Arda Fazla

Ph.D. Student
Purdue University
afazla (at) purdue.edu

About Me

I am a 2nd year Ph.D. student at Purdue University’s Machine Intelligence and Networked Data Science Group (MINDS) Laboratory advised by Prof. Abolfazl Hashemi.

Research Interests

Optimization Theory: Convergence analysis, variance-reduction methods, multi-objective optimization
LLM Training: Pretraining optimization, adaptive training schedules
Trustworthy Deep Learning: Robustness, fairness, memorization

News

[May 2026] Our paper on coreset selection under spurious correlations is accepted to UAI 2026.

Publications

PREPRINT

Lower Bounds and Proximally Anchored SGD for Non-Convex Minimization Under Unbounded Variance

Arda Fazla, Ege Can Kaya, Antesh Upadhyay, Abolfazl Hashemi

Under Review, 2026.

PDF

PREPRINT

Beyond Bounded Variance: Variance-Reduced Normalized Methods for Nonconvex Optimization under Blum-Gladyshev Noise

Antesh Upadhyay, Arda Fazla, Abolfazl Hashemi

Under Review, 2026.

PDF

UAI

Mitigating Spurious Correlations with Memorization-Guided Dataset De-Biasing

Arda Fazla, Abolfazl Hashemi

Conference on Uncertainty in Artificial Intelligence (UAI), 2026.

PDF Code Poster

Services

Conference Reviewers

Journal Reviewers

Powered by Jekyll and Minimal Light theme.