Fahim Faisal

Fahim Faisal, Antonios Anastasopoulos · cmu

Large pretrained multilingual models, trained on dozens of languages, have delivered promising results due to cross-lingual learning capabilities on variety of language tasks. Further adapting these models to specific languages, especially ones unseen during pre-training, is an important goal towards expanding the coverage of language technologies. In this study, we show how we can use language phylogenetic information to improve cross-lingual transfer leveraging closely related languages in a structured, linguistically-informed manner. We perform adapter-based training on languages from diverse language families (Germanic, Uralic, Tupian, Uto-Aztecan) and evaluate on both syntactic and semantic tasks, obtaining more than 20% relative performance improvements over strong commonly used baselines, especially on languages unseen during pre-training.

Fahim Faisal, Antonios Anastasopoulos · cmu

Pretrained language models (PLMs) often fail to fairly represent target users from certain world regions because of the under-representation of those regions in training datasets. With recent PLMs trained on enormous data sources, quantifying their potential biases is difficult, due to their black-box nature and the sheer scale of the data sources. In this work, we devise an approach to study the geographic bias (and knowledge) present in PLMs, proposing a Geographic-Representation Probing Framework adopting a self-conditioning method coupled with entity-country mappings. Our findings suggest PLMs' representations map surprisingly well to the physical world in terms of country-to-country associations, but this knowledge is unequally shared across languages. Last, we explain how large PLMs despite exhibiting notions of geographical proximity, over-amplify geopolitical favouritism at inference time.

Md Mushfiqur Rahman, Fardin Ahsan Sakib, Fahim Faisal et al. · cmu

Choosing an appropriate tokenization scheme is often a bottleneck in low-resource cross-lingual transfer. To understand the downstream implications of text representation choices, we perform a comparative analysis on language models having diverse text representation modalities including 2 segmentation-based models (\texttt{BERT}, \texttt{mBERT}), 1 image-based model (\texttt{PIXEL}), and 1 character-level model (\texttt{CANINE}). First, we propose a scoring Language Quotient (LQ) metric capable of providing a weighted representation of both zero-shot and few-shot evaluation combined. Utilizing this metric, we perform experiments comprising 19 source languages and 133 target languages on three tasks (POS tagging, Dependency parsing, and NER). Our analysis reveals that image-based models excel in cross-lingual transfer when languages are closely related and share visually similar scripts. However, for tasks biased toward word meaning (POS, NER), segmentation-based models prove to be superior. Furthermore, in dependency parsing tasks where word relationships play a crucial role, models with their character-level focus, outperform others. Finally, we propose a recommendation scheme based on our findings to guide model selection according to task and language requirements.

Md Mahfuz Ibn Alam, Ruoyu Xie, Fahim Faisal et al. · cmu

This report describes GMU's sentiment analysis system for the SemEval-2023 shared task AfriSenti-SemEval. We participated in all three sub-tasks: Monolingual, Multilingual, and Zero-Shot. Our approach uses models initialized with AfroXLMR-large, a pre-trained multilingual language model trained on African languages and fine-tuned correspondingly. We also introduce augmented training data along with original training data. Alongside finetuning, we perform phylogeny-based adapter tuning to create several models and ensemble the best models for the final submission. Our system achieves the best F1-score on track 5: Amharic, with 6.2 points higher F1-score than the second-best performing system on this track. Overall, our system ranks 5th among the 10 systems participating in all 15 tracks.

Fatin Abrar Shams, Rashed Hasan Ratul, Ahnaf Islam Naf et al.

Superconductors have been among the most fascinating substances, as the fundamental concept of superconductivity as well as the correlation of critical temperature and superconductive materials have been the focus of extensive investigation since their discovery. However, superconductors at normal temperatures have yet to be identified. Additionally, there are still many unknown factors and gaps of understanding regarding this unique phenomenon, particularly the connection between superconductivity and the fundamental criteria to estimate the critical temperature. To bridge the gap, numerous machine learning techniques have been established to estimate critical temperatures as it is extremely challenging to determine. Furthermore, the need for a sophisticated and feasible method for determining the temperature range that goes beyond the scope of the standard empirical formula appears to be strongly emphasized by various machine-learning approaches. This paper uses a stacking machine learning approach to train itself on the complex characteristics of superconductive materials in order to accurately predict critical temperatures. In comparison to other previous accessible research investigations, this model demonstrated a promising performance with an RMSE of 9.68 and an R2 score of 0.922. The findings presented here could be a viable technique to shed new insight on the efficient implementation of the stacking ensemble method with hyperparameter optimization (HPO).

