208 research outputs found
Py-Graph: An Easy-To-Use Interface for Building Graph-Based QSAR Models
No full textGraph Neural Networks (GNNs) have been proved to be effective for the prediction of molecules’ properties using the
molecular graph. These models not only allow to obtain high classification and regression performances, but, with the
help of Explainable Artificial Intelligence, they can be useful for identifying the structural motifs that modulate the
biological activity.
The 3d-qsar.com portal is an online platform that allows users to build several ligand-based and structure-based
models, bridging the need of computational skills.
We implemented a novel application, called Py-Graph, to build and explain GNNs for classification and regression tasks.
Py-Graph represents the first graphic interface that allows users to build QSAR models with GNNs and visualize the parts
of the molecules that most contribute to the prediction
Explainable AI in drug design: self-interpretable graph neural network for molecular property prediction using concept whitening
No full textMolecular properties' prediction is a fundamental task in the field of drug discovery. Several works focus on using graph neural networks as they allow to directly use molecular graph representations. Although they have been successfully applied in a variety of applications, these models lack in the transparency of the decision process. In this work, we adapt the concept whitening explainability method to graph neural networks. This approach allows building an inherently interpretable model, by aligning the axes of the latent space with known concepts of interest, thus providing a straightforward way of extracting them. We test the models on the BBBP dataset from MoleculeNet. Starting from previous works, we identify the most significant molecular
properties to be used as concepts to perform the classification. We show that the
addition of concept whitening layers brings an improvement in both classification performance and interpretability. Finally, we show how to obtain both structural and conceptual explanations for the predictions
Molecule Generation from Input-Attribution over Graph Convolutional Networks
No full textIt is well known that Drug Design is often a costly process both in terms of time and economic effort. While good Quantitative Structure-Activity Relationship models (QSAR) can help predicting molecular properties without the need to synthesize them, it is still required to come up with new molecules to be tested. This is mostly done in lack of tools to determine which modifications are more promising or which aspects of a molecule are more influential for the final activity/property. Here we present an automatic process which involves Graph Convolutional Network models and input-attribution methods to generate new molecules. We also explore the problems of over-optimization and applicability, recognizing them as two important aspects in the practical use of such automatic tools
Prototype-based Interpretable Graph Neural Networks
No full textGraph neural networks have proved to be a key tool for dealing with many problems and domains such as chemistry, natural language processing and social networks. While the structure of the layers is simple, it is difficult to identify the patterns learned by the graph neural network. Several works propose post-hoc methods to explain graph predictions, but few of them try to generate interpretable models. Conversely, the topic of the interpretable models is highly investigated in image recognition. Given the similarity between image and graph domains, we analyze the adaptability of prototype-based neural networks for graph and node classification. In particular, we investigate the use of two interpretable networks, ProtoPNet and TesNet, in the graph domain. We show that the adapted networks manage to reach better or higher accuracy scores than their respective black-box models and comparable performances with state-of-the-art self-explainable models. Showing how to extract ProtoPNet and TesNet explanations on graph neural networks, we further study how to obtain global and local explanations for the trained models. We then evaluate the explanations of the interpretable models by comparing them with post-hoc approaches and self-explainable models. Our findings show that the application of TesNet and ProtoPNet to the graph domain produces qualitative predictions while improving their reliability and transparency
Memory Replay For Continual Learning With Spiking Neural Networks
No full textTwo of the most impressive features of biological neural networks are their high energy efficiency and their ability to continuously adapt to varying inputs. On the contrary, the amount of power required to train top-performing deep learning models rises as they become more complex. This is the main reason for the increasing research interest in spiking neural networks, which mimic the functioning of the human brain achieving similar performances to artificial neural networks, but with much lower energy costs. However, even this type of network is not provided with the ability to incrementally learn new tasks, with the main obstacle being catastrophic forgetting. This paper investigates memory replay as a strategy to mitigate catastrophic forgetting in spiking neural networks. Experiments are conducted on the MNIST-split dataset in both class-incremental learning and task-free continual learning scenarios
Understanding Deep RL agent decisions: a novel interpretable approach with trainable prototypes
No full textDeep reinforcement learning (DRL) models have shown great promise in various applications, but their practical adoption in critical domains is limited due to their opaque decision-making processes. To address this challenge, explainable AI (XAI) techniques aim to enhance transparency and interpretability of black-box models. However, most current interpretable systems focus on supervised learning problems,
leaving reinforcement learning relatively unexplored.
