If you liked this video, get the Book on Evolutionary Algorithms (With Python Notebooks) datacrayon.com/shop/product/practical-evolutionary-algorithms-book/
Thank you - glad you've found it useful! After some delay I've just released the vide on on the Hypervolume Indicator ua-cam.com/video/cR4r1aNPBkQ/v-deo.html
Every time I find someone who genuienly knows about what they are speaking I really appreciate that 'cause it's rare, thank you so much for creating this content, waiting for more !!
This is everything that a video lecture must comprise of! I'm just sitting here in awe for the fusion of non-assumptive explanations, structuring of information, Goldilocks pace and reducing complex information to layman terms. You sir, got my respect.
Just ordered your book on Practical Evolution Algorithm. Your explanation of the evolutionary process of genetic algorithm is so simple and easy to understand. Thank you.
I'm doing my PhD in Computer vision and trying to improve a recently proposed Evolutionary algorithm in (GANs), which is my current research area. The article was published recently in IEEE transactions on Evolutionary Computation. i'm trying to build a strong intuition for evolutionary computation and your introductory video has helped me a lot, Thank you so much Doc!!!!
I see the whole system frontloaded with specified complexity and creative intelligence and then evolution acting on that existing information. My point is that it still starts with creative intelligence with evolutionary algorithms programmed in to the system. Very facinating, smart and advanced work, but to me that testifies of an external intelligence encoding the system to run as it should but over time, when the computer system is left to itself, then still transrciption errors through virusses might creep in and causes the overall system to malfunction.
well, sure, but in a computer simulation everything of information is made by intelligence from the start, and still we can not always see what the answer would be, and also, in a biological world, the information would be there, since matter, energy, and chemistry exist
Great illustration. Given the wide variety of applicability of GA at times it’s become very difficult to encode the population. I would like you to give a similar example in terms of social network influence maximizations.
Excellent presentation...It is so explanatory and useful for someone who want to get into that field! The introduction that everyone would find useful for evolutionary algortithms :) Keep up the good work!
Excellent introduction. Subscribed, looking forward to more follow-up videos. I love that you added further reading resources if you make more videos please keep that. As for the follow along part, maybe you could ask those people to pause the video and work it out, or alternatively include an annotation for those that want to skip it.
Shouldn't we call this Micro-evolution Algorithm or Adaptation Algorithm? It allows living forms to change for survival, it's called adaptation. This has been proven by experiments both in the lab (fruit flies) and found in nature (finches). What about macro-evolution? Issues: Where did the initial population come from? Who or what decided their original characteristics? How is the characteristic information pass down? Who or what decides when a species is completed or finished?
@@ShahinRostamiI'm currently in my final year of my computer science undergrad and my final year project is on 'Evolutionary Reinforcement Learning', neither of which I knew much about before starting. This video really helped me get started on understanding the evolutionary algorithms side of things!
Thanks Dr. Rostami. I'm going through the book on Reinforcement Learning by Sutton and a little knowledge of evolutionary methods is highly resourceful in its reading
This is an excellent video. I'm working on an evolutionary algorithm to produce a process model that reflects a "real process". Big help to me so far. Any chance you could put up perhaps a walk-through of some actual code for intermediate-level coders?
Hello Stan - thank you for your interest in the video and I'm glad to see you're working within the area. Talking through some code may be difficult because there are so many evolutionary algorithms out there, not to mention the multiple implementations of these algorithms in different languages. I can recommend some frameworks which have many of these implemented - i.e. jMetal (Java) or PlatEMO (MATLAB). I hope this helps
Congratulations and thanks for posting. Very very very good. One small detail, your population of cartoons has not 6 but 7 parameters, as head width and length vary independently.
Thank you kind comment. You are absolutely right - I didn't separate them to keep the example simple, and I wanted to make one of their heads wide to demonstrate an undesirable mutation :)
Hey, I want to make an essay on things like holding a religious faith while accepting genetic algorithms that work towards things like aerodynamic optimization. Because the main problem is that you can’t hold a belief and be a scientist in the sense of applying evolutionary algorithms. I just want a simple reply.