Fahim Faisal, Orevaoghene Ahia, Aarohi Srivastava et al.

Language technologies should be judged on their usefulness in real-world use cases. An often overlooked aspect in natural language processing (NLP) research and evaluation is language variation in the form of non-standard dialects or language varieties (hereafter, varieties). Most NLP benchmarks are limited to standard language varieties. To fill this gap, we propose DIALECTBENCH, the first-ever large-scale benchmark for NLP on varieties, which aggregates an extensive set of task-varied variety datasets (10 text-level tasks covering 281 varieties). This allows for a comprehensive evaluation of NLP system performance on different language varieties. We provide substantial evidence of performance disparities between standard and non-standard language varieties, and we also identify language clusters with large performance divergence across tasks. We believe DIALECTBENCH provides a comprehensive view of the current state of NLP for language varieties and one step towards advancing it further. Code/data: https://github.com/ffaisal93/DialectBench

Fahim Faisal, Md Mushfiqur Rahman, Antonios Anastasopoulos

There has been little systematic study on how dialectal differences affect toxicity detection by modern LLMs. Furthermore, although using LLMs as evaluators ("LLM-as-a-judge") is a growing research area, their sensitivity to dialectal nuances is still underexplored and requires more focused attention. In this paper, we address these gaps through a comprehensive toxicity evaluation of LLMs across diverse dialects. We create a multi-dialect dataset through synthetic transformations and human-assisted translations, covering 10 language clusters and 60 varieties. We then evaluated three LLMs on their ability to assess toxicity across multilingual, dialectal, and LLM-human consistency. Our findings show that LLMs are sensitive in handling both multilingual and dialectal variations. However, if we have to rank the consistency, the weakest area is LLM-human agreement, followed by dialectal consistency. Code repository: \url{https://github.com/ffaisal93/dialect_toxicity_llm_judge}

Fahim Faisal, Antonios Anastasopoulos

This report presents GMUNLP's participation to the Dialect-Copa shared task at VarDial 2024, which focuses on evaluating the commonsense reasoning capabilities of large language models (LLMs) on South Slavic micro-dialects. The task aims to assess how well LLMs can handle non-standard dialectal varieties, as their performance on standard languages is already well-established. We propose an approach that combines the strengths of different types of language models and leverages data augmentation techniques to improve task performance on three South Slavic dialects: Chakavian, Cherkano, and Torlak. We conduct experiments using a language-family-focused encoder-based model (BERTić) and a domain-agnostic multilingual model (AYA-101). Our results demonstrate that the proposed data augmentation techniques lead to substantial performance gains across all three test datasets in the open-source model category. This work highlights the practical utility of data augmentation and the potential of LLMs in handling non-standard dialectal varieties, contributing to the broader goal of advancing natural language understanding in low-resource and dialectal settings. Code:https://github.com/ffaisal93/dialect_copa

Fahim Faisal, Antonios Anastasopoulos

The capacity and effectiveness of pre-trained multilingual models (MLMs) for zero-shot cross-lingual transfer is well established. However, phenomena of positive or negative transfer, and the effect of language choice still need to be fully understood, especially in the complex setting of massively multilingual LMs. We propose an \textit{efficient} method to study transfer language influence in zero-shot performance on another target language. Unlike previous work, our approach disentangles downstream tasks from language, using dedicated adapter units. Our findings suggest that some languages do not largely affect others, while some languages, especially ones unseen during pre-training, can be extremely beneficial or detrimental for different target languages. We find that no transfer language is beneficial for all target languages. We do, curiously, observe languages previously unseen by MLMs consistently benefit from transfer from almost any language. We additionally use our modular approach to quantify negative interference efficiently and categorize languages accordingly. Furthermore, we provide a list of promising transfer-target language configurations that consistently lead to target language performance improvements. Code and data are publicly available: https://github.com/ffaisal93/neg_inf