This paper extends the work of PW-Net, an interpretable wrapper model for DRL agents inspired by image classification methodologies. We introduce Shared-PW-Net, an interpretable deep learning model that features a fully trainable prototype layer. Unlike PW-Net, Shared-PW-Net does not rely on
pre-existing prototypes. Instead, it leverages the concept of ProtoPool to automatically learn general prototypes assigned to actions during training. Additionally, we propose a novel prototype initialization method that significantly improves the model’s performance.
Through extensive experimentation, we demonstrate that our Shared-PW-Net achieves the same reward performance as existing methods without requiring human intervention. Our model’s fully trainable prototype layer, coupled with the innovative prototype initialization approach, contributes to a clearer and more interpretable decision-making process. The code for this work is publicly available for further exploration and applications
Semi-Supervised GCN for learning Molecular Structure-Activity Relationships
No full textSince the introduction of artificial intelligence in medicinal chemistry, the
necessity has emerged to analyse how molecular property variation is modulated
by either single atoms or chemical groups. In this paper, we propose to train
graph-to-graph neural network using semi-supervised learning for attributing
structure-property relationships. As initial case studies we apply the method
to solubility and molecular acidity while checking its consistency in
comparison with known experimental chemical data. As final goal, our approach
could represent a valuable tool to deal with problems such as activity cliffs,
lead optimization and de-novo drug design
On the Mechanism of Asymmetric Epoxidation of Enones Catalyzed by alpha,alpha-L-Diarylprolinols: A Theoretical Insight
No full textThe mechanism of the asymmetric epoxidation of enones with tert-butyl hydroperoxide promoted by a,a-L-diarylprolinols has been studied by second order MollerPlesset perturbation theory (MP2) and density functional theory (DFT) computations. The non-covalent activation of the reactants, through an effective network of hydrogen bonding interactions, initially hypothesized on the basis of the available experimental data, is shown to constitute an energetically viable pathway. According to the non-covalent route, the reaction follows a two-step nucleophilic epoxidation mechanism, with the first oxa-Michael addition being the rate- and stereoselectivity-determining step. Formation of the (2R,3S)-enantiomer of the epoxide derived from trans-chalcone is predicted to be energetically favoured, in agreement with the experimental findings
Enantioselective conjugate addition of malononitrile to chalcones promoted by α,α-L-diaryl prolinols: Noncovalent versus covalent catalysis?
No full textThe enantioselective conjugate addition of malononitrile to trans-chalcones has been investigated as a case study using easily available α,α-L-diaryl prolinols as promoters. Both experimental and computational results are consistent with a bifunctional noncovalent mode of activation of the reactive partners, provided by the secondary amine and hydroxyl groups of the promoter as general base and acid catalysis. This most energetically affordable pathway predicts predominant formation of the R-configured adducts, which is in agreement with the experimental findings. The present study addresses, for the first time, the ability of proline derivatives to assist product formation through noncovalent catalysis, in analogy to the mode of action typically recognized for Cinchona alkaloids
Corrigendum: Platinated Nucleotides are Substrates for the Human Mitochondrial Deoxynucleotide Carrier (DNC) and DNA Polymerase γ: Relevance for the Development of New Platinum-Based Drugs (ChemistrySelect, (2016), 1, (4633-4637), 10.1002/slct.201600961)
No full textIn agreement with all authors of this paper the order of the authors and the contribution “Paola Lunetti[+], Alessandro Romano[+], Chiara Carrisi, Daniela Antonucci, Tiziano Verri, Giuseppe E. De Benedetto, Vincenza Dolce, Francesco P. Fanizzi, Michele Benedetti,* and Loredana Capobianco.* [+] These authors contributed equally to this paper. * Corresponding authors: Michele Benedetti and Loredana Capobianco, Department of Biological and Environmental Sciences and Technologies, University of Salento, 73100 Lecce.” is corrected to read the original order of the first submitted version of the paper: “Chiara Carrisi[+], Alessandro Romano[+], Paola Lunetti, Daniela Antonucci, Tiziano Verri, Giuseppe E. De Benedetto, Vincenza Dolce, Francesco P. Fanizzi, Michele Benedetti,* and Loredana Capobianco.* [+] These authors contributed equally to this paper. * Corresponding authors: Loredana Capobianco and Michele Benedetti, Department of Biological and Environmental Sciences and Technologies, University of Salento, 73100 Lecce.” For completeness the following individual contributions of the authors were added in the Supporting Information: “Author contributions: L.C., M.B., C.C., A.R. designed research; C.C., A.R., P.L. and D.A. performed research; L.C., M.B., C.C., A.R., V.D. and P.L. analysed data; F.P.F., A.R., L.C. and M.B. wrote the paper; L.C., M.B., T.V., G.D.B. and F.P.F. active discussion paper revision.”
- …
