Non-Commutative Spectral Geometry for Adaptive Quantum-Classical Drug-Target Interaction Prediction
Keywords: Drug-Target Interaction Prediction，UDA
TL;DR: Our approach, Non-Commutative Geometric Adaptation for Molecular Interactions (NCGAMI), reframes the DTI prediction problem within the context of a non-commutative pharmacological manifold
Abstract: Drug-target interactions (DTIs) are fundamental and intricate processes essential for the advancement of drug discovery and design. We present a groundbreaking unified framework for drug-target interaction (DTI) prediction that seamlessly integrates advanced concepts from non-commutative geometry, optimal transport theory, and quantum information science. Our approach, Non-Commutative Geometric Adaptation for Molecular Interactions (NCGAMI), reframes the DTI prediction problem within the context of a non-commutative pharmacological manifold, enabling a profound synthesis of classical and quantum perspectives. By leveraging the spectral action principle, we develop a novel domain adaptation technique that minimizes a geometrically motivated functional, yielding optimal transport maps between pharmacological domains. We establish a deep connection between our framework and non-equilibrium statistical mechanics through a fluctuation theorem for domain adaptation, providing fundamental insights into the thermodynamics of the adaptation process. Our unified variational objective, formulated using geometric quantization, incorporates quantum relative entropy and Liouville volume forms, bridging information-theoretic and geometric aspects of the problem. We introduce a quantum adiabatic optimization algorithm for solving this objective, guaranteeing convergence to the optimal solution under specified conditions. Furthermore, we prove that the algebra of observables generated by our model forms a hyperfinite type III$_1$ factor, revealing a profound link between the algebraic structure of DTI prediction and the geometry of optimal transport. This result enables us to characterize the modular automorphism group governing the evolution of adapted distributions. Extensive numerical experiments demonstrate that NCGAMI significantly outperforms existing state-of-the-art methods across a wide range of DTI prediction tasks, achieving unprecedented accuracy and robustness.
Primary Area: transfer learning, meta learning, and lifelong learning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 13634
Loading