Fahim Faisal, Yinkai Wang, Antonios Anastasopoulos

As language technologies become more ubiquitous, there are increasing efforts towards expanding the language diversity and coverage of natural language processing (NLP) systems. Arguably, the most important factor influencing the quality of modern NLP systems is data availability. In this work, we study the geographical representativeness of NLP datasets, aiming to quantify if and by how much do NLP datasets match the expected needs of the language speakers. In doing so, we use entity recognition and linking systems, also making important observations about their cross-lingual consistency and giving suggestions for more robust evaluation. Last, we explore some geographical and economic factors that may explain the observed dataset distributions. Code and data are available here: https://github.com/ffaisal93/dataset_geography. Additional visualizations are available here: https://nlp.cs.gmu.edu/project/datasetmaps/.

Fahim Faisal, Sharlina Keshava, Md Mahfuz ibn Alam et al.

Question answering (QA) systems are now available through numerous commercial applications for a wide variety of domains, serving millions of users that interact with them via speech interfaces. However, current benchmarks in QA research do not account for the errors that speech recognition models might introduce, nor do they consider the language variations (dialects) of the users. To address this gap, we augment an existing QA dataset to construct a multi-dialect, spoken QA benchmark on five languages (Arabic, Bengali, English, Kiswahili, Korean) with more than 68k audio prompts in 24 dialects from 255 speakers. We provide baseline results showcasing the real-world performance of QA systems and analyze the effect of language variety and other sensitive speaker attributes on downstream performance. Last, we study the fairness of the ASR and QA models with respect to the underlying user populations. The dataset, model outputs, and code for reproducing all our experiments are available: https://github.com/ffaisal93/SD-QA.

Fahim Faisal, Antonios Anastasopoulos

Human knowledge is collectively encoded in the roughly 6500 languages spoken around the world, but it is not distributed equally across languages. Hence, for information-seeking question answering (QA) systems to adequately serve speakers of all languages, they need to operate cross-lingually. In this work we investigate the capabilities of multilingually pre-trained language models on cross-lingual QA. We find that explicitly aligning the representations across languages with a post-hoc fine-tuning step generally leads to improved performance. We additionally investigate the effect of data size as well as the language choice in this fine-tuning step, also releasing a dataset for evaluating cross-lingual QA systems. Code and dataset are publicly available here: https://github.com/ffaisal93/aligned_qa

Anjali Kantharuban, Aarohi Srivastava, Fahim Faisal et al.

Existing sentence representations primarily encode what a sentence says, rather than how it is expressed, even though the latter is important for many applications. In contrast, we develop sentence representations that capture style and dialect, decoupled from semantic content. We call this the task of idiolectal representation learning. We introduce IDIOLEX, a framework for training models that combines supervision from a sentence's provenance with linguistic features of a sentence's content, to learn a continuous representation of each sentence's style and dialect. We evaluate the approach on dialects of both Arabic and Spanish. The learned representations capture meaningful variation and transfer across domains for analysis and classification. We further explore the use of these representations as training objectives for stylistically aligning language models. Our results suggest that jointly modeling individual and community-level variation provides a useful perspective for studying idiolect and supports downstream applications requiring sensitivity to stylistic differences, such as developing diverse and accessible LLMs.

Fahim Faisal, Kaiqiang Song, Song Wang et al.

While reinforcement learning has advanced the reasoning abilities of Large Language Models (LLMs), these gains are largely confined to English, creating a significant performance disparity across languages. To address this, we introduce Pivot-Based Reinforcement Learning with Semantically Verifiable Rewards (PB-RLSVR), a novel framework that enhances multilingual reasoning by circumventing the need for human-annotated data in target languages. Our approach employs a high-performing English LLM as a "pivot" model to generate reference responses for reasoning tasks. A multilingual model is then rewarded based on the semantic equivalence of its responses to the English reference, effectively transferring the pivot model's reasoning capabilities across languages. We investigate several cross-lingual semantic reward functions, including those based on embeddings and machine translation. Extensive experiments on a suite of multilingual reasoning benchmarks show that our method significantly narrows the performance gap between English and other languages, substantially outperforming traditional PPO baselines. Specifically, our PB-RLSVR framework improves the average multilingual performance of Llama-3.1-8B-Instruct and Qwen3-32B by 16.41% and 10.17%, respectively, demonstrating a powerful and data-efficient approach to building truly multilingual reasoning agents.

