constraints.

Jittarin Jetwiriyanon, Teo Susnjak, Surangika Ranathunga

Many universities face increasing financial pressure and rely on accurate forecasts of commencing enrolments. However, enrolment forecasting in higher education is often data-sparse; annual series are short and affected by reporting changes and regime shifts. Popular classical approaches can be unreliable, as parameter estimation and model selection are unstable with short samples, and structural breaks degrade extrapolation. Recently, TSFMs have provided zero-shot priors, delivering strong gains in annual, data-sparse institutional forecasting under leakage-disciplined covariate construction. We benchmark multiple TSFM families in a zero-shot setting and test a compact, leakage-safe covariate set and introduce the Institutional Operating Conditions Index (IOCI), a transferable 0-100 regime covariate derived from time-stamped documentary evidence available at each forecast origin, alongside Google Trends demand proxies with stabilising feature engineering. Using an expanding-window backtest with strict vintage alignment, covariate-conditioned TSFMs perform on par with classical benchmarks without institution-specific training, with performance differences varying by cohort and model.

Teo Susnjak, Peter Hwang, Napoleon H. Reyes et al.

This research pioneers the use of fine-tuned Large Language Models (LLMs) to automate Systematic Literature Reviews (SLRs), presenting a significant and novel contribution in integrating AI to enhance academic research methodologies. Our study employed the latest fine-tuning methodologies together with open-sourced LLMs, and demonstrated a practical and efficient approach to automating the final execution stages of an SLR process that involves knowledge synthesis. The results maintained high fidelity in factual accuracy in LLM responses, and were validated through the replication of an existing PRISMA-conforming SLR. Our research proposed solutions for mitigating LLM hallucination and proposed mechanisms for tracking LLM responses to their sources of information, thus demonstrating how this approach can meet the rigorous demands of scholarly research. The findings ultimately confirmed the potential of fine-tuned LLMs in streamlining various labor-intensive processes of conducting literature reviews. Given the potential of this approach and its applicability across all research domains, this foundational study also advocated for updating PRISMA reporting guidelines to incorporate AI-driven processes, ensuring methodological transparency and reliability in future SLRs. This study broadens the appeal of AI-enhanced tools across various academic and research fields, setting a new standard for conducting comprehensive and accurate literature reviews with more efficiency in the face of ever-increasing volumes of academic studies.

Nimesh Ariyarathne, Harshani Bandara, Yasith Heshan et al.

Mathematics is often perceived as a complex subject by students, leading to high failure rates in exams. To improve Mathematics skills, it is important to provide sample questions for students to practice problem-solving. Manually creating Math Word Problems (MWPs) is time consuming for tutors, because they have to type in natural language while adhering to grammar and spelling rules of the language. Early techniques that use pre-trained Language Models for MWP generation either require a tutor to provide the initial portion of the MWP, and/or additional information such as an equation. In this paper, we present an MWP generation system (MathWiz) based on Large Language Models (LLMs) that overcomes the need for additional input - the only input to our system is the number of MWPs needed, the grade and the type of question (e.g.~addition, subtraction). Unlike the existing LLM-based solutions for MWP generation, we carried out an extensive set of experiments involving different LLMs, prompting strategies, techniques to improve the diversity of MWPs, as well as techniques that employ human feedback to improve LLM performance. Human and automated evaluations confirmed that the generated MWPs are high in quality, with minimal spelling and grammar issues. However, LLMs still struggle to generate questions that adhere to the specified grade and question type requirements.

Hannah Liu, Muxin Tian, Iqra Ali et al.

Sentence simplification aims to make complex text more accessible by reducing linguistic complexity while preserving the original meaning. However, progress in this area remains limited for mid-resource and low-resource languages due to the scarcity of high-quality data. To address this gap, we introduce the OasisSimp dataset, a multilingual dataset for sentence-level simplification covering five languages: English, Sinhala, Tamil, Pashto, and Thai. Among these, no prior sentence simplification datasets exist for Thai, Pashto, and Tamil, while limited data is available for Sinhala. Each language simplification dataset was created by trained annotators who followed detailed guidelines to simplify sentences while maintaining meaning, fluency, and grammatical correctness. We evaluate eight open-weight multilingual Large Language Models (LLMs) on the OasisSimp dataset and observe substantial performance disparities between high-resource and low-resource languages, highlighting the simplification challenges in multilingual settings. The OasisSimp dataset thus provides both a valuable multilingual resource and a challenging benchmark, revealing the limitations of current LLM-based simplification methods and paving the way for future research in low-resource sentence simplification. The dataset is available at https://OasisSimpDataset.github.io/.

