Daniel Saeedi

Daniel Saeedi

PhD candidate in ECE at Georgia Tech. AI/ML — mechanistic interpretability, agentic AI, and generative models for science.

Publications

You can also find my articles on my Google Scholar profile.

Preprints

Autonomous Multi-Agent Reconstruction of Prebiotic Reaction Networks Reveals Organizational Features of Amino Acid Synthesis

Published in Preprint, 2026

ASTRA, a multi-agent AI system, reconstructs prebiotic reaction networks for the twenty canonical amino acids and reveals system-level organization in their synthesis.

Recommended citation: D. Saeedi, N. Pokhrel, L. Gao, C. Wen, E. Bruce, J. C. Aponte, A. Stockton, A. Aghazadeh. (2026). "Autonomous Multi-Agent Reconstruction of Prebiotic Reaction Networks Reveals Organizational Features of Amino Acid Synthesis." Preprint.

Journal Articles

Discriminating Abiotic and Biotic Organics in Meteorite and Terrestrial Samples Using Machine Learning on Mass Spectrometry Data

Published in PNAS Nexus, 4(11), pgaf334, 2025

Machine learning on two-dimensional gas chromatography–mass spectrometry data distinguishes biotic from abiotic organics in meteorites and terrestrial samples.

Recommended citation: D. Saeedi, D. Buckner, T. Walton, J. C. Aponte, A. Aghazadeh. (2025). "Discriminating abiotic and biotic organics in meteorite and terrestrial samples using machine learning on mass spectrometry data." PNAS Nexus, 4(11), pgaf334.
Download Paper

Conference & Workshop Papers

Generative Priors for Cryo-EM Image Reconstruction

Published in ICML 2026 Workshop on Generative and Agentic AI for Biology (GenBio), 2026

* Equal contribution.

Recommended citation: D. Saeedi*, Z. Shabeeb*, D. Tsui*, V. Jamali, A. Aghazadeh. (2026). "Generative Priors for Cryo-EM Image Reconstruction." ICML 2026 Workshop on Generative and Agentic AI for Biology (GenBio).
Download Paper

Protein Circuit Tracing via Cross-layer Transcoders

Published in International Conference on Machine Learning (ICML), 2026

Discovering computational circuits inside protein language models with cross-layer transcoders.

Recommended citation: D. Tsui, K. Talreja, D. Saeedi, A. Aghazadeh. (2026). "Protein Circuit Tracing via Cross-layer Transcoders." International Conference on Machine Learning (ICML).
Download Paper

cryoSENSE: Compressive Sensing Enables High-throughput Microscopy with Sparse and Generative Priors on the Protein Cryo-EM Image Manifold

Published in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026

Generative priors on the cryo-EM image manifold enable higher-throughput microscopy via compressive sensing.

Recommended citation: D. Saeedi*, Z. Shabeeb*, D. Tsui*, V. Jamali, A. Aghazadeh. (2026). "cryoSENSE: Compressive Sensing Enables High-throughput Microscopy with Sparse and Generative Priors on the Protein Cryo-EM Image Manifold." IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).
Download Paper

GT-NLP at SemEval-2025 Task 11: EmoRationale — Evidence-Based Emotion Recognition via Retrieval-Augmented Generation

Published in Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025), 2025

Links: Paper (ACL Anthology)

Recommended citation: D. Saeedi, A. Kheirandish, S. Saeedi, H. Sahour, A. A. Panahi, I. Naeeni. (2025). "GT-NLP at SemEval-2025 Task 11: EmoRationale, Evidence-Based Emotion Recognition via Retrieval-Augmented Generation." SemEval-2025.
Download Paper

Machine Learning Analysis of Mass Spectrometry Data Can Differentiate Organic Distributions in Meteoritic and Terrestrial Geologic Samples

Published in 56th Lunar and Planetary Science Conference (LPSC), Abstract 2729 (Oral Presentation), 2025

Links: Abstract (NASA NTRS, PDF)

Recommended citation: D. K. Buckner, D. Saeedi, T. Walton, J. C. Aponte, A. Aghazadeh. (2025). "Machine Learning Analysis of Mass Spectrometry Data Can Differentiate Organic Distributions in Meteoritic and Terrestrial Geologic Samples." 56th Lunar and Planetary Science Conference (LPSC), Abstract 2729.
Download Paper

AstroAgents: Agentic AI for Scientific Discovery

Published in ICLR 2025 Workshop on Agentic AI for Science / AMS 2025 (Oral Presentation), 2025

A multi-agent AI system that generates scientific hypotheses from mass spectrometry data.

Recommended citation: D. Saeedi, D. K. Buckner, J. C. Aponte, A. Aghazadeh. (2025). "AstroAgents: Agentic AI for Scientific Discovery." ICLR 2025 Workshop on Agentic AI for Science / AMS 2025 (Oral).
Download Paper