Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Muhammad Rafi has a total of 56 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Generative AI in Industry 5.0: A Critical Examination of Risks, Threats and Challenges

    byĀ Sameer Faisal, Shehnila Zardari, Muhammad Rafi, Arsalan Khan, Mehar Ullah

    Generative AI in Industry 5.0 enables human-centric, sustainable manufacturing through predictive maintenance, real-time decision-making, and custom design. However, these opportunities bring serious risks: data leakage, algorithmic bias, opaque black-box systems, privacy violations, and regulatory conflicts. This paper presents a Systematic Literature Review (SLR) that identifies and delineates these threats, drawing on risks inherited from Industry 4.0. Based on this analysis, we propose a risk management framework embedding best practices to address ethical, operational, and compliance concerns. The framework guides organizations and decision-makers to responsibly harness Generative AI's potential while safeguarding the humanistic values of Industry 5.0.

  • Open Access

    Agentic table talk

    byĀ Farhan Dhanani, Muhammad Rafi, Uzair Ahmad

    The article presents a heuristic-driven solution called Agentic Table Talk, which utilizes general-purpose Large Language Models (LLMs) to solve the problem of Table Question Answering (Tabular Q/A). The goal of this task is to answer the user query based on the provided table. Earlier approaches required whole tables to prompt LLMs. As tables can grow infinitely large, directly incorporating entire tables into prompts becomes infeasible due to the finite context window of LLMs, which can only handle a limited number of input tokens. In contrast, some approaches are entirely driven by Retrieval Augmented Generation (RAG) to ground the LLMs with K relevant records where the value of K is predefined and hard coded. The main drawback of such approaches is that they do not yield satisfactory results for Tabular Q/A, where data is structured and cellular. Additionally, the meaning of individual cells is often context-dependent and can vary substantially based on the values of neighboring cells. To address these challenges, this article proposes a novel combination of a lightweight heuristic called RAGular SubGraph Retrieval combined with the ReAct Agentic Framework powered by general-purpose LLMs. In this combination, the proposed heuristic adopts the core principles of traditional RAG while placing greater emphasis on preserving the neighborhood relationships of the selected cells. The ReAct Agent receives the user query along with the filtered tabular data as a subgraph that is assumed to be part of a larger, unknown graph. The output from the heuristic guides the subsequent ReAct Agent with a focused input context that consists of only relevant cells and their neighborhood, based on the given input query. The primary objective of the ReAct Agent is to use explicit reasoning to identify the facts necessary to answer the provided user query. This may involve reasoning over the known subgraph or taking actions (such as re-triggering the heuristic with a slightly modified query or finding the closest K-Adjacent neighbors of a relevant cell) to explore new, unexplored cells based on the structure of the already known regions. Both components work together iteratively to discover relevant facts for generating an authentic answer to the user query based on the provided reference table, until the process successfully locates all the necessary information or sufficient confidence in the query’s un-answerability is achieved. The proposed approach is evaluated using general-purpose LLMs of varying sizes on benchmarked AIT-QA ( https://github.com/IBM/AITQA ) and HiTab ( https://github.com/microsoft/HiTab ) datasets, along with a Synthetic Dataset (SD) ( https://github.com/FarhanDhanani/AgenticTableTalk/tree/main/dataset/synthetic ) featuring large tables. The evaluation examines how heuristic-driven input contexts affect the reasoning abilities of LLMs of varying sizes in managing complex tables and compares these results to existing methodologies. The codebase is publicly available on Git (Agentic Table Talk Approach Available At GIT: https://github.com/FarhanDhanani/AgenticTableTalk ) repository, allowing all experiments to be easily re-executed for reproducing the results.

  • Open Access

    Trust-Aware Multimodal Multi-Agent Coordination for Secure UAV Swarms in 6G Networks

    byĀ Hafiz Muhammad Attaullah, Muhammad Ehsan, Muhammad Shoaib Siddiqui, Muhammad Rafi

  • Open Access

    Consensus at a Turning Point: Analyzing Ethereum’s 2022 Migration From PoW to PoS

    byĀ Shahbaz Siddiqui, Sufian Hameed, Muhammed Rafi, Syed Attique Shah, Dirk Draheim