Aneesha Fernando, Surangika Ranathunga, Kristin Stock et al.

Georeferencing text documents has typically relied on either gazetteer-based methods to assign geographic coordinates to place names, or on language modelling approaches that associate textual terms with geographic locations. However, many location descriptions specify positions relatively with spatial relationships, making geocoding based solely on place names or geo-indicative words inaccurate. This issue frequently arises in biological specimen collection records, where locations are often described through narratives rather than coordinates if they pre-date GPS. Accurate georeferencing is vital for biodiversity studies, yet the process remains labour-intensive, leading to a demand for automated georeferencing solutions. This paper explores the potential of Large Language Models (LLMs) to georeference complex locality descriptions automatically, focusing on the biodiversity collections domain. We first identified effective prompting patterns, then fine-tuned an LLM using Quantized Low-Rank Adaptation (QLoRA) on biodiversity datasets from multiple regions and languages. Our approach outperforms existing baselines with an average, across datasets, of 65% of records within a 10 km radius, for a fixed amount of training data. The best results (New York state) were 85% within 10km and 67% within 1km. The selected LLM performs well for lengthy, complex descriptions, highlighting its potential for georeferencing intricate locality descriptions.

Nisansa de Silva, Surangika Ranathunga

This paper presents the first-ever Sinhala physical common sense reasoning dataset created as part of Global PIQA. It contains 110 human-created and verified data samples, where each sample consists of a prompt, the corresponding correct answer, and a wrong answer. Most of the questions refer to the Sri Lankan context, where Sinhala is an official language.

Yomal De Mel, Kasun Wickramasinghe, Nisansa de Silva et al.

Due to reasons of convenience and lack of tech literacy, transliteration (i.e., Romanizing native scripts instead of using localization tools) is eminently prevalent in the context of low-resource languages such as Sinhala, which have their own writing script. In this study, our focus is on Romanized Sinhala transliteration. We propose two methods to address this problem: Our baseline is a rule-based method, which is then compared against our second method where we approach the transliteration problem as a sequence-to-sequence task akin to the established Neural Machine Translation (NMT) task. For the latter, we propose a Transformer-based Encode-Decoder solution. We witnessed that the Transformer-based method could grab many ad-hoc patterns within the Romanized scripts compared to the rule-based method. The code base associated with this paper is available on GitHub - https://github.com/kasunw22/Sinhala-Transliterator/

Surangika Ranathunga, Asanka Ranasinghea, Janaka Shamala et al.

This paper presents a multi-way parallel English-Tamil-Sinhala corpus annotated with Named Entities (NEs), where Sinhala and Tamil are low-resource languages. Using pre-trained multilingual Language Models (mLMs), we establish new benchmark Named Entity Recognition (NER) results on this dataset for Sinhala and Tamil. We also carry out a detailed investigation on the NER capabilities of different types of mLMs. Finally, we demonstrate the utility of our NER system on a low-resource Neural Machine Translation (NMT) task. Our dataset is publicly released: https://github.com/suralk/multiNER.

Surangika Ranathunga, Rumesh Sirithunga, Himashi Rathnayake et al.

Text Simplification is a task that has been minimally explored for low-resource languages. Consequently, there are only a few manually curated datasets. In this paper, we present a human curated sentence-level text simplification dataset for the Sinhala language. Our evaluation dataset contains 1,000 complex sentences and corresponding 3,000 simplified sentences produced by three different human annotators. We model the text simplification task as a zero-shot and zero resource sequence-to-sequence (seq-seq) task on the multilingual language models mT5 and mBART. We exploit auxiliary data from related seq-seq tasks and explore the possibility of using intermediate task transfer learning (ITTL). Our analysis shows that ITTL outperforms the previously proposed zero-resource methods for text simplification. Our findings also highlight the challenges in evaluating text simplification systems, and support the calls for improved metrics for measuring the quality of automated text simplification systems that would suit low-resource languages as well. Our code and data are publicly available: https://github.com/brainsharks-fyp17/Sinhala-Text-Simplification-Dataset-and-Evaluation

