α-MOP: Molecule optimization with α-divergence
… -Molecular OPtimization (α-MOP), which uses α-divergence to … Evaluated on multiple datasets,
α-MOP obtains success rate … and RL based molecule optimization methods. Index Terms—…
α-MOP obtains success rate … and RL based molecule optimization methods. Index Terms—…
A comprehensive review of molecular optimization in artificial intelligence‐based drug discovery
… [60] proposed a method named α‐MOP to unify the molecular mapping‐based method and
Reinforcement learning (RL) using α‐divergence. By controlling the α value in the objective …
Reinforcement learning (RL) using α‐divergence. By controlling the α value in the objective …
Mimosa: Multi-constraint molecule sampling for molecule optimization
… To allow for flexible and efficient molecule optimization on multiple properties, we propose
a new sampling based molecule optimization framework named MultI-constraint MOlecule …
a new sampling based molecule optimization framework named MultI-constraint MOlecule …
An open unified deep graph learning framework for discovering drug leads
… : To provide stronger interpretability, the molecule optimization in this paper utilizes graph
node classification logic to change the element symbols of one single atom at one time, …
node classification logic to change the element symbols of one single atom at one time, …
Disentangle VAE for Molecular Generation
Y Wang, Q Song - arXiv preprint arXiv:2202.06794, 2022 - arxiv.org
… a molecule with low solubility as an input, we want to generate a molecule that is not only
similar to the input molecule, … a new molecule that is similar to the given molecule. Second, we …
similar to the input molecule, … a new molecule that is similar to the given molecule. Second, we …
Accelerating Stochastic Probabilistic Inference
M Liu, S Bu - arXiv preprint arXiv:2203.07585, 2022 - arxiv.org
… with stochastic optimization, following noisy estimates of the natural gradient. However,
almost all the state-of-the-art SVI algorithms are based on first-order optimization algorithm and …
almost all the state-of-the-art SVI algorithms are based on first-order optimization algorithm and …
Splitting numerical integration for matrix completion
Q Song - arXiv preprint arXiv:2202.06482, 2022 - arxiv.org
… Departing from a perspective of differential systems, we study the solution of the optimization
problem in (6). In particular, we consider an alternative formulation for Problem (2) in the …
problem in (6). In particular, we consider an alternative formulation for Problem (2) in the …