Taki Hasan Rafi

Bio

I am a Ph.D. Candidate in Computer Science at Hanyang University (HYU), where I am working at the Data Intelligence Lab with Prof. Dong-Kyu Chae. Before joining HYU, I received my bachelor's degree in Electrical Engineering from Ahsanullah University of Science and Technology, Bangladesh in 2021. My Ph.D. is supported by Brain Korea 21 (NRF) and Samsung Electronics! I occasionally collaborate with researchers from Oracle, USA.

My research focuses on data-efficient and robust learning, particularly in scenarios where models must generalize to unseen or shifting environments with limited access to target domain data. I develop methods for test-time adaptation, robust data selection, and dynamic model adjustment to enhance performance in real-world settings. Additionally, I explore the safety and reliability of AI systems, including LLMs and VLMs, by building evaluation benchmarks, addressing cultural and linguistic biases, and improving robustness in low-resource languages. Ultimately, my goal is to create adaptive, trustworthy AI systems that perform reliably across diverse data distributions and user contexts.

Recent News

One paper gets accepted at CIKM 2025! thanks to the Oracle team Amit and Hitesh.

Three papers get accepted at CSCW 2025! thanks to Wasi and Wahid.

Two papers get accepted at ACL 2025! including one Main and one Findings.

Successfully passed the Ph.D. Candidacy Exam!

Received a PhD research Intersnship offer from Infosys for 6 months residency. But declined the offer due to visa issue.

Serving as a Session Chair at NAACL 2025!

Serving as a Reviewer at ACM MM 2025, ACL 2025 (Industry Track), ACL ARR!

One paper gets accepted at NAACL 2025!

Three papers get accepted at DASFAA 2025!

One paper gets accepted at WWW 2025!

Selected Recent Papers

[C7] Towards Robust Continual Test-Time Adaptation via Neighbor Filtration
Taki Hasan Rafi, Amit Agarwal, Hitesh Patel, Dong-Kyu Chae
In. CIKM 2025, Seoul, South Korea

[C6] Crowdsource, Crawl, or Generate? Creating SEA-VL, a Multicultural Vision-Language Dataset for Southeast Asia
Samuel Cahyawijaya, Holy Lovenia,..., Taki Hasan Rafi,..., Peerat Limkonchotiwat
In. ACL 2025 (Main), Vienna, Austria

[C5] MVTamperBench: Evaluating Robustness of Vision-Language Models
Amit Agarwal, Srikant Panda, Angeline Charles, Bhargava Kumar, Hitesh Patel, Priyaranjan Pattnayak, Taki Hasan Rafi, Tejaswini Kumar, Hansa Meghwani, Karan Gupta, Dong-Kyu Chae
In. ACL 2025 (Findings), Vienna, Austria

[C4] SweEval: Do LLMs Really Swear? A Safety Benchmark for Testing Limits for Enterprise Use
Hitesh Laxmichand Patel, Amit Agarwal, Arion Das, Bhargava Kumar, Srikant Panda, Priyaranjan Pattnayak, Taki Hasan Rafi, Tejaswini Kumar, Dong-Kyu Chae
In. NAACL 2025 (Industry), New Mexico, USA

[C3] Instance-Aware Test-Time Adaptation for Domain Generalization
Taki Hasan Rafi, Serbeter Karlo, Amit Agarwal, Hitesh Patel, Bhargava Kumar, and Dong-Kyu Chae
In. DASFAA 2025, Singapore City, Singapore

[C2] Gaussian Regularization in Neural Graph Learning
Azmine Toushik Wasi, Taki Hasan Rafi, Dong-Kyu Chae
In. DASFAA 2025, Singapore City, Singapore

[C1] GReFEL: Geometry-Aware Reliable Facial Expression Learning under Bias and Imbalanced Data Distribution
Azmine Toushik Wasi*, Taki Hasan Rafi*, Serbeter Karlo, Raima Islam, and Dong-Kyu Chae (*=Equal contributions)
In. ACCV 2024, Hanoi, Vietnam