On the Complementary Role of Data Assimilation and Machine Learning: An Example Derived from Air Quality Analysis

In artificial intelligence, abstraction is commonly used to account for the use of various levels of details in a given representation language or the ability to change from one level to another while preserving useful properties. Abstraction has been mainly studied in problem solving, theorem proving, knowledge representation (in particular for spatial and temporal reasoning) and machine learning. In such contexts, abstraction is defined as a mapping between formalisms that reduces the computational complexity of the task at stake. By analysing the notion of abstraction from an information quantity point of view, we pinpoint the differences and the complementary role of reformulation and abstraction in any representation change. We contribute to extending the existing semantic theories of abstraction to be grounded on perception, where the notion of information quantity is easier to characterize formally. In the author's view, abstraction is best represented using abstraction operators, as they provide semantics for classifying different abstractions and support the automation of representation changes. The usefulness of a grounded theory of abstraction in the cartography domain is illustrated. Finally, the importance of explicitly representing abstraction for designing more autonomous and adaptive systems is discussed.

Download Full-text

Urban Air Quality Analysis and Prediction Using Machine Learning

2019 1st International Conference on Advanced Technologies in Intelligent Control, Environment, Computing & Communication Engineering (ICATIECE) ◽

10.1109/icatiece45860.2019.9063845 ◽

2019 ◽

Author(s):

K. Nandini ◽

G. Fathima

Keyword(s):

Machine Learning ◽

Air Quality ◽

Urban Air Quality ◽

Quality Analysis ◽

Urban Air

Download Full-text

What Can We Expect from Data Assimilation for Air Quality Forecast? Part I: Quantification with Academic Test Cases

Journal of Atmospheric and Oceanic Technology ◽

10.1175/jtech-d-18-0002.1 ◽

2019 ◽

Vol 36 (2) ◽

pp. 269-279 ◽

Cited By ~ 1

Author(s):

Laurent Menut ◽

Bertrand Bessagnet

Keyword(s):

Air Quality ◽

Data Assimilation ◽

Nitrogen Dioxide ◽

Atmospheric Composition ◽

Quality Analysis ◽

Test Cases ◽

Ozone Formation ◽

Air Quality Forecast ◽

Maximum Benefit ◽

Academic Test

Abstract Data assimilation has been successfully used for meteorology for many years and is now used more and more for atmospheric composition issues (air quality analysis and forecast). The data assimilation of pollutants remains difficult and its deployment is currently in progress. It is thus difficult to have quantitative knowledge of what we can expect as the maximum benefit. In this study we propose a simple framework to make this quantification. In this first part, the gain of data assimilation is quantified using academic but realistic test cases over an urbanized polluted area and during a summertime period favorable to ozone formation. Different data assimilation configurations are tested, corresponding to different amounts of data available for assimilation. For ozone (O3) and nitrogen dioxide (NO2), it is shown that the benefit resulting from data assimilation lasts from a few hours to a possible maximum of 60 and 21 h, respectively. Maps of the number of hours are presented, spatializing the benefit of data assimilation.

Download Full-text

Air Quality Forecast through Integrated Data Assimilation and Machine Learning

Proceedings of the 11th International Conference on Agents and Artificial Intelligence ◽

10.5220/0007555207870793 ◽

2019 ◽

Cited By ~ 1

Author(s):

Hai Lin ◽

Jianbing Jin ◽

Jaap van den Herik

Keyword(s):

Machine Learning ◽

Air Quality ◽

Data Assimilation ◽

Air Quality Forecast ◽

Quality Forecast

Download Full-text

Exchange Spin Coupling from Gaussian Process Regression

10.26434/chemrxiv.12589541.v3 ◽

2020 ◽

Author(s):

Marc Philipp Bahlke ◽

Natnael Mogos ◽

Jonny Proppe ◽

Carmen Herrmann

Keyword(s):

Machine Learning ◽

Gaussian Process ◽

Gaussian Process Regression ◽

Molecular Magnets ◽

Molecular Structures ◽

Spin Coupling ◽

Structure Property ◽

Data Set ◽

Uncertainty Estimates

Heisenberg exchange spin coupling between metal centers is essential for describing and understanding the electronic structure of many molecular catalysts, metalloenzymes, and molecular magnets for potential application in information technology. We explore the machine-learnability of exchange spin coupling, which has not been studied yet. We employ Gaussian process regression since it can potentially deal with small training sets (as likely associated with the rather complex molecular structures required for exploring spin coupling) and since it provides uncertainty estimates (“error bars”) along with predicted values. We compare a range of descriptors and kernels for 257 small dicopper complexes and find that a simple descriptor based on chemical intuition, consisting only of copper-bridge angles and copper-copper distances, clearly outperforms several more sophisticated descriptors when it comes to extrapolating towards larger experimentally relevant complexes. Exchange spin coupling is similarly easy to learn as the polarizability, while learning dipole moments is much harder. The strength of the sophisticated descriptors lies in their ability to linearize structure-property relationships, to the point that a simple linear ridge regression performs just as well as the kernel-based machine-learning model for our small dicopper data set. The superior extrapolation performance of the simple descriptor is unique to exchange spin coupling, reinforcing the crucial role of choosing a suitable descriptor, and highlighting the interesting question of the role of chemical intuition vs. systematic or automated selection of features for machine learning in chemistry and material science.

Download Full-text