H. W. K. Aravinda, Rashad Sirajudeen, Samith Karunathilake et al.

Low-resource languages such as Sinhala are often overlooked by open-source Large Language Models (LLMs). In this research, we extend an existing multilingual LLM (Llama-3-8B) to better serve Sinhala. We enhance the LLM tokenizer with Sinhala specific vocabulary and perform continual pre-training on a cleaned 10 million Sinhala corpus, resulting in the SinLlama model. This is the very first decoder-based open-source LLM with explicit Sinhala support. When SinLlama was instruction fine-tuned for three text classification tasks, it outperformed base and instruct variants of Llama-3-8B by a significant margin.

Surangika Ranathunga, Nisansa de Silva, Menan Velayuthan et al.

We conducted a detailed analysis on the quality of web-mined corpora for two low-resource languages (making three language pairs, English-Sinhala, English-Tamil and Sinhala-Tamil). We ranked each corpus according to a similarity measure and carried out an intrinsic and extrinsic evaluation on different portions of this ranked corpus. We show that there are significant quality differences between different portions of web-mined corpora and that the quality varies across languages and datasets. We also show that, for some web-mined datasets, Neural Machine Translation (NMT) models trained with their highest-ranked 25k portion can be on par with human-curated datasets.

Robert Spencer, Surangika Ranathunga, Mikael Boulic et al.

This study investigates the application of Transfer Learning (TL) on Transformer architectures to enhance building energy consumption forecasting. Transformers are a relatively new deep learning architecture, which has served as the foundation for groundbreaking technologies such as ChatGPT. While TL has been studied in the past, prior studies considered either one data-centric TL strategy or used older deep learning models such as Recurrent Neural Networks or Convolutional Neural Networks. Here, we carry out an extensive empirical study on six different data-centric TL strategies and analyse their performance under varying feature spaces. In addition to the vanilla Transformer architecture, we also experiment with Informer and PatchTST, specifically designed for time series forecasting. We use 16 datasets from the Building Data Genome Project 2 to create building energy consumption forecasting models. Experimental results reveal that while TL is generally beneficial, especially when the target domain has no data, careful selection of the exact TL strategy should be made to gain the maximum benefit. This decision largely depends on the feature space properties such as the recorded weather features. We also note that PatchTST outperforms the other two Transformer variants (vanilla Transformer and Informer). Our findings advance the building energy consumption forecasting using advanced approaches like TL and Transformer architectures.

Bastin Tony Roy Savarimuthu, Surangika Ranathunga, Stephen Cranefield

Software agents, both human and computational, do not exist in isolation and often need to collaborate or coordinate with others to achieve their goals. In human society, social mechanisms such as norms ensure efficient functioning, and these techniques have been adopted by researchers in multi-agent systems (MAS) to create socially aware agents. However, traditional techniques have limitations, such as operating in limited environments often using brittle symbolic reasoning. The advent of Large Language Models (LLMs) offers a promising solution, providing a rich and expressive vocabulary for norms and enabling norm-capable agents that can perform a range of tasks such as norm discovery, normative reasoning and decision-making. This paper examines the potential of LLM-based agents to acquire normative capabilities, drawing on recent Natural Language Processing (NLP) and LLM research. We present our vision for creating normative LLM agents. In particular, we discuss how the recently proposed "LLM agent" approaches can be extended to implement such normative LLM agents. We also highlight challenges in this emerging field. This paper thus aims to foster collaboration between MAS, NLP and LLM researchers in order to advance the field of normative agents.

Tong Su, Xin Peng, Sarubi Thillainathan et al.

Parameter-efficient fine-tuning (PEFT) methods are increasingly vital in adapting large-scale pre-trained language models for diverse tasks, offering a balance between adaptability and computational efficiency. They are important in Low-Resource Language (LRL) Neural Machine Translation (NMT) to enhance translation accuracy with minimal resources. However, their practical effectiveness varies significantly across different languages. We conducted comprehensive empirical experiments with varying LRL domains and sizes to evaluate the performance of 8 PEFT methods with in total of 15 architectures using the SacreBLEU score. We showed that 6 PEFT architectures outperform the baseline for both in-domain and out-domain tests and the Houlsby+Inversion adapter has the best performance overall, proving the effectiveness of PEFT methods.

