Today, I had the honor and pleasure to chair the International Workshop on Benchmarking of Computational Intelligence Algorithms (BOCIA) at the Tenth International Conference on Advanced Computational Intelligence (ICACI 2018), which took place from March 29 to 31, 2018 in Xiamen [[厦门]], Fujian [[福建省]], China. Our program was packed with seven very interesting presentations on the field, each of which discussing another aspect or application where algorithm performance was mined using modern statistics. Even after working on this field for several years, the ideas proposed today were still novel to me and I found the presentations quite inspiring. Indeed, each presentation led to an exciting discussion, so the workshop could indeed become a platform for the exchange of thoughts.

I want to wholeheartedly thank all of our authors for their research contributions and insightful presentations and the audience for their valuable comments and discussions. It should be mentioned that our workshop was the very last session at ICACI, yet was attended very well. This also speaks for the great organization of ICACI, whose chairs managed to attract several top-level keynote and tutorial presenters and an audience who deeply cares about research on Computational Intelligence. Of course, I would also like to thank my co-chairs and the our program committee, without whom organizing our workshop would have been impossible.

]]>The Black-Box-Optimization Benchmarking (BBOB) methodology introduced by the long-standing and successful BBOB-GECCO workshops series has become a well-established standard for benchmarking continuous optimization algorithms. The aim of this workshop is to develop a similar standard methodology for the benchmarking of black-box optimization algorithms for discrete and combinatorial domains. The goal of this first edition of the BB-DOB workshop series is to define a suitable set of benchmark functions for discrete and combinatorial optimization problems.

]]>The conference program of the Tenth International Conference on Advanced Computational Intelligence (ICACI 2018), where our International Workshop on Benchmarking of Computational Intelligence Algorithms (BOCIA) takes place, has been released (pdf). The paper presentation of our workshop will take place on March 31st, 2018 from 09:45 to 11:30 in room Guang-Yi [[广益厅]]. Each talk will have a total of 15 minutes, 12 of which are for presentation and 3 for questions and answers. All authors should bring their presentation either in PPT, PPTX or PDF format on a USB stick and hand them to the session chair before the session starts. This will ensure that all presentations can take place in a timely manner without too many delays caused by switching and plugging in laptops. We are looking forward to meet you in Xiamen [[厦门]], Fujian [[福建省]], China!

- Yiqin Zhang, Fenlin Liu, Hongyan Jia, Jicang Lu, and Chunfang Yang. "Optimization of Rich Model based on Fisher Criterion for Image Steganalysis" in
*Proceedings of the Tenth International Conference on Advanced Computational Intelligence (ICACI 2018),*March 29-31, 2018, Xiamen, Fujian, China, IEEE, ISBN: 978-1-5386-4362-4, pages 809—814 - Zimian Wei, Yawen Cui, Wenjing Yang, Yin Li, Xiaodong Yi, Huadong Dai, and Xiaotian Zhou. "A Research on Metric Learning in Computer Vision and Pattern Recognition" in
*Proceedings of the Tenth International Conference on Advanced Computational Intelligence (ICACI 2018),*March 29-31, 2018, Xiamen, Fujian, China, IEEE, ISBN: 978-1-5386-4362-4, pages 815—820 - Haisheng Sun, Chuang Liu, Rui Xu, and Huaping Chen. "An Improved Estimation of Distribution Algorithm for Cloud Computing Resource Scheduling" in
*Proceedings of the Tenth International Conference on Advanced Computational Intelligence (ICACI 2018),*March 29-31, 2018, Xiamen, Fujian, China, IEEE, ISBN: 978-1-5386-4362-4, pages 821—826 - Meng Zhao and Jinlong Li. "Tuning the Hyper-parameters of CMA-ES with Tree-structured Parzen Estimators" in
- Jialing Li and Fei Han. "A Hybrid Multi-swarm Particle Swarm Optimization with One-Dimensional Chaotic Search Strategy" in
- Rong Zhao, Yanpeng Qu, Ansheng Deng, and Reyer Zwiggelaar. "A Density-based Discretization Method With Inconsistency Evaluation" in
- Qi Qi, Thomas Weise, and Bin Li. "Optimization Algorithm Behavior Modeling: A Study on the Traveling Salesman Problem" in

Here you can find a short workshop report.

]]>The Consulate General of Germany in Shanghai supports German companies and researchers from the provinces Anhui, Jiangsu, and Zhejiang by organizing a *German Science Circle* which meets in Shanghai. These meetings are always very interesting and good opportunities for exchanging thoughts. Today, I attended such a meeting with the focus on the collaboration of Max Planck Institutes [MPG] with Chinese groups and researchers. Two top-level researchers, Professor Dr. Klaus Müllen, Director of the Max Planck Institute for Polymer Research in Mainz and Professor Dr. Markus Antonietti, Director of the Max Planck Institute for Colloids and Interfaces in Potsdam-Golm, gave their thoughts in impulse talks about "Joint Innovation and Technology". Here I want to summarize what I learned at this meeting and add a few thoughts.

