Publications
E. Rahmati, Nona Ghazizadeh, Z. Sourati, N. Rouhani, and M. Dehghani. Abstraction as a Memory-Efficient Inductive Bias for Continual Learning. ARR March 2026 Submission, 2026. (Under review)
A. S. Ziabari*, Nona Ghazizadeh*, Z. Sourati, F. Karimi-Malekabadi, P. Piray, and M. Dehghani. Reasoning on a Spectrum: Aligning LLMs to System 1 and System 2 Thinking. COLM 2026 Submission, 2025. (Under review)
P. Golazizian, E. Rahmati, J. Trager, Z. Sourati, Nona Ghazizadeh, G. Chochlakis, J. J. Alcocer, K. Bennett, A. V. Devnani, P. Hejabi, H. G. Muttram, A. K. Padte, M. Saadatinia, C. Wu, A. S. Ziabari, M. Sierra-Arevalo, N. Weller, S. Narayanan, B. A. T. Graham, and M. Dehghani. The Subjectivity of Respect in Police Traffic Stops: Modeling Community Perspectives in Body-Worn Camera Footage. ACL ARR January 2026 Submission, 2026. (Under review)
M. M. Abootorabi, Nona Ghazizadeh, A. Dalili, A. Ghahramani, M. Dehghani, and E. Asgari. AIMA at SemEval-2024 Task 10: History-Based Emotion Recognition in Hindi-English Code-Mixed Conversations. In Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024), Mexico City, Mexico, June 2024. Association for Computational Linguistics.
A. Ghahramani Kure, M. Dehghani, M. M. Abootorabi, Nona Ghazizadeh, S. A. Dalili, and E. Asgari. AIMA at SemEval-2024 Task 3: Simple Yet Powerful Emotion Cause Pair Analysis. In Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024), Mexico City, Mexico, June 2024. Association for Computational Linguistics.
M. Akhi and Nona Ghazizadeh. HDNA: A graph-based change detection in HTML pages (Deface Attack Detection), 2023.
Research Experience
Research Assistant at University of Southern California
Modeling Cognitive Biases and Fast/Slow Thinking in LLMs
Modeling human fast thinking (System 1) and slow thinking (System 2) to understand cognitive biases in large language models, with the goal of manipulating human behavior based on their thinking type.
Mechanistic Representations of Structure-Mapping Theory
Investigated mechanistic representations of Gentner’s Structure-Mapping Theory in LLMs by constructing a multi-domain analogical reasoning dataset and extracting structural mapping vectors through contrastive activation analysis.
Interdisciplinary Review on Regularized Reinforcement Learning
Conducting an interdisciplinary review on regularized reinforcement learning with an emphasis on unified computational frameworks in cognitive neuroscience, robotics, control theory, and NLP alignment.
Cognitive-Inspired Self-Evaluation Reasoning Framework
Developing a cognitive-inspired self-evaluation reasoning framework using reflection token tagging and eye-tracking data analysis to identify decision-making points of cognitive reflection.
Abstraction-Augmented Training (AAT) for Continual Learning
Designed Abstraction-Augmented Training (AAT), a replay-free online continual learning method that improves stability and generalization via abstraction-aware loss optimization.
Perspective-Aware Modeling of LAPD Body-Worn Camera Footage
Developed a perspective-aware modeling framework and a multi-perspective dataset of LAPD body-worn camera footage to analyze subjective interpretations of respect using procedural justice-grounded rubrics.
Under the Supervision of Prof. Morteza Dehghani at Brain and Cognitive Science Department.
August 2024 – Present
Research Assistant at Sharif University of Technology
Classified emotions in code-mixed conversations
Classified emotions in code-mixed conversations using pre-trained large models and GRU networks to integrate prior/future context and sequential flow.
Extracted emotion-cause pairs in multi-modal contexts
Extracted emotion-cause pairs in multi-modal conversational contexts by constructing a causality matrix using transformer-based encoders and prompt engineering.
Multilingual Medical and News Question Answering (RAG)
Implemented Retrieval-Augmented Generation (RAG) models for multilingual medical, drug, and news question-answering systems. It introduces a model that utilizes PubMed QA, PubMed summarization, and news datasets for answering questions.
Bug Issues Ranker Based on Traceability and call graphs
Ranked GitHub project issues using software engineering methods, creating call graphs with traceability and applying NLP techniques for embedding similarity.
Node Identification and weighted tree structures for HTML pages (HDNA)
Developed a novel approach for node identification and weighted tree structures for HTML pages to capture changes in dynamically generated DOM trees (HTML DNA).
Under the Supervision of Dr. Ehsaneddin Asgari and Prof. Abbas Heydarnoori at Computer Engineering Department.
August 2022 – July 2024
Research Assistant at University of New South Wales Sydney
Automated Medical Report Generation for FFA Images
Developed an automated medical report generation system for Fundus Fluorescein Angiography (FFA) images using visual feature extractors, cross-modal memory, and reinforcement learning.
Under the Supervision of Dr. Imran Razzak
June 2023 – September 2023