    The development of blockchain consensus mechanisms has put sustainability, scalability, and efficiency first in the list of cryptocurrency studies. As Bitcoin still uses the energy-intensive Proof-of-Work (PoW) algorithm, Ethereum is in the process of switching from PoW to Proof-of-Stake (PoS) with the Ethereum Merge in September 2022. The cause of this change was the rising energy prices, bottlenecks in scalability, and increased concerns among the global population regarding the environmental effects of PoW-based mining. Our study provides a rationale for the Ethereum transition by comparing and contrasting the volume of transactions, network performance through hash rate, and electricity consumption of Bitcoin and Ethereum in their PoW stages.With the help of both historical trend and predictive modelling in the form of the ARIMA framework, the obtained results showed that Bitcoin and Ethereum had different dynamics of growth under PoW, which was increased hash and energy expenditures but limited throughput. The ARIMA-based analysis achieved a mean absolute error (MAE) of 27.76, a root mean square error (RMSE) of 31.16, and a mean absolute percentage error (MAPE) of 26.08, which provided an overall forecast accuracy of 73.9%. These measures show that there is a medium predictive reliability, and volatility is mainly caused by sudden changes in mining intensity and market forces. The predictions further indicated that under PoW, Ethereum would have experienced a steep rise in operational energy consumption and congestion of its network transactions, which would have weakened network sustainability.Results thus confirm that the switch of Ethereum to PoS was not only an environmental need but also a strategic optimization of a transaction to scale higher, operational overhead reduction, and alignment with sustainable and ESG (Environmental, Social, and Governance) goals on a global scale. This piece leads to a better comprehension of the trade-offs between PoW and PoS, as well as the Ethereum consensus shift as the turning point in the evolution of blockchains—establishing a precedent of other cryptocurrencies that can select a balance between performance, security, and sustainability.

  • Open Access

    BUSTED at AraGenEval Shared Task: A Comparative Study of Transformer-Based Models for Arabic AI-Generated Text Detection

    byĀ Ali Zain, Sareem Farooqui, Muhammad Rafi

    This paper details our submission to the AraGenEval Shared Task on Arabic AI-generated text detection, where our team, BUSTED, secured 5th place. We investigated the effectiveness of three pre-trained transformer models: AraELECTRA, CAMeLBERT, and XLM-RoBERTa. Our approach involved fine-tuning each model on the provided dataset for a binary classification task. Our findings revealed a surprising result: the multilingual XLM-RoBERTa model achieved the highest performance with an F1 score of 0.7701, outperforming the specialized Arabic models. This work underscores the complexities of AI-generated text detection and highlights the strong generalization capabilities of multilingual models.

  • Federated unlearning using diffusive noise injection

    byĀ Muhammad Mustafa Ali Usmani, Muhammad Atif Tahir, Humna Faisal, Muhammad Rafi


  • Open Access

    A Multi-Strategy Approach for AI-Generated Text Detection

    byĀ Ali Zain, Sareem Farooqui, Muhammad Rafi

    This paper presents presents three distinct systems developed for the M-DAIGT shared task on detecting AI generated content in news articles and academic abstracts. The systems includes: (1) A fine-tuned RoBERTa-base classifier, (2) A classical TF-IDF + Support Vector Machine (SVM) classifier , and (3) An Innovative ensemble model named Candace, leveraging probabilistic features extracted from multiple Llama-3.2 models processed by a customTransformer encoder.The RoBERTa-based system emerged as the most performant, achieving near-perfect results on both development and test sets.

  • Open Access

    TransFINN ā€œTransparent Feature Integrated Neural Network for Text Feature Selection and Classificationā€

    byĀ Saif Ur Rehman, Ramsha Jat, Muhammad Rafi, Jaroslav Frnda

    This paper introduces TransFINN, a transparent extension artificial neural network that combines transparent feature selection with text classification. TransFINN provides clear insights into which terms most influence its decisions, addressing the ā€œblack–boxā€ limitation of conventional artificial neural networks. We evaluate TransFINN on AG News classification dataset and Yahoo! Answers Topic classification dataset, Two topic classification benchmark datasets. Results demonstrate that TransFINN achieves competitive classification accuracy and macro F1-score against baseline models while significantly improving feature-selection transparency. These findings indicate that TransFINN successfully narrows the gap between the prediction capability of artificial neural networks and the explainability needed for applications where the latter becomes paramount. This work opens up new directions for further research to explore how complex topic classification task could be improved, especially on religious data where a greater semantic insight is required.


  • Open Access

    BUSTED at ARATECT Shared Task: A Comparative Study of Transformer-Based Models for Arabic AI-Generated Text Detection

    byĀ Ali Zain, Sareem Farooqui, Muhammad Rafi

    Ali Zain, Sareem Farooqui, Muhammad Rafi. Proceedings of The Third Arabic Natural Language Processing Conference: Shared Tasks. 2025.