Aloka Fernando, Surangika Ranathunga

Multilingual Pre-trained Language models (multiPLMs), trained on the Masked Language Modelling (MLM) objective are commonly being used for cross-lingual tasks such as bitext mining. However, the performance of these models is still suboptimal for low-resource languages (LRLs). To improve the language representation of a given multiPLM, it is possible to further pre-train it. This is known as continual pre-training. Previous research has shown that continual pre-training with MLM and subsequently with Translation Language Modelling (TLM) improves the cross-lingual representation of multiPLMs. However, during masking, both MLM and TLM give equal weight to all tokens in the input sequence, irrespective of the linguistic properties of the tokens. In this paper, we introduce a novel masking strategy, Linguistic Entity Masking (LEM) to be used in the continual pre-training step to further improve the cross-lingual representations of existing multiPLMs. In contrast to MLM and TLM, LEM limits masking to the linguistic entity types nouns, verbs and named entities, which hold a higher prominence in a sentence. Secondly, we limit masking to a single token within the linguistic entity span thus keeping more context, whereas, in MLM and TLM, tokens are masked randomly. We evaluate the effectiveness of LEM using three downstream tasks, namely bitext mining, parallel data curation and code-mixed sentiment analysis using three low-resource language pairs English-Sinhala, English-Tamil, and Sinhala-Tamil. Experiment results show that continually pre-training a multiPLM with LEM outperforms a multiPLM continually pre-trained with MLM+TLM for all three tasks.

Sarubi Thillainathan, Songchen Yuan, En-Shiun Annie Lee et al.

Fine-tuning multilingual sequence-to-sequence large language models (msLLMs) has shown promise in developing neural machine translation (NMT) systems for low-resource languages (LRLs). However, conventional single-stage fine-tuning methods struggle in extremely low-resource NMT settings, where training data is very limited. This paper contributes to artificial intelligence by proposing two approaches for adapting msLLMs in these challenging scenarios: (1) continual pre-training (CPT), where the msLLM is further trained with domain-specific monolingual data to compensate for the under-representation of LRLs, and (2) intermediate task transfer learning (ITTL), a method that fine-tunes the msLLM with both in-domain and out-of-domain parallel data to enhance its translation capabilities across various domains and tasks. As an application in engineering, these methods are implemented in NMT systems for Sinhala, Tamil, and English (six language pairs) in domain-specific, extremely low-resource settings (datasets containing fewer than 100,000 samples). Our experiments reveal that these approaches enhance translation performance by an average of +1.47 bilingual evaluation understudy (BLEU) score compared to the standard single-stage fine-tuning baseline across all translation directions. Additionally, a multi-model ensemble further improves performance by an additional BLEU score.

Aloka Fernando, Nisansa de Silva, Menan Velyuthan et al.

Parallel Data Curation (PDC) techniques aim to filter out noisy parallel sentences from web-mined corpora. Ranking sentence pairs using similarity scores on sentence embeddings derived from Pre-trained Multilingual Language Models (multiPLMs) is the most common PDC technique. However, previous research has shown that the choice of the multiPLM significantly impacts the quality of the filtered parallel corpus, and the Neural Machine Translation (NMT) models trained using such data show a disparity across multiPLMs. This paper shows that this disparity is due to different multiPLMs being biased towards certain types of sentence pairs, which are treated as noise from an NMT point of view. We show that such noisy parallel sentences can be removed to a certain extent by employing a series of heuristics. The NMT models, trained using the curated corpus, lead to producing better results while minimizing the disparity across multiPLMs. We publicly release the source code and the curated datasets.

Thevin Senath, Kumuthu Athukorala, Ransika Costa et al.

