Research

In my doctoral research, I focused on mathematical modeling and numerical method for solving strongly correlated fermion problems. I applied that knowledge to study unconventional superconductivity in strongly correlated materials and for modeling classical and quantum phase transitions in heavy-fermion systems.

When working for TypeScore (2016-2018), I worked on model robustness and on methods of learning from heterogeneous data.

When at Fetch.AI (2018-2020), I was one of the lead researchers in a group working on consensus algorithms and decentralized protocol security. I also led the company’s effort in finding applications of machine learning in decentralized systems.

When working at USC, my main focus was on developing novel federated training protocols and exploring applications of machine learning for scientific concept discovery, especially in the fields of materials science.

Publications

• X. Li, H. Tang, S. Chen, Z. Wang, A. Maravi and M. Abram, Context Matters: Data-Efficient Augmentation of Large Language Models for Scientific Applications, arXiv:2312.07069 (2023).
• D. Stripelis, M. Abram, and J.-L. Ambite, Performance weighting for robust federated learning against corrupted sources, arXiv:2205.01184 (2022).
• M. Abram, K. Burghardt, R. Dingreville, J. A. Stewart, G. Ver Steeg, and A. Galstyan, Inferring topological transitions in pattern-forming processes with self-supervised learning, Npj Comput. Mater. 8, 205 (2022).
• Y. Yao, C. Cao, D. Khanna, M. Agarwal, S. Haas, and M. Abram, Emulating quantum dynamics with neural networks via knowledge distillation, Front. Mater. 9, 1060744 (2022).
• J. Ward, M. Abram, and D. Honerkamp, Distributed computer system and method of operation thereof, European Patent no. EP3929742, Bulletin 2021/52 (2021).
• M. Abram, D. Honerkamp, J. Ward, J.-M. Wong, Democratising blockchain: A minimal agency consensus model (2019), presented as working paper at Tokenomics 2019, preprint available at arXiv:2006.05390 (2020).
• Y. Ye, D. Qiu, J. Ward, and M. Abram, Model-free real-time autonomous energy management for a residential multi-carrier energy system: A deep reinforcement learning approach, Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence (IJCAI) (2020).
• M. Abram, M. Zegrodnik, and J. Spałek, Antiferromagnetism, charge density wave, and d-wave superconductivity in the extended t-J-U model: role of intersite Coulomb interaction and a critical overview of renormalized mean field theory, J. Phys. Condens. Matter 29, 365602 (2017).
• M. Abram, M. M. Wysokiński, and J. Spałek, Tricritical wings in UGe2: A microscopic interpretation, J. Magn. Magn. Mater. 400, 27 (2016)
• M. M. Wysokiński, M. Abram, and J. Spałek, Criticalities in the itinerant ferromagnet UGe2, Phys. Rev. B 91, 081108 (2015)
• M. M. Wysokiński, M. Abram, and J. Spałek, Ferromagnetism in UGe2: A microscopic model, Phys. Rev. B 90, 081114(R) (2014)
• M. Abram, t–J–U Model in Mean-Field Approximation: Coexistence of Superconductivity and Antiferromagnetism, Acta Phys. Pol. A 126, A–25 (2014)
• A. Kapanowski and M. Abram, Model of hard spheroplatelets near a hard wall, Phys. Rev. E 89, 062503 (2014)
• M. Abram, J. Kaczmarczyk, J. Jędrak, and J. Spałek, d-wave superconductivity and its coexistence with antiferromagnetism in the t–J–U model: Statistically consistent Gutzwiller approach, Phys. Rev. B 88, 094502 (2013)

Service

Since 2018, I regularly review articles for New Journal of Physics, Journal of Physics: Material, Journal of Physics: Condensed Matter, Physica Scripta, Machine Learning: Science and Technology, IOP SciNotes, and Superconductor Science and Technology.

In 2020, I was awarded the Outstanding Reviewer title by the IOP Journal of Physics: Condensed Matter journal.

Comercial projects

• I was one of the main developer behind the Open Risk Exchange (ORX), the first online portal offering free credit scores for businesses in the United Kingdom. The score is built with help of complex statistical modeling and machine learning algorithms combined with a wide variety of structured and unstructured data sources.
• I am one of the main co-authors of a novel protocol that aims to limit the one-shot monopoly of round-leaders. The protocol was designed to offer a high throughput while reducing the scale of some attacks that traditional blockchains are prone to, like an undercutting attack or Flash Boys 2.0 exploit.

To learn more about my research and work experience, check the CV section of this website.