Prediction and Interpretable Visualization of Synthetic Reactions Using Graph Convolutional Networks

2019 ◽  
Author(s):  
Shoichi Ishida ◽  
Kei Terayama ◽  
ryosuke kojima ◽  
Kiyosei Takasu ◽  
Yasushi Okuno
Keyword(s):  
Black Box ◽  
Data Driven ◽  
Research Groups ◽  
Reaction Prediction ◽  
Convolutional Networks ◽  
Learning Techniques ◽  
Synthetic Reactions ◽  
Retrosynthetic Analysis ◽  
Data Driven Approach ◽  
Extended Connectivity

<div>Recently, many research groups have been addressing data-driven approaches for reaction prediction and retrosynthetic analysis. Although the performances of the data-driven approach have progressed due to recent advances of machine learning and deep learning techniques, problems such as improving capability of reaction prediction and the black-box problem of neural networks still persist for practical use by chemists. To expand data-driven approaches to chemists, we focused on two challenges: improvement of reaction prediction and interpretability of the prediction. In this paper, we propose an interpretable prediction framework using Graph Convolutional Networks (GCN) for reaction prediction and Integrated Gradients (IGs) for visualization of contributions to the prediction to address these challenges. As a result, our model showed better performances than the approach using Extended-Connectivity Fingerprint (ECFP). Furthermore, IGs based visualization of the GCN prediction successfully highlighted reaction-related atoms.</div>

Prediction and Interpretable Visualization of Synthetic Reactions Using Graph Convolutional Networks

2019 ◽  
Author(s):  
Shoichi Ishida ◽  
Kei Terayama ◽  
ryosuke kojima ◽  
Kiyosei Takasu ◽  
Yasushi Okuno
Keyword(s):  
Black Box ◽  
Data Driven ◽  
Research Groups ◽  
Reaction Prediction ◽  
Convolutional Networks ◽  
Learning Techniques ◽  
Synthetic Reactions ◽  
Retrosynthetic Analysis ◽  
Data Driven Approach ◽  
Extended Connectivity

<div>Recently, many research groups have been addressing data-driven approaches for reaction prediction and retrosynthetic analysis. Although the performances of the data-driven approach have progressed due to recent advances of machine learning and deep learning techniques, problems such as improving capability of reaction prediction and the black-box problem of neural networks still persist for practical use by chemists. To expand data-driven approaches to chemists, we focused on two challenges: improvement of reaction prediction and interpretability of the prediction. In this paper, we propose an interpretable prediction framework using Graph Convolutional Networks (GCN) for reaction prediction and Integrated Gradients (IGs) for visualization of contributions to the prediction to address these challenges. As a result, our model showed better performances than the approach using Extended-Connectivity Fingerprint (ECFP). Furthermore, IGs based visualization of the GCN prediction successfully highlighted reaction-related atoms.</div>

scholarly journals EEGG: An analytic brain-computer interface algorithm

2021 ◽  
Author(s):  
Gang Liu ◽  
Jing Wang
Keyword(s):  
Brain Regions ◽  
Poor Quality ◽  
Classification Performance ◽  
Black Box ◽  
Data Driven ◽  
Left Hand ◽  
Convolutional Networks ◽  
Data Driven Approach ◽  
Traditional Sense ◽  
The Brain

<div><div><b>Objective.</b> In the traditional sense, the modeling approaches can be divided into white-box (physics-based), black-box (data-driven), and gray-box (the combination of physics-based and data-driven). Because the human brain is a black box itself, the EEG-BCI algorithm is generally a data-driven approach. It generates a black-box or gray-box (e.g., "Visualizing convolutional networks") model. However, one black- or gray-box cannot completely explain the brain. This paper presents the first analytic "white-box" EEG-BCI algorithm using Gang neurons (EEGG).</div><div><br></div><div><b>Approach.</b> Independent and interactive components of neurons or brain regions can fully describe the brain. This paper constructed a relationship frame based on the independent and interactive compositions for intention recognition and analysis using a novel dendrite module of Gang neuron. A total of 4,906 EEG data about motor imagery (MI) of left-hand and right-hand movements from 26 subjects were obtained from GigaDB. Firstly, this paper explored EEGG's classification performance according to cross-subject accuracy. Secondly, this paper transformed the EEGG model into a relation spectrum expressing independent and interactive components of brain regions. Then, the relation spectrum was verified through the known ERD/ERS phenomenon. Finally, this paper explored the previously unreachable further BCI-based analysis of brain.</div><div><br></div><div><b>Main results.</b> (1) EEGG was more robust than typical "CSP+" algorithms for the data of poor quality [AUC:0.825±0.074(EEGG)>0.745±0.094(CSP+LDA)/0.591±0.104(CSP+Bayes)/0.750±0.091(CSP+SVM), p<0.001]. (2) The relation spectrum showed the known ERD/ERS phenomenon. (3) Interestingly, EEGG showed that brain regions' interactive components put a brake on ERD/ERS effects for classification (p<0.001). This means that generating fine hand intention needs more centralized activation in the brain.</div><div><br></div><div><b>Significance.</b> EEGG decomposed the biological EEG-intention system of this paper into the relation spectrum inheriting Taylor series, rather than the fuzzy interpretation of outputs, which offers a novel frame for analysis of the brain.</div></div><div><p></p></div>