In this paper, we address the challenge of recipe personalization through ingredient substitution. We make use of Large Language Models (LLMs) to build an ingredient substitution system designed to predict plausible substitute ingredients within a given recipe context. Given that the use of LLMs for this task has been barely done, we carry out an extensive set of experiments to determine the best LLM, prompt, and the fine-tuning setups. We further experiment with methods such as multi-task learning, two-stage fine-tuning, and Direct Preference Optimization (DPO). The experiments are conducted using the publicly available Recipe1MSub corpus. The best results are produced by the Mistral7-Base LLM after fine-tuning and DPO. This result outperforms the strong baseline available for the same corpus with a Hit@1 score of 22.04. Thus we believe that this research represents a significant step towards enabling personalized and creative culinary experiences by utilizing LLM-based ingredient substitution.

Akesh Gunathilake, Nadil Karunarathna, Tharusha Bandaranayake et al.

Spell correction is still a challenging problem for low-resource languages (LRLs). While pretrained language models (PLMs) have been employed for spell correction, their use is still limited to a handful of languages, and there has been no proper comparison across PLMs. We present the first empirical study on the effectiveness of PLMs for spell correction, which includes LRLs. We find that Large Language Models (LLMs) outperform their counterparts (encoder-based and encoder-decoder) when the fine-tuning dataset is large. This observation holds even in languages for which the LLM is not pre-trained. We release LMSpell, an easy- to use spell correction toolkit across PLMs. It includes an evaluation function that compensates for the hallucination of LLMs. Further, we present a case study with Sinhala to shed light on the plight of spell correction for LRLs.

Sameeah Noreen Hameed, Surangika Ranathunga, Raj Prasanna et al.

Large-scale disasters can often result in catastrophic consequences on people and infrastructure. Situation awareness about such disaster impacts generated by authoritative data from in-situ sensors, remote sensing imagery, and/or geographic data is often limited due to atmospheric opacity, satellite revisits, and time limitations. This often results in geo-temporal information gaps. In contrast, impact-related social media posts can act as "geo-sensors" during a disaster, where people describe specific impacts and locations. However, not all locations mentioned in disaster-related social media posts relate to an impact. Only the impacted locations are critical for directing resources effectively. e.g., "The death toll from a fire which ripped through the Greek coastal town of #Mati stood at 80, with dozens of people unaccounted for as forensic experts tried to identify victims who were burned alive #Greecefires #AthensFires #Athens #Greece." contains impacted location "Mati" and non-impacted locations "Greece" and "Athens". This research uses Large Language Models (LLMs) to identify all locations, impacts and impacted locations mentioned in disaster-related social media posts. In the process, LLMs are fine-tuned to identify only impacts and impacted locations (as distinct from other, non-impacted locations), including locations mentioned in informal expressions, abbreviations, and short forms. Our fine-tuned model demonstrates efficacy, achieving an F1-score of 0.69 for impact and 0.74 for impacted location extraction, substantially outperforming the pre-trained baseline. These robust results confirm the potential of fine-tuned language models to offer a scalable solution for timely decision-making in resource allocation, situational awareness, and post-disaster recovery planning for responders.

Tyler A. Chang, Catherine Arnett, Abdelrahman Eldesokey et al. · uw

To date, there exist almost no culturally-specific evaluation benchmarks for large language models (LLMs) that cover a large number of languages and cultures. In this paper, we present Global PIQA, a participatory commonsense reasoning benchmark for over 100 languages, constructed by hand by 335 researchers from 65 countries around the world. The 116 language varieties in Global PIQA cover five continents, 14 language families, and 23 writing systems. In the non-parallel split of Global PIQA, over 50% of examples reference local foods, customs, traditions, or other culturally-specific elements. We find that state-of-the-art LLMs perform well on Global PIQA in aggregate, but they exhibit weaker performance in lower-resource languages (up to a 37% accuracy gap, despite random chance at 50%). Open models generally perform worse than proprietary models. Global PIQA highlights that in many languages and cultures, everyday knowledge remains an area for improvement, alongside more widely-discussed capabilities such as complex reasoning and expert knowledge. Beyond its uses for LLM evaluation, we hope that Global PIQA provides a glimpse into the wide diversity of cultures in which human language is embedded.

Imalsha Puranegedara, Themira Chathumina, Nisal Ranathunga et al.

