Scholar
Amin Banayeeanzade
Google Scholar ID: QnujJhEAAAAJ
Graduate Research Assistant, University of Southern California
Citations & Impact
Citations
42
H-index
3
i10-index
1
Publications
7
Co-authors
9
list available
Publications
Resume (English only)
Academic Achievements
- 1. Psychological Steering in LLMs: An Evaluation of Effectiveness and Trustworthiness, arXiv, 2025.
- 2. Mechanistic Interpretability of Emotion Inference in Large Language Models, ACL, 2025.
- 3. Theoretical Insights into Overparameterized Models in Multi-Task and Replay-Based Continual Learning, TMLR, 2025.
- 4. GABRIL: Gaze-Based Regularization for Mitigating Causal Confusion in Imitation Learning, IROS 2025.
- 5. Hybrid Learners Do Not Forget: A Brain-Inspired Neuro-Symbolic Approach to Continual Learning, arXiv, 2025.
- 6. A Distinct Unsupervised Reference Model from The Environment Helps Continual Learning, arXiv, 2023.
- 7. Generative vs. Discriminative: Rethinking The Meta-Continual Learning, NeurIPS, 2021.
Research Experience
- 1. Conducting research on the trustworthiness of large language models at USC.
- 2. Developed brain-inspired algorithms for the meta-continual learning problem at Sharif University of Technology.
Education
- 1. Ph.D. student in Computer Science at the University of Southern California (USC), advised by Sai Praneeth Karimireddy.
- 2. Master's in Computer Engineering from Sharif University of Technology, supervised by Mahdieh Soleymani.
- 3. B.Sc. in Electrical Engineering from Sharif University of Technology, jointly supervised by Mahdi Shabany and Zahra Kavehvash.
Background
- Research interests include the trustworthiness of large language models (LLMs) and agentic systems, particularly inference-time failure modes such as hallucinations, insufficient diversity, and reasoning breakdowns. Through data-centric analysis, interpretability tools, and targeted synthetic data generation, aims to establish principled mechanisms for diagnosing, attributing, and correcting these weaknesses, thereby improving real-world robustness and safety.
Miscellany
- Completed an automatic object detection project under clothing in millimeter-wave images as a senior AI researcher at Basir Wave Tech.