Fahim Faisal

Modern NLP breakthrough includes large multilingual models capable of performing tasks across more than 100 languages. State-of-the-art language models came a long way, starting from the simple one-hot representation of words capable of performing tasks like natural language understanding, common-sense reasoning, or question-answering, thus capturing both the syntax and semantics of texts. At the same time, language models are expanding beyond our known language boundary, even competitively performing over very low-resource dialects of endangered languages. However, there are still problems to solve to ensure an equitable representation of texts through a unified modeling space across language and speakers. In this survey, we shed light on this iterative progression of multilingual text representation and discuss the driving factors that ultimately led to the current state-of-the-art. Subsequently, we discuss how the full potential of language democratization could be obtained, reaching beyond the known limits and what is the scope of improvement in that space.

Yueqi Song, Catherine Cui, Simran Khanuja et al.

Despite the major advances in NLP, significant disparities in NLP system performance across languages still exist. Arguably, these are due to uneven resource allocation and sub-optimal incentives to work on less resourced languages. To track and further incentivize the global development of equitable language technology, we introduce GlobalBench. Prior multilingual benchmarks are static and have focused on a limited number of tasks and languages. In contrast, GlobalBench is an ever-expanding collection that aims to dynamically track progress on all NLP datasets in all languages. Rather than solely measuring accuracy, GlobalBench also tracks the estimated per-speaker utility and equity of technology across all languages, providing a multi-faceted view of how language technology is serving people of the world. Furthermore, GlobalBench is designed to identify the most under-served languages, and rewards research efforts directed towards those languages. At present, the most under-served languages are the ones with a relatively high population, but nonetheless overlooked by composite multilingual benchmarks (like Punjabi, Portuguese, and Wu Chinese). Currently, GlobalBench covers 966 datasets in 190 languages, and has 1,128 system submissions spanning 62 languages.

Ishrak Jahan Ratul, Ummay Habiba Wani, Mirza Muntasir Nishat et al.

Bone Marrow Transplant, a gradational rescue for a wide range of disorders emanating from the bone marrow, is an efficacious surgical treatment. Several risk factors, such as post-transplant illnesses, new malignancies, and even organ damage, can impair long-term survival. Therefore, technologies like Machine Learning are deployed for investigating the survival prediction of BMT receivers along with the influences that limit their resilience. In this study, an efficient survival classification model is presented in a comprehensive manner, incorporating the Chi-squared feature selection method to address the dimensionality problem and Hyper Parameter Optimization (HPO) to increase accuracy. A synthetic dataset is generated by imputing the missing values, transforming the data using dummy variable encoding, and compressing the dataset from 59 features to the 11 most correlated features using Chi-squared feature selection. The dataset was split into train and test sets at a ratio of 80:20, and the hyperparameters were optimized using Grid Search Cross-Validation. Several supervised ML methods were trained in this regard, like Decision Tree, Random Forest, Logistic Regression, K-Nearest Neighbors, Gradient Boosting Classifier, Ada Boost, and XG Boost. The simulations have been performed for both the default and optimized hyperparameters by using the original and reduced synthetic dataset. After ranking the features using the Chi-squared test, it was observed that the top 11 features with HPO, resulted in the same accuracy of prediction (94.73%) as the entire dataset with default parameters. Moreover, this approach requires less time and resources for predicting the survivability of children undergoing BMT. Hence, the proposed approach may aid in the development of a computer-aided diagnostic system with satisfactory accuracy and minimal computation time by utilizing medical data records.

Rafid Ameer Mahmud, Fahim Faisal, Saaduddin Mahmud et al.

