Explaining reaction coordinates of alanine dipeptide isomerization obtained from deep neural networks using Explainable Artificial Intelligence (XAI)

Abstract

A method for obtaining appropriate reaction coordinates is required to identify transition states distinguishing product and reactant in complex molecular systems. Recently, abundant research has been devoted to obtaining reaction coordinates using artificial neural networks from deep learning literature, where many collective variables are typically utilized in the input layer. However, it is difficult to explain the details of which collective variables contribute to the predicted reaction coordinates owing to the complexity of the nonlinear functions in deep neural networks. To overcome this limitation, we used Explainable Artificial Intelligence (XAI) methods of the Local Interpretable Model-agnostic Explanation (LIME) and the game theory-based framework known as Shapley Additive exPlanations (SHAP). We demonstrated that XAI enables us to obtain the degree of contribution of each collective variable to reaction coordinates that is determined by nonlinear regressions with deep learning for the committor of the alanine dipeptide isomerization in vacuum. In particular, both LIME and SHAP provide important features to the predicted reaction coordinates, which are characterized by appropriate dihedral angles consistent with those previously reported from the committor test analysis. The present study offers an AI-aided framework to explain the appropriate reaction coordinates, which acquires considerable significance when the number of degrees of freedom increases.

Variational deep learning of equilibrium transition path ensembles

“…14,18 Significant advancements in methods to parameterize the time independent committor have been made by resolving this model along physically motivated, predetermined order parameters. 21,23,28,[62][63][64][65][66][67][68] As we show, choosing among a large number of internal coordinates without consideration of their correlation or coupling risks neglecting important aspects of the transition path ensemble. This is because internal coordinates do not form an orthogonal set of coordinates, and collective motions such as the rotations of a single dihedral angle can be coupled with the motions of angles and other dihedrals.…”

Using Diffusion Maps to Analyze Reaction Dynamics for a Hydrogen Combustion Benchmark Dataset

“…A range of procedures for selecting the local scales have been proposed. 27,28 At an abstract level, the application of the normalization defined in (9) or (12) to the locally scaled kernel (13) can be viewed as a special case of the weighted kernel matrix 27 (15) For example, the weighting factor used in ( 9) is w i = 1/p i . It is in (12).…”

“…To find an approximate solution, one can exploit the asymptotic property of the diffusion map (11) and can compute an approximate committor function q by solving the following system of linear equations 2,30 (21) where c and b represent the set of indices for configurations in the complement of A ∪ B and in the metastable region B, respectively. Furthermore, if local distance scales are highly variable for given snapshots, one can use the kernel defined with different local scales (15) for solving the eq 21.…”

How Beneficial or Threatening is Artificial Intelligence?

“…Additionally, 200 million protein 3D structures were predicted and published by Deep Mind and EMBL [5]. In addition, quantum chemical calculations based on the density functional theory (DFT) can be realized using AI [4], and a chemical reaction coordinate of alanine dipeptide isomerization was interpreted using explainable AI (XAI) [6]. There are many other examples like these.…”

“…AI has realized the explanation for the chemical reaction coordinate of alanine dipeptide isomerization [6]. Meuwly reviewed the ML method application [8].…”