CCS591: Research Methodology & Empirical Methods in Computer Science by Fadratul Hafinaz Hassan

Please Enter Your Research Title

fadratul — 2025-10-14 07:17:37 UTC

Project Description

Dataset Availability

Your Contact

Attention-Augmented Multimodal Fusion Network for Fake News Detection

masarvghadi1 — 2026-04-17 02:49:16 UTC

Description: This project proposes a novel deep learning architecture for fake news detection that integrates textual and visual information using an attention-based fusion mechanism.

Traditional approaches rely heavily on text-only models (e.g., transformers like BERT), while some multimodal approaches use simple feature concatenation. These methods fail to capture cross-modal interactions effectively.

Dataset: FakeNewsNet

Contact: masarvghadi@usm.my

A Congestion-Aware and Link-Stable Routing Protocol for Wireless Sensor Networks in Simulated Environments

masarvghadi1 — 2026-04-17 02:58:17 UTC

Description: This project proposes a novel routing protocol for Wireless Sensor Networks (WSNs) that improves data delivery performance by combining:

adaptive next-hop selection
link-quality-aware routing
traffic-aware load balancing

Contact: masarvghadi@usm.my

An Interactive Learning System for Education

2026-04-17 05:53:00 UTC

This project aims to transform traditional manual learning methods into an interactive and immersive experience using a 3D environment. To provide real-world relevance, selected elements of digital forensics are incorporated, including basic concepts such as identifying potential digital evidence, handling it appropriately, and understanding the importance of maintaining its integrity throughout the process. By combining interactive technology with contextual forensic concepts, the project seeks to enhance student engagement, improve practical understanding, and bridge the gap between conventional teaching approaches and modern digital learning environments.

Digital Transformation

vaithegy@usm.my

Dual-Path Feature Fusion Network with Adaptive Attention for Image Classification

masarvghadi1 — 2026-04-17 07:00:23 UTC

Description: This project proposes a novel deep learning architecture for image classification that enhances feature representation by introducing a dual-path feature extraction mechanism combined with adaptive attention-based fusion.

Dataset: CIFAR-10 / CIFAR-100, Fashion-MNIST or Intel Image Classification Dataset

Contact: masarvghadi@usm.my

ramonausmmy — 2026-04-29 07:12:00 UTC

Explainable AI for Toxic Comment Detection in Multilingual Malaysian Social Media

This project develops a machine learning model to detect toxic comments in social media, focusing on the Malaysian context where users mix Bahasa Melayu, English, and informal language.

The goal is to build an accurate and interpretable system for safer and more transparent content moderation.

Dataset:
Jigsaw Toxic Comment Classification (Kaggle)
(Optional: multilingual/Malaysian-style text)

Contact : ramona@usm.my

Explainable AI for Scam Detection in Social Media Posts

ramonausmmy — 2026-04-29 07:13:38 UTC

Description:
Build a model to detect scam-related posts using textual and behavioural features. Apply XAI to identify patterns such as urgency language, suspicious keywords, and emotional tone.

Dataset:
Fraud / Scam Text Dataset / simulated Malaysian scam messages