Large Language Models (LLMs) excel in English, but their performance degrades significantly on low-resource languages (LRLs) due to English-centric training. While methods like LangBridge align LLMs with multilingual encoders such as the Massively Multilingual Text-to-Text Transfer Transformer (mT5), they typically use only the final encoder layer. We propose a novel architecture that fuses all intermediate layers, enriching the linguistic information passed to the LLM. Our approach features two strategies: (1) a Global Softmax weighting for overall layer importance, and (2) a Transformer Softmax model that learns token-specific weights. The fused representations are mapped into the LLM's embedding space, enabling it to process multilingual inputs. The model is trained only on English data, without using any parallel or multilingual data. Evaluated on XNLI, IndicXNLI, Sinhala News Classification, and Amazon Reviews, our Transformer Softmax model significantly outperforms the LangBridge baseline. We observe strong performance gains in LRLs, improving Sinhala classification accuracy from 71.66% to 75.86% and achieving clear improvements across Indic languages such as Tamil, Bengali, and Malayalam. These specific gains contribute to an overall boost in average XNLI accuracy from 70.36% to 71.50%. This approach offers a scalable, data-efficient path toward more capable and equitable multilingual LLMs.

Rishemjit Kaur, Arshdeep Singh Bhankhar, Jashanpreet Singh Salh et al.

Enabling farmers to access accurate agriculture-related information in their native languages in a timely manner is crucial for the success of the agriculture field. Publicly available general-purpose Large Language Models (LLMs) typically offer generic agriculture advisories, lacking precision in local and multilingual contexts. Our study addresses this limitation by generating multilingual (English, Hindi, Punjabi) synthetic datasets from agriculture-specific documents from India and fine-tuning LLMs for the task of question answering (QA). Evaluation on human-created datasets demonstrates significant improvements in factuality, relevance, and agricultural consensus for the fine-tuned LLMs compared to the baseline counterparts.

Jittarin Jetwiriyanon, Teo Susnjak, Surangika Ranathunga

This study investigates zero-shot forecasting capabilities of Time Series Foundation Models (TSFMs) for macroeconomic indicators. We apply TSFMs to forecasting economic indicators under univariate conditions, bypassing the need for train bespoke econometric models using and extensive training datasets. Our experiments were conducted on a case study dataset, without additional customisation. We rigorously back-tested three state-of-the-art TSFMs (Chronos, TimeGPT and Moirai) under data-scarce conditions and structural breaks. Our results demonstrate that appropriately engineered TSFMs can internalise rich economic dynamics, accommodate regime shifts, and deliver well-behaved uncertainty estimates out of the box, while matching state-of-the-art multivariate models on this domain. Our findings suggest that, without any fine-tuning, TSFMs can match or exceed classical models during stable economic conditions. However, they are vulnerable to degradation in performances during periods of rapid shocks. The findings offer guidance to practitioners on when zero-shot deployments are viable for macroeconomic monitoring and strategic planning.

Likai Tang, Niruth Bogahawatta, Yasod Ginige et al.

Large Language Models (LLMs) are acquiring a wider range of capabilities, including understanding and responding in multiple languages. While they undergo safety training to prevent them from answering illegal questions, imbalances in training data and human evaluation resources can make these models more susceptible to attacks in low-resource languages (LRL). This paper proposes a framework to automatically assess the multilingual vulnerabilities of commonly used LLMs. Using our framework, we evaluated six LLMs across eight languages representing varying levels of resource availability. We validated the assessments generated by our automated framework through human evaluation in two languages, demonstrating that the framework's results align with human judgments in most cases. Our findings reveal vulnerabilities in LRL; however, these may pose minimal risk as they often stem from the model's poor performance, resulting in incoherent responses.

Charitha Rathnayake, P. R. S. Thilakarathna, Uthpala Nethmini et al.

Bilingual lexicons play a crucial role in various Natural Language Processing tasks. However, many low-resource languages (LRLs) do not have such lexicons, and due to the same reason, cannot benefit from the supervised Bilingual Lexicon Induction (BLI) techniques. To address this, unsupervised BLI (UBLI) techniques were introduced. A prominent technique in this line is structure-based UBLI. It is an iterative method, where a seed lexicon, which is initially learned from monolingual embeddings is iteratively improved. There have been numerous improvements to this core idea, however they have been experimented with independently of each other. In this paper, we investigate whether using these techniques simultaneously would lead to equal gains. We use the unsupervised version of VecMap, a commonly used structure-based UBLI framework, and carry out a comprehensive set of experiments using the LRL pairs, English-Sinhala, English-Tamil, and English-Punjabi. These experiments helped us to identify the best combination of the extensions. We also release bilingual dictionaries for English-Sinhala and English-Punjabi.