scholarly journals Data-driven Chemical Reaction Prediction and Retrosynthesis

2019 ◽  
Vol 73 (12) ◽  
pp. 997-1000
Author(s):  
Vishnu H Nair ◽  
Philippe Schwaller ◽  
Teodoro Laino
Keyword(s):  
Chemical Reaction ◽  
Data Availability ◽  
Data Driven ◽  
High Tech ◽  
Full Potential ◽  
Material Synthesis ◽  
Reaction Prediction ◽  
Strong Focus ◽  
Retrosynthetic Analysis ◽  
Synthetic Routes

The synthesis of organic compounds, which is central to many areas such as drug discovery, material synthesis and biomolecular chemistry, requires chemists to have years of knowledge and experience. The development of technologies with the potential to learn and support experts in the design of synthetic routes is a half-century-old challenge with an interesting revival in the last decade. In fact, the renewed interest in artificial intelligence (AI), driven mainly by data availability, is profoundly changing the landscape of computer-aided chemical reaction prediction and retrosynthetic analysis. In this article, we briefly review different approaches to predict forward reactions and retrosynthesis, with a strong focus on data-driven ones. While data-driven technologies still need to demonstrate their full potential compared to expert rule-based systems in synthetic chemistry, the acceleration experienced in the last decade is a convincing sign that where we use software today, there will be AI tomorrow. This revolution will help and empower bench chemists, driving the transformation of chemistry towards a high-tech business over the next decades.

scholarly journals Built Environment Factors Affecting Bike Sharing Ridership: Data-Driven Approach for Multiple Cities

2019 ◽  
Vol 2673 (12) ◽  
pp. 55-68 ◽  
Author(s):  
David Duran-Rodas ◽  
Emmanouil Chaniotakis ◽  
Constantinos Antoniou
Keyword(s):  
Built Environment ◽  
National Level ◽  
Data Driven ◽  
Machine Learning Techniques ◽  
Data Set ◽  
Factors Affecting ◽  
Environment Factors ◽  
Learning Techniques ◽  
Bike Sharing ◽  
Data Driven Approach

Identification of factors influencing ridership is necessary for policy-making, as well as, when examining transferability and aspects of performance and reliability. In this work, a data-driven method is formulated to correlate arrivals and departures of station-based bike sharing systems with built environment factors in multiple cities. Ridership data from stations of multiple cities are pooled in one data set regardless of their geographic boundaries. The method bundles the collection, analysis, and processing of data, as well as, the model’s estimation using statistical and machine learning techniques. The method was applied on a national level in six cities in Germany, and also on an international level in three cities in Europe and North America. The results suggest that the model’s performance did not depend on clustering cities by size but by the relative daily distribution of the rentals. Selected statistically significant factors were identified to vary temporally (e.g., nightclubs were significant during the night). The most influencing variables were related to the city population, distance to city center, leisure-related establishments, and transport-related infrastructure. This data-driven method can help as a support decision-making tool to implement or expand bike sharing systems.

scholarly journals AI-Driven Synthetic Route Design with Retrosynthesis Knowledge

