About Me
I am a PhD candidate in Mathematics at TU Berlin, working in the Interactive Optimization & Learning (IOL) lab under the supervision of Prof. Sebastian Pokutta. My research focuses on developing agentic AI systems for mathematical and scientific discovery, with particular attention to the interplay between autonomous reasoning and interactive, human-guided exploration. In parallel, I study foundational questions in optimization theory. Prior to this, I worked on learning-augmented algorithms that incorporate predictive models into classical methods from computational algebra.
Recent Publications
Neural Sum-of-Squares: Certifying the Nonnegativity of Polynomials with Transformers
arXiv preprint (under review), 2025
@article{pelleriti2025neural,
title={Neural Sum-of-Squares: Certifying the Nonnegativity of Polynomials with Transformers},
author={Pelleriti, Nico and Spiegel, Christoph and Liu, Shiwei and Mart{\'i}nez-Rubio, David and Zimmer, Max and Pokutta, Sebastian},
journal={arXiv preprint arXiv:2510.13444},
year={2025}
}Approximating Latent Manifolds in Neural Networks via Vanishing Ideals
International Conference on Machine Learning (ICML), 2025
Computational Algebra with Attention: Transformer Oracles for Border Basis Algorithms
Conference on Neural Information Processing Systems (NeurIPS), 2025
@inproceedings{kera2025computational,
title={Computational Algebra with Attention: Transformer Oracles for Border Basis Algorithms},
author={Kera, Hiroshi and Pelleriti, Nico and Ishihara, Yuki and Zimmer, Max and Pokutta, Sebastian},
booktitle={Conference on Neural Information Processing Systems (NeurIPS)},
year={2025},
note={Equal contribution: Kera and Pelleriti}
}