Hybrid Modeling Workshop (July 1st - July 3rd, 2020)

The Novasign hybrid modeling workshop 2020 will take place from July 1st to July 3rd at the University of Natural Resources and Life Sciences (BOKU) in Vienna. During these three days, an intense workshop will cover current topics about hybrid modeling, such as a general introduction, the possibilities and benefits of using hybrid models and their practical application. For the whole course, the latest version of the Novasign hybrid modeling toolbox will be used to get in touch with all the provided features.

What can you expect?

Who should attend?

The workshop is a great opportunity for everyone who wants to first get in touch with the subject or to educate oneself further. Basic knowledge and a modeling background are not mandatory. The provided program is suited for every level of knowledge, regardless of being a beginner in (hybrid) modeling or already possessing in-depth knowledge. Therefore, this workshop is directed and recommended for everyone, e.g., (PhD) students, senior scientists in academia and the industry, who are interested in modeling to enhance the efficiency of their processes and to those who want to train themself in a growing and highly beneficial field to enhance their market value.

What is hybrid modeling?

In general, there are two ways of modeling, called parametric (white box) and non-parametric (black box) modeling, hereafter referred to as white-box and black-box. Black-box models refer to purely data-driven approaches, for which no further process knowledge is needed and the model structure is inferred from data, e.g., artificial neural networks. However, once the model is used for predictions out of the characterized space, it lacks the ability to extrapolate. On the other side, the structure of a parametric model is based on knowledge and empirical considerations and therefore is a more suitable match for extrapolation. Since this type of model solely assume a mechanistic trend and does not account for empiric observations, also the predictability often ends up inaccurately. 

However, Semi-parametric (hybrid) models allow for the integration of knowledge from different domains, e.g. fundamental process knowledge and process data. Hereby, it is possible to benefit from the positive aspects and make up for the respective drawbacks. Such a model can either combine the black and white box parts in a parallel or serial structure. The difference is the order in which the different parts and their respective weights are taken into account in the overall model. These features generally make it a more cost-effective modeling approach to deal with complex problems, e.g., the black-box can be used to estimate the known-unknowns used in the white-box. As a result, the behavior and deviations of, i.e., a process and temporary deviations can be understood.

These features enable hybrid models to be well-suitable and highly interesting candidates in the QbD and PAT concept for scaling and bioprocess characterization studies as well as development and optimization.

Which topics will be covered?

This comprehensive workshop covers present and relevant topics one has to know to understand the fundamentals and advanced subjects of hybrid modeling as well as its practical application. Starting with general questions, e.g., 'What is a hybrid model?'  'What are the advantages?' 'How can I generate a hybrid model?', the course will rapidly proceed to more specific topics such as 'Are there ways to make my hybrid model more robust?' and 'How can I enhance the performance and model quality?'.

In addition, useful tips answering all these questions to develop high performing hybrid models will be provided during these lectures and also during the hands-on part. The hands-on part will take up the most time of the workshop and will give the participants the opportunity to directly apply and understand the learned content in more detail. Therefore, the Novasign hybrid modeling toolbox will be used with in-house generated data to provide showcases for both, upstream and downstream processing. First, toolbox basics, i.e., data import, visualization and plotting will be explained followed by the showcases.

The upstream showcase will cover hybrid modeling on a completely characterized three-dimensional design space utilizing a comprehensive 20L Escherichia coli fed-batch study.

The downstream showcase will be about TFF.

With these, current topics such as process characterization, development and optimization according to the QbD and PAT principles will be understood and accessible for basic and industrial applications. This broad range of topics provides the participants the needed knowledge to apply hybrid models on their own in their respective fields after the end of the Novasign hybrid modeling workshop.

Who is joining?

By now the confirmed participants comprise people from: ... List of Universities, Companies,...

 

the lectures will be held by international specialists which also visited the well-established hybrid modeling summer school promoted by ESBES. This setting provides the participants with the optimal opportunity to get in contact with some of the most well-known experts in process development, characterization, optimization and hybrid modeling from both, industry and academia. The larger and practical hands-on part of the workshop will be held, guided and tutored by Novasign to provide the best support to get to know the toolbox and all implied features.

During the workshop and the dinner events, there will be enough time to get in contact with all these people to extend the personal network and to further have the opportunity to get access to the powerful and generic Novasign hybrid modeling toolbox.

Further opportunities around the workshop

The organized dinner events of the first two workshop days already allow short guided trips through the beautiful city of Vienna to collect some impressions. Especially in summer, the city has many (outdoor) activities to offer with its high amount of museums, green areas and due to its location next to the Danube and Donauinsel. Most of these are for free and a great opportunity to visit if the trip to our workshop is prolonged for the weekend. It is not pure coincidence that Vienna was voted the most livable city for ten times in a row, but we recommend to experience this on your own.

In addition, during the hands-on part of the workshop, there will be the 'Novasign hybrid modeling challenge' at which every participant can attend. The goal will be to use the gained knowledge and skills to develop a hybrid model with the provided data sets. All participants can type the performance of their best hybrid model in a sheet, in which they will be ranked,  to have the opportunity to win the following prizes:

1st place: 1x Bose Bluetooth Earphones

2nd place: 1x Fitbit Smartwatch

3rd place: 1x 100€ Amazon Voucher

Program of the Workshop

July 1st - Day 1 (Conference Room 12)

July 2nd - Day 2 (Conference Room 12)

July 3rd - Day 3 (Conference Room 12)

Registration

Registration Fees

Super Early Bird*

Industry

Academia

1299€ (5 left)                  689€ (5 left)                    449€ (5 left)

Early Bird**

    1399€                                789€                               499€

Regular

    1499€                               889€                               549€

Students 

* first registrations

** until April 4th, 2020

Registration deadline: June 5th, 2020 

Important Information

The entire workshop will take place in the conference rooms 12 and 13 (red dots) at the University of Natural Resources and Life Sciences (BOKU)

Adress: Muthgasse 11, 1190, Vienna, Austria

  • All participants are encouraged to bring their own laptop

  • For the hands-on part of the workshop, the Novasign hybrid modeling toolbox will be used and therefore installed in the participants' personal computers

  • Note: The Novasign hybrid modeling toolbox

    • runs only in Windows

    • It is free of charge!

  • a limited number of Windows computers can be provided if necessary email to benjamin.bayer@boku.ac.at

Partners and Sponsors