On March 6 and 13, 2018, I had the chance to give two one-hour introduction lessons on German culture to the students of the Nanmen Primary School [[南门小学||http://www.hfnx.com/]] [[(森林公园校区)]]. Before going there, I wondered how and what should I teach elementary school children about German culture? What could they be interested to learn? This was an interesting challenge.

We started by finding out where Germany is. I wanted to put this in a context of something the school children can relate to, so we began a virtual journey to Germany in [[Hefei|en|wiki]] [[合肥||wiki]]. We looked at satellite images and "zoomed" farther and farther out, until other cities and the whole of the [[Anhui|en|wiki]] [[安徽||wiki]] province and then all of China and its neighboring countries became visible. We then flew over the globe until arriving in Europe, where we zoomed in again on Germany. We then compared the size and population of Germany to Anhui and China. Comparing with Anhui makes sense, because here the scales are quite similar: Germany is about 2.5 times as big there are eight German people for every seven Anhui people. China is 27 times as big as Germany and has 17.5 times as many people. We then learned how German people look like and what clothes they wear. We looked at photos of typical German people of various professions, ranging from scientists, elementary school children, bakers, police people, fire fighters, sports persons, cashiers in a super market, farmers, and fisher people. The children liked that we have many different hair colors, such as blond, brown, black, and red. Finally, we talked about Christmas and the traditions around it, including Christmas trees and Santa Claus. And then our first lesson was already over.

In the second lesson, we continued to look into the German holidays related to Children and discussed Silvester/New Year, Eastern, Fasching (the German carnival), the Children's day, Mother's day, and how birthdays are celebrated. Finally, we found out what German people eat: Different from Chinese, who like to warm/freshly cooked food for all three meals, German people usually warm only for lunch and cold food for breakfast and dinner. Then, we eat bread covered with butter on top of which we put either sausages or cheese. A typical German lunch often involves potatoes and, again, meat. After having discussed typical German dishes (and that Chinese often perceive them as too sour), the second lesson was over, too.

Teaching a class of primary school children about my culture was a very nice experience. The children were very curious, attentive, and asked interesting questions. Also I want to thank the teachers Mrs. Yang [[杨老师]] and Mrs. Cheng [[程老师]] for their support and for translating the class.

I also found that this primary school to be a very good example about the resources and efforts that China is investing in education. It does not only have a very nice and green campus, but also offers many extracurricular activities to the pupils, including, for instance, singing, dancing, painting, playing musical instruments — and the international culture class. And for each activity, there is a special, dedicated and nicely decorated room with the right equipment.

]]>Today, the Nobel laureat Prof. Mario J. Molina visited our Hefei University [[合肥学院]] and gave the talk "Global Sustainability and Climate Change: Science, Policy and Risks" in our library, just across our institute's building. Prof. Molina is one of the world's leading experts on climate and sustainability. He is professor at University of California, San Diego and the Center for Atmospheric Sciences at the Scripps Institution of Oceanography. He received a Nobel Prize in chemistry for his research on the impact of CFC on the ozon layer and was one of the consultants advising the Obama administration in climate issues. Today, he was awarded the honorary professor title of our university. Here, I want to summarize his highly interesting talk on the challenges we face when trying to achieve sustainability, i.e., an economy and way of life which does not threat the well-being of future generations.

]]>Today, we published the first version of a new `R`

package at github.com/thomasWeise/dataTransformeR for normalizing and transforming numerical data.

When we fit models to data, we often do not want to use the raw data as-is. Instead, we usually want to fit models to normalized or log-scaled versions of the data. If all data elements are in `[0,1]`

, this makes it easier to pick initial parameter values for models. If there are exponential relationships present in data sets, we may want to get rid of them by log-scaling the data. This means that, after the models have been fitted, we need to transform the model back by applying the inverse of the data transformation to the model.

This package uses our functionComposeR package to construct and apply such bijective transformations. The core of this package are the `Transformation`

and `TransformedData`

S4 classes and the routines to construct instances of them.

Today, we published the first version of a new `R`

package at github.com/thomasWeise/functionComposeR for composing and canonicalizing functions. When we combine functions in `R`

in the form of `g(f(x))`

, we have the problem that the result is rarely human readable. This results from two problems. The first problem is that variables inside the function are evaluated in the environment of the function and even if they are constants, they will remain as variables. Thus, when printing a function `f(x)`

, I may sometimes something like a `k*x`

inside, but may not know the value of `k`

, even though it may be perfectly known in the function's environment and a constant. The second problem is that this also applies to nested functions, so there may be something like `f=function(x) x+g(x)`

where `g`

is a well-defined function, but printing `f`

will not reveal the nature of `g`

. Both of these issues also make evaluating the functions slower, as we could resolve the variables to constants and inline the nested functions' bodies, but instead evaluate them as variables and function calls, respectively. With our new package, we try to solve all of these issues at once. We provide a tool for combining functions and one for canonicalizing functions, i.e., for resolving all resolve-able components of a function.

The pary secretary of our university, Prof. Dr. Jingmin CAI [[蔡敬民党委书记]] therefore visited the quarters of foreign exchange students and foreign staff to wish everyone a Happy New Year. We wholeheartedly join Prof. Cai in his wishes and hope that all our university's members and their families will have a happy, healthy, and successful new year!

Happy New Year!

]]>