2020 ◽  
Author(s):  
Shoichi Ishida ◽  
Kei Terayama ◽  
Ryosuke Kojima ◽  
Kiyosei Takasu ◽  
Yasushi Okuno
Keyword(s):  
Deep Neural Networks ◽  
Data Driven ◽  
Machine Learning Techniques ◽  
Synthetic Route ◽  
Learning Techniques ◽  
Route Design ◽  
Human Interventions ◽  
Synthesis Planning ◽  
Retrosynthetic Analysis ◽  
Synthetic Routes

<div>Computer-aided synthesis planning (CASP) aims to assist chemists in performing retrosynthetic analysis for which they exploit their experiments, intuition, and knowledge. Recent breakthroughs in machine learning techniques, including deep neural networks, have significantly improved data-driven synthetic route designs without human interventions. However, such CASP applications are yet to incorporate retrosynthesis knowledge sufficiently into their algorithms to reflect chemists' way of thinking flexibly. In this study, we developed a hybrid CASP application of data-driven techniques and various retrosynthesis knowledge called "ReTReK" that integrates the knowledge as adjustable parameters into an evaluation for promising search directions. Experimental results showed that ReTReK successfully searched synthetic routes based on the specified retrosynthesis knowledge, and the results indicated that the synthetic routes searched with the knowledge were preferred to those without knowledge. The concept of integrating retrosynthesis knowledge as adjustable parameters into data-driven CASP applications is expected to contribute to further their development and spread them to chemists widely. </div>

scholarly journals Colour and Texture Descriptors for Visual Recognition: A Historical Overview

2021 ◽  
Vol 7 (11) ◽  
pp. 245
Author(s):  
Francesco Bianconi ◽  
Antonio Fernández ◽  
Fabrizio Smeraldi ◽  
Giulia Pascoletti
Keyword(s):  
Deep Learning ◽  
Intelligent Systems ◽  
Visual Recognition ◽  
Early Years ◽  
Data Driven ◽  
Traditional Methods ◽  
Research Attention ◽  
Convolutional Networks ◽  
Advantages And Disadvantages ◽  
Data Driven Approach

Colour and texture are two perceptual stimuli that determine, to a great extent, the appearance of objects, materials and scenes. The ability to process texture and colour is a fundamental skill in humans as well as in animals; therefore, reproducing such capacity in artificial (`intelligent’) systems has attracted considerable research attention since the early 70s. Whereas the main approach to the problem was essentially theory-driven (`hand-crafted’) up to not long ago, in recent years the focus has moved towards data-driven solutions (deep learning). In this overview we retrace the key ideas and methods that have accompanied the evolution of colour and texture analysis over the last five decades, from the `early years’ to convolutional networks. Specifically, we review geometric, differential, statistical and rank-based approaches. Advantages and disadvantages of traditional methods vs. deep learning are also critically discussed, including a perspective on which traditional methods have already been subsumed by deep learning or would be feasible to integrate in a data-driven approach.

scholarly journals AI-Driven Synthetic Route Design with Retrosynthesis Knowledge

2020 ◽  
Author(s):  
Shoichi Ishida ◽  
Kei Terayama ◽  
Ryosuke Kojima ◽  
Kiyosei Takasu ◽  
Yasushi Okuno
Keyword(s):  
Deep Neural Networks ◽  
Data Driven ◽  
Machine Learning Techniques ◽  
Synthetic Route ◽  
Learning Techniques ◽  
Route Design ◽  
Human Interventions ◽  
Synthesis Planning ◽  
Retrosynthetic Analysis ◽  
Synthetic Routes

<div>Computer-aided synthesis planning (CASP) aims to assist chemists in performing retrosynthetic analysis for which they exploit their experiments, intuition, and knowledge. Recent breakthroughs in machine learning techniques, including deep neural networks, have significantly improved data-driven synthetic route designs without human interventions. However, such CASP applications are yet to incorporate retrosynthesis knowledge sufficiently into their algorithms to reflect chemists' way of thinking flexibly. In this study, we developed a hybrid CASP application of data-driven techniques and various retrosynthesis knowledge called "ReTReK" that integrates the knowledge as adjustable parameters into an evaluation for promising search directions. Experimental results showed that ReTReK successfully searched synthetic routes based on the specified retrosynthesis knowledge, and the results indicated that the synthetic routes searched with the knowledge were preferred to those without knowledge. The concept of integrating retrosynthesis knowledge as adjustable parameters into data-driven CASP applications is expected to contribute to further their development and spread them to chemists widely. </div>

Sign in / Sign up

Export Citation Format

Share Document