Jiangshu Du, Yibo Wang, Wenting Zhao et al.

This work is motivated by two key trends. On one hand, large language models (LLMs) have shown remarkable versatility in various generative tasks such as writing, drawing, and question answering, significantly reducing the time required for many routine tasks. On the other hand, researchers, whose work is not only time-consuming but also highly expertise-demanding, face increasing challenges as they have to spend more time reading, writing, and reviewing papers. This raises the question: how can LLMs potentially assist researchers in alleviating their heavy workload? This study focuses on the topic of LLMs assist NLP Researchers, particularly examining the effectiveness of LLM in assisting paper (meta-)reviewing and its recognizability. To address this, we constructed the ReviewCritique dataset, which includes two types of information: (i) NLP papers (initial submissions rather than camera-ready) with both human-written and LLM-generated reviews, and (ii) each review comes with "deficiency" labels and corresponding explanations for individual segments, annotated by experts. Using ReviewCritique, this study explores two threads of research questions: (i) "LLMs as Reviewers", how do reviews generated by LLMs compare with those written by humans in terms of quality and distinguishability? (ii) "LLMs as Metareviewers", how effectively can LLMs identify potential issues, such as Deficient or unprofessional review segments, within individual paper reviews? To our knowledge, this is the first work to provide such a comprehensive analysis.

Surangika Ranathunga, Nisansa de Silva, Dilith Jayakody et al.

We analysed a sample of NLP research papers archived in ACL Anthology as an attempt to quantify the degree of openness and the benefit of such an open culture in the NLP community. We observe that papers published in different NLP venues show different patterns related to artefact reuse. We also note that more than 30% of the papers we analysed do not release their artefacts publicly, despite promising to do so. Further, we observe a wide language-wise disparity in publicly available NLP-related artefacts.

Isuru Boyagane, Oshadha Katulanda, Surangika Ranathunga et al.

Computer system log data is commonly used in system monitoring, performance characteristic investigation, workflow modeling and anomaly detection. Log data is inherently unstructured or semi-structured, which makes it harder to understand the event flow or other important information of a system by reading raw logs. The process of structuring log files first identifies the log message groups based on the system events that triggered them, and extracts an event template to represent the log messages of each event. This paper introduces a novel method to extract event templates from raw system log files, by using the vector space model commonly used in the field of Information Retrieval to vectorize log data and group log messages into event templates based on their vector similarity. Template extraction process is further enhanced with the use of character and length based filters. When evaluated on publicly available real-world log data benchmarks, this proposed method outperforms all the available state-of-the-art systems in terms of accuracy and robustness.

Piyumal Demotte, Surangika Ranathunga

Text classification systems based on contextual embeddings are not viable options for many of the low resource languages. On the other hand, recently introduced capsule networks have shown performance in par with these text classification models. Thus, they could be considered as a viable alternative for text classification for languages that do not have pre-trained contextual embedding models. However, current capsule networks depend upon spatial patterns without considering the sequential features of the text. They are also sub-optimal in capturing the context-level information in longer sequences. This paper presents a novel Dual-State Capsule (DS-Caps) network-based technique for text classification, which is optimized to mitigate these issues. Two varieties of states, namely sentence-level and word-level, are integrated with capsule layers to capture deeper context-level information for language modeling. The dynamic routing process among capsules was also optimized using the context-level information obtained through sentence-level states. The DS-Caps networks outperform the existing capsule network architectures for multiple datasets, particularly for tasks with longer sequences of text. We also demonstrate the superiority of DS-Caps in text classification for a low resource language.

Charith Rajitha, Lakmali Piyarathne, Dilan Sachintha et al.

Document alignment techniques based on multilingual sentence representations have recently shown state of the art results. However, these techniques rely on unsupervised distance measurement techniques, which cannot be fined-tuned to the task at hand. In this paper, instead of these unsupervised distance measurement techniques, we employ Metric Learning to derive task-specific distance measurements. These measurements are supervised, meaning that the distance measurement metric is trained using a parallel dataset. Using a dataset belonging to English, Sinhala, and Tamil, which belong to three different language families, we show that these task-specific supervised distance learning metrics outperform their unsupervised counterparts, for document alignment.

