ARTIFICIAL NEURONAL NETWORKS: A DEEPER INSIGHT INTO IONIC LIQUIDS


DIRECTORS: Kenneth R. Seddon and José S. Torrecilla

 

OBJECTIVES:

  • Detailed description of the main features and progress within the ionic liquid field.
  • Explanation of the principal characteristics of artificial neural networks.
  • Design of computerized models based on artificial neural networks using experimental data.
  • Comprehension of the main differences between the current models in the ionic liquid field.

SUMMARY:

The course will offer a general view of the combination of two interesting research fields within the international scientific community. First of all, ionic liquids represent a rapidly expanding chemical market with great interest in research due to the great number of new synthesized compounds and novel applications. On the other hand, computerized artificial intelligence embodies a group of tools that are suitable for supporting a wide variety of research areas.

The novelty and high progress rate of the ionic liquid field implies a great amount of difficulty to understand their chemical behaviour and calculate their physicochemical properties. Both these aspects heavily limit the design of useful new ionic liquid-related applications, and, therefore, it should be resolved. This is the reason why many research groups which work with ionic liquids measure their physicochemical properties and their variation with the surrounding conditions. Despite of all this work, there is still a lack of knowledge in this regard that can be avoided to a great extent with mathematical tools. To do so, mathematical models which can accurately define the behaviour of ionic liquids, as well as estimate their properties, are required. And going a bit further, these models could be trained to estimate properties for hypothetical ionic liquids, yet to be synthesised, to assist in their chemical design.

Currently, there are predictive models based on mathematical algorithms that are employed to describe the nature of various compounds (covalent, salts, and others). These models can be applied to ionic liquids in some cases, offering acceptable results. Nevertheless, the main problem behind these systems is that they are not designed purposely for ionic liquids. On the other hand, we count with models that can be designed specifically for ionic liquids, and that in some situations use experimental data. This is the case of artificial neural networks. Given the excellent results attained when combining ionic liquids with neural networks, these models have been developed in numerous fields and to fulfil dissimilar tasks such as property estimation, contamination assessment, ionic liquid design, chemometric tool creation, and so on.

During this course, the reason why applications involving ionic liquids are increasing on a daily basis will be explained, together with the main differences they have versus other chemical compounds. Additionally, the mathematical tools currently employed will be studied and compared with intelligent models. The main characteristics and advantages that artificial neural networks provide against other empirical algorithms will also be analysed. Finally, to get the most out of the neural network lectures, the students will design their own artificial neural network utilizing a real ionic liquid-related database. This way, the main goal of the course will be fulfilled, which is learning how to design an artificial neural network for an ionic liquid application.

 

REGISTRATION:
 
If you are interested to attend to our course click here     

 


 


 


 


 


 

Prices

  • Student (UCM or Queens University)  40 Euros
  • Student (other Universities) 100 Euros
  • Professors, Researchers and rest of attendees 160 Euros

Registration form

   Click here 

 

Programme

Click here

Other information

Travel agency

Tourist information