Machine learning of generic reactions: 1. Scope of the project; the GRAMS program

Jauffret, Philippe; Hanser, Thierry; Tonnelier, Christian; Kaufmann, G.

doi:10.1016/0898-5529(90)90059-h

Cited by 11 publications

(5 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Section: Methodsmentioning

confidence: 99%

“…The method must be able to mine several thousands of graphs containing hundred thousand bonds. For this reason, the algorithm should have a linear complexity in the number of examples, contrasting with bottom-up graph-based learning methods 14,15,29 The discoVery of formable bonds consists in deciding whether the hypothesis "the input bond b is formable" is true given the position of b in G and given the set E of examples. Closely related to this problem, the ranking of formable bonds consists in deciding which one of two input bonds b 1 and b 2 in the input molecule G is the easiest to form.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Graph-Mining Algorithm for the Evaluation of Bond Formability

Pennerath

Niel

Vismara

et al. 2010

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

The formability of a bond in a target molecule is a bond property related to the problem of finding a reaction that synthesizes the target by forming the bond: the easier this problem, the higher the formability. Bond formability provides an interesting piece of information that might be used for selecting strategic bonds during a retrosynthesic analysis or for assessing synthetic accessibility in virtual screening. The article describes a graph-mining algorithm called GemsBond that evaluates formability of bonds by mining structural environments contained in several thousand molecular graphs of reaction products. When tested on reaction databases, GemsBond recognizes most formed bonds in reaction products and provides explanations consistent with knowledge in organic synthesis.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Graph-Mining Algorithm for the Evaluation of Bond Formability

Pennerath

Niel

Vismara

et al. 2010

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Amazon Echo or Google Home, that are considered to be intelligence since they are capable of comprehending the natural language questions and provide the feedback using the same natural language. D. Petrelli, A. Dadzie and V. Lanfranchi in [10] posit that many users of the AI technology do not have enough insights into the interior operations to comprehend the manner in which they have reached their output. As a result, this makes it challenging for humans to trust in advancements, learn from it and be capable of correcting the predictions of the upcoming situations.…”

Section: Literature Reviewmentioning

confidence: 99%

Artificial Intelligence for Smart Systems Critical Analysis of the Human Centered Approach

2021

JCNS

View full text Add to dashboard Cite

A program for Artificial Intelligence (AI) is knowledge as intelligent agent, which typically interacts with the ecosystem. This agent is capable of identifying the status of the ecosystem using the sensors before affecting the state via the actuators. We call the smart systems "agents” whenever they are able to make some decisions on their own with respect on particular goals. On the other hand, Machine Learning (ML) signifies a specific strategy meant to design smart systems whereby these systems can adapt to specific behaviors with respect to data. In the modern age, humans are rapidly collaborating with ML and AI systems. The AI that is human-based is a perspective of ML and AI, which algorithms have to be established with the awareness that they are a major segment of the massive system incorporating human. In this paper, we have presented a research that means that AI systems understand humans with respect to their socio-cultural aspects and that AI system assist humans comprehend them. We also present an argument of the challenges of social responsibility e.g. transparency, interpretability, accountability and fairness.

show abstract

“…The concept of a superimposed reaction skeleton graph was introduced by Yuri Kiho 16 and George Vladutz 17 and was later reinvented by Shinsaku Fujita as an "imaginary transition structure". 18 Gérard Kaufmann and co-workers [19][20][21] at the University of Strasbourg called it CGR and used it in reaction classification, reaction rules, and synthesis design. Varnek later used CGR to describe reactions in descriptor-based chemical space and to develop machine-learning models.…”

Section: Condensed Graph Of Reaction: An Efficient Approach To Reactions Miningmentioning

confidence: 99%

National Institutes of Health (NIH) Workshop on Reaction Informatics

Warr¹

2021

Preprint

View full text Add to dashboard Cite

The virtual workshop took place on May 18-20, 2021. It was a follow-up from the December 2020 NIH Workshop on Ultra Large Chemistry Databases. A major theme emerging from the December 2020 workshop was the fact that all the databases of a billion or more structures are virtual. For each virtual molecule the question then arises of whether, or how, it can be synthesized. The organizers therefore assembled speakers to give presentations about how reaction-related data are represented, captured, managed in databases, analyzed, used for drug design, applied in robotics, and exchanged locally as well as globally. This report summarizes talks from 27 practitioners in the reaction informatics field. The aim is to represent as accurately as possible the information that was delivered by the speakers; the report does not seek to be evaluative. The themes, in the order used for this report, were reaction representations, file formats, and standards; sources of reaction data; AI and machine learning applications of reaction-related data in de novo drug design, synthetic accessibility, synthesis planning, reaction prediction etc.; and automation and progression toward autonomous synthesis.

show abstract

Machine learning of generic reactions: 1. Scope of the project; the GRAMS program

Cited by 11 publications

References 3 publications

Graph-Mining Algorithm for the Evaluation of Bond Formability

Graph-Mining Algorithm for the Evaluation of Bond Formability

Artificial Intelligence for Smart Systems Critical Analysis of the Human Centered Approach

National Institutes of Health (NIH) Workshop on Reaction Informatics

Contact Info

Product

Resources

About