Machine Learning - Support Vector Machines

Entertaining talk. Great examples. Very good explanation of the underlying theory, and even some very good advice.

Thanks Sjoerd, great talk!

Very interested talk with good examples. The only thing I missed was the answer to why, for your example problems, a support vector machine is the best / a good approach in favor of other ml algorithms.

Very live talk, kept attention for the whole time, interesting concept and examples.

Great talk with interesting examples.

Good talk!

Interesting talk Sjoerd, well done :) Basic ML concepts were explained from the ground, examples were really, really good. My only proposition is about changing Alice in Wonderland case to live-demo. It would be really great to see how data could be classified by PHP in realtime :)

Thanks for your talk!

Oh, one more thing... I need to add one slide to my presentation, which will be inspired by you talk :)

Liked this talk very much, good sense of humour

Awesome talk, simple explanation of a complex subject.
More depth would've been nice but there was not enough time for that I think.

Some things that could be improved: explain the theory (as it states in the abstract); the current talk was more like "we have an SVM implementation, I put things in it, I get things out of it, I did machine learning". From the talk, it was not clear how SVM's even worked/how you could implement them, which was exactly what I was expecting after reading the abstract. It was clear, however, how you could use them.
The speaker seemed to simplify the process too much or did not have a deep understanding of the subject matter himself.

Really enjoyable talk, you are very capable of transferring your own enthousiasm on the subject.

The presentation was done well, but the talk lacked some depth to me. Mariusz Gil's talk on friday went a bit more clearly into the mechanics of the data prep and use of the learning machines.

Very well done. I liked the way Sjoerd explained a very complex subject in a way that everyone could understand.

Interesting topic presented well!

Very interesting and entertaining talk. Maybe OCR example should be a little bit shorter, but my overall impression is ok. Both ML talk on DPC were very useful, need to dig more into it.

Enthusiastic speaker. Great and detailed examples on the problems you attempted to solve.

I would have expected to see some more insight into the fundamental statistics involved. Some significant choices the speaker made in the data-prep phase can be controversial if done without any explanation, and an audience question afterwards reflected that.

Entertaining talk in how to achieve the 'next level' in machine learning with SVM and how heavily your solution depends on the preparation of your dataset.

Well built presentation, interesting subject and an interesting approach to the subject. The examples were well thought out and emphasised the basics greatly.