Multi-agent Markov Decision Process (MMDP) has been an effective way of modelling sequential decision making algorithms for multi-agent cooperative environments. A number of algorithms based on centralized and decentralized planning have been developed in this domain. However, dynamically changing environment, coupled with exponential size of the state and joint action space, make it difficult for these algorithms to provide both efficiency and scalability. Recently, Centralized planning algorithm FV-MCTS-MP and decentralized planning algorithm \textit{Alternate maximization with Behavioural Cloning} (ABC) have achieved notable performance in solving MMDPs. However, they are not capable of adapting to dynamically changing environments and accounting for the lack of communication among agents, respectively. Against this background, we introduce a simulation based online planning algorithm, that we call SiCLOP, for multi-agent cooperative environments. Specifically, SiCLOP tailors Monte Carlo Tree Search (MCTS) and uses Coordination Graph (CG) and Graph Neural Network (GCN) to learn cooperation and provides real time solution of a MMDP problem. It also improves scalability through an effective pruning of action space. Additionally, unlike FV-MCTS-MP and ABC, SiCLOP supports transfer learning, which enables learned agents to operate in different environments. We also provide theoretical discussion about the convergence property of our algorithm within the context of multi-agent settings. Finally, our extensive empirical results show that SiCLOP significantly outperforms the state-of-the-art online planning algorithms.

Junayed Mahmud, Fahim Faisal, Raihan Islam Arnob et al.

Automated source code summarization is a popular software engineering research topic wherein machine translation models are employed to "translate" code snippets into relevant natural language descriptions. Most evaluations of such models are conducted using automatic reference-based metrics. However, given the relatively large semantic gap between programming languages and natural language, we argue that this line of research would benefit from a qualitative investigation into the various error modes of current state-of-the-art models. Therefore, in this work, we perform both a quantitative and qualitative comparison of three recently proposed source code summarization models. In our quantitative evaluation, we compare the models based on the smoothed BLEU-4, METEOR, and ROUGE-L machine translation metrics, and in our qualitative evaluation, we perform a manual open-coding of the most common errors committed by the models when compared to ground truth captions. Our investigation reveals new insights into the relationship between metric-based performance and model prediction errors grounded in an empirically derived error taxonomy that can be used to drive future research efforts

Nazim Choudhury, Fahim Faisal, Matloob Khushi

Literature-based knowledge discovery process identifies the important but implicit relations among information embedded in published literature. Existing techniques from Information Retrieval and Natural Language Processing attempt to identify the hidden or unpublished connections between information concepts within published literature, however, these techniques undermine the concept of predicting the future and emerging relations among scientific knowledge components encapsulated within the literature. Keyword Co-occurrence Network (KCN), built upon author selected keywords (i.e., knowledge entities), is considered as a knowledge graph that focuses both on these knowledge components and knowledge structure of a scientific domain by examining the relationships between knowledge entities. Using data from two multidisciplinary research domains other than the medical domain, capitalizing on bibliometrics, the dynamicity of temporal KCNs, and a Long Short Term Memory recurrent neural network, this study proposed a framework to successfully predict the future literature-based discoveries - the emerging connections among knowledge units. Framing the problem as a dynamic supervised link prediction task, the proposed framework integrates some novel node and edge-level features. Temporal importance of keywords computed from both bipartite and unipartite networks, communities of keywords, built upon genealogical relations, and relative importance of temporal citation counts used in the feature construction process. Both node and edge-level features were input into an LSTM network to forecast the feature values for positive and negatively labeled non-connected keyword pairs and classify them accurately. High classification performance rates suggest that these features are supportive both in predicting the emerging connections between scientific knowledge units and emerging trend analysis.

Fahim Faisal, Shafkat Ahmed Bhuiyan, Abu Raihan Mostofa Kamal

A method to identify probable diseases from the unstructured textual input (eg, health forum posts) by incorporating a lexicographic and semantic feature based two-phase text classification module and a symptom-disease correlation-based similarity measurement module. One notable aspect of my approach was to develop a competent algorithm to extract all inherent features from the data source to make a better decision.