Surangika Ranathunga, En-Shiun Annie Lee, Marjana Prifti Skenduli et al.

Neural Machine Translation (NMT) has seen a tremendous spurt of growth in less than ten years, and has already entered a mature phase. While considered as the most widely used solution for Machine Translation, its performance on low-resource language pairs still remains sub-optimal compared to the high-resource counterparts, due to the unavailability of large parallel corpora. Therefore, the implementation of NMT techniques for low-resource language pairs has been receiving the spotlight in the recent NMT research arena, thus leading to a substantial amount of research reported on this topic. This paper presents a detailed survey of research advancements in low-resource language NMT (LRL-NMT), along with a quantitative analysis aimed at identifying the most popular solutions. Based on our findings from reviewing previous work, this survey paper provides a set of guidelines to select the possible NMT technique for a given LRL data setting. It also presents a holistic view of the LRL-NMT research landscape and provides a list of recommendations to further enhance the research efforts on LRL-NMT.

Dilan Sachintha, Lakmali Piyarathna, Charith Rajitha et al.

Multilingual sentence representations pose a great advantage for low-resource languages that do not have enough data to build monolingual models on their own. These multilingual sentence representations have been separately exploited by few research for document and sentence alignment. However, most of the low-resource languages are under-represented in these pre-trained models. Thus, in the context of low-resource languages, these models have to be fine-tuned for the task at hand, using additional data sources. This paper presents a weighting mechanism that makes use of available small-scale parallel corpora to improve the performance of multilingual sentence representations on document and sentence alignment. Experiments are conducted with respect to two low-resource languages, Sinhala and Tamil. Results on a newly created dataset of Sinhala-English, Tamil-English, and Sinhala-Tamil show that this new weighting mechanism significantly improves both document and sentence alignment. This dataset, as well as the source-code, is publicly released.

Lahiru Senevirathne, Piyumal Demotte, Binod Karunanayake et al.

Due to the high impact of the fast-evolving fields of machine learning and deep learning, Natural Language Processing (NLP) tasks have further obtained comprehensive performances for highly resourced languages such as English and Chinese. However Sinhala, which is an under-resourced language with a rich morphology, has not experienced these advancements. For sentiment analysis, there exists only two previous research with deep learning approaches, which focused only on document-level sentiment analysis for the binary case. They experimented with only three types of deep learning models. In contrast, this paper presents a much comprehensive study on the use of standard sequence models such as RNN, LSTM, Bi-LSTM, as well as more recent state-of-the-art models such as hierarchical attention hybrid neural networks, and capsule networks. Classification is done at document-level but with more granularity by considering POSITIVE, NEGATIVE, NEUTRAL, and CONFLICT classes. A data set of 15059 Sinhala news comments, annotated with these four classes and a corpus consists of 9.48 million tokens are publicly released. This is the largest sentiment annotated data set for Sinhala so far.

Aloka Fernando, Surangika Ranathunga, Gihan Dias

Out of vocabulary (OOV) is a problem in the context of Machine Translation (MT) in low-resourced languages. When source and/or target languages are morphologically rich, it becomes even worse. Bilingual list integration is an approach to address the OOV problem. This allows more words to be translated than are in the training data. However, since bilingual lists contain words in the base form, it will not translate inflected forms for morphologically rich languages such as Sinhala and Tamil. This paper focuses on data augmentation techniques where bilingual lexicon terms are expanded based on case-markers with the objective of generating new words, to be used in Statistical machine Translation (SMT). This data augmentation technique for dictionary terms shows improved BLEU scores for Sinhala-English SMT.

Vijini Liyanage, Surangika Ranathunga

Existing approaches for automatically generating mathematical word problems are deprived of customizability and creativity due to the inherent nature of template-based mechanisms they employ. We present a solution to this problem with the use of deep neural language generation mechanisms. Our approach uses a Character Level Long Short Term Memory Network (LSTM) to generate word problems, and uses POS (Part of Speech) tags to resolve the constraints found in the generated problems. Our approach is capable of generating Mathematics Word Problems in both English and Sinhala languages with an accuracy over 90%.