Maksim Zhdanov
I'm a Research Scientist at AIRI.
At the moment, I'm a 4-year BSc computer science student at University of Science and Technology MISIS (Moscow).
My passion is the AI research. I believe that the state-of-the-art technologies will shape our tomorrow as we continuosly understand the laws of our world better and better. I'm interested in AI4Science, particularly right now I'm focused on:
- Generative modeling of biomolecular interactions including flexible all-atom docking models and AlphaFold-like co-folding models
- Reward fine-tuning of generative models and test-time steering. Especially, the design of methods and rewards for stable molecular dynamics and biomolecular modeling
- More efficient architectures for large folding models
I'm also generally interested in the topic of self-supervised learning.
My past experience includes conducting general machine learning research in T-Bank AI Lab mostly in the areas of uncertainty estimation and out-of-distribution detection. Previously, I worked on image/video generative models at X-Labs AI.
In the past, I participated in numerous machine learning competitions.
From cold Siberia
Links
[GitHub, Google Scholar, CV, LinkedIn, Research Gate]
Selected Papers
Electrostatics from Laplacian Eigenbasis for Neural Network Interatomic Potentials
M. Zhdanov, V. Kurenkov
International Conference on Machine Learning (ICML) GenBio Workshop, 2025
[PDF, ArXiv,
Code]
Identity Curvature Laplace Approximation for Improved Out-of-Distribution Detection
M. Zhdanov, S. Dereka, S. Kolesnikov
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025 (previosuly AAAI 2024 Deployable AI Workshop)
[PDF, ArXiv, Poster,
Code]
Diversifying Deep Ensembles: A Saliency Map Approach for Enhanced OOD Detection, Calibration, and Accuracy
S. Dereka, I. Karpukhin, M. Zhdanov, S. Kolesnikov
IEEE International Conference on Image Processing (ICIP), 2024
[PDF, ArXiv,
Code]
Machine learning-assisted close-set X-ray diffraction phase identification of transition metals
M. Zhdanov, A. Zhdanov
International Conference on Learning Representations (ICLR) ML4Materials Workshop, 2023
[PDF, ArXiv,
Poster, Code]
Invited Talks
2024, Research4Kids CIS trip (7 citites).
2023, MISIS.
2023, Tinkoff STEM University.
News
06/2025: Phi-Module was accepted at The 2nd Workshop on Generative AI and Biology (ICML 2025)!
05/2025: Our new pre-print on the electrostatics for neural network interatomic potentials is available.
04/2025: I'm attending ICLR 2025 in Singapore!
11/2024: Our new paper on identity curvature Laplace approximation is accepted to WACV 2024!
Meet us in Tucson, Arizona in February!
09/2024: I'm joining AIRI to work on multidisciplinary problems of AI! Stay tuned for new updates!
06/2024: Our paper on the diversification of deep ensembles is accepted to IEEE ICIP 2024!
Meet us in Adu Dhabi in October!
04/2024: I've just deadlifted 200 KG!
04/2024: We've visited 7 cities of CIS with our research talks.
02/2024: I'm joining Tinkoff Research as a Research Scientist!
12/2023: Our paper on Laplace approximation is accepted to AAAI 2024 DAI Workshop. Stay tuned for updates!
10/2023: I gave a talk about our paper on deep ensembles in MISIS (in Russian).
10/2023: Pre-print of our paper on the diversification of deep ensembles is available.
07/2023: I gave a talk on the presentation day of Tinkoff STEM University.
06/2023: Pre-print of our paper on image retrieval generalization is available.
04/2023: MISIS media highlighted our paper.
03/2023: Our paper on machine learning for X-Ray diffraction is accepted to ICLR 2023
ML4Materials workshop.
01/2023: I'm starting my research internship at Tinkoff AI Lab.
09/2022: I'm starting my undegraduate studies at MISIS university!