Ive been struggling on working on my own algorithm. It technically evolves until the desired fitness but the way the mutations occur is too inefficient. What would you recommend looking at for inspiration?
Hello - happy to have a look at your approach and make a recommendation. Send me a description of the problem using the contact form on www.shahinrostami.com and I'll get back to you soon!
1st time learning about GA. I think the speed is about right. Color coding of body parts makes explanation of cross over algorithm clear. it would be nice to have extra text when new terms are mentioned like cross over or mutation although they are common terms. thanks for the video, hope to see more! Subbed.
Thanks, great explanation, would love to see a more realistic example. Does such evalutionary algo's have overlapping use cases with deep learning methods?
Hello - thank you for the kind comment. If you're interested in more realistic examples I would recommend reading up on multi-objective optimisation (I have a recent video on the subject). There are some overlaps between EA's and Deep Learning techniques. EA's have been used to train neural networks for a while, with the benefit of being able to optimise the structure (number of hidden layers/neurons per layer) as well as the weights/biases.
I am not sure how to say this but.. although your computational work is clear and insightful your assumption that these models which demonstrate how the algorithm of evolution works (agreed) also verifies biological evolution of life is incorrect, for several reasons. Would you or anyone here like to know why this is the case?
Shouldn't the chromosome crossover Variation stage occur *before the Variation stage as it does in nature through organisms different genders being defined before they are sexualy mature?
Hello Paul. Each Evolutionary Algorithm (EA) uses its own distinctive set of operators for variation. Those which fall under the area of Genetic Algorithms typically include some type of recombination (or crossover) operator as part of that set. I hope this helps and that the video was useful.
Hello Paul, the selection stage may involve one or more selection operators which all work differently. For the demonstration in the video I chose to keep it simple and went with keeping the best half of the population as parents for the next generation. This actually introduces elitism too, and the size of your selected parent population can introduce different selection pressures.
If you liked this video, get the Book on Evolutionary Algorithms (With Python Notebooks) datacrayon.com/shop/product/practical-evolutionary-algorithms-book/
A shame that such an impressive introduction to a rather complex field has so few views. Keep it up, great work!
I'm glad you enjoyed it - I will be recording another video soon
AGREED!
Thank you - glad you've found it useful! After some delay I've just released the vide on on the Hypervolume Indicator ua-cam.com/video/cR4r1aNPBkQ/v-deo.html
The simplicity of the illustrated example makes it simple to digest the concepts. Thanks Shahin.
That's an amazing introduction ..thank you so much !
Thank you Ammar, I'm glad you found it useful.
Every time I find someone who genuienly knows about what they are speaking I really appreciate that 'cause it's rare, thank you so much for creating this content, waiting for more !!
Very straight forward and yet deep. It's intersting to see how darwin's theory of evolution proves itself even in computer science.
As a first year, I'm really looking forward to your unit in the second Semester! This makes me look forward to it even more!
This is everything that a video lecture must comprise of! I'm just sitting here in awe for the fusion of non-assumptive explanations, structuring of information, Goldilocks pace and reducing complex information to layman terms. You sir, got my respect.
You have a FANTASTIC speaking / presenting voice. No "adjustment" to the format needed - you nailed it!
This is one of the best and simplest way of describing EAs. Thank you professor for such a nice introduction
Just ordered your book on Practical Evolution Algorithm. Your explanation of the evolutionary process of genetic algorithm is so simple and easy to understand. Thank you.
It’s my pleasure, I’m glad you found it useful!
This is best introductory tutorial for Evolutionary algorithm. Thank you Dr. Rostami.
I'm doing my PhD in Computer vision and trying to improve a recently proposed Evolutionary algorithm in (GANs), which is my current research area. The article was published recently in IEEE transactions on Evolutionary Computation. i'm trying to build a strong intuition for evolutionary computation and your introductory video has helped me a lot, Thank you so much Doc!!!!
@Dr.Shahin Rostami i would be grateful and it would be of great help, if we can communicate via e-mail for the purpose of sharing ideas.
Thank you sir ❤
Thanks, Shahin, this is a well-detailed presentation.
Respected Professor,
Thanks a lot for explaining such a difficult concept in such a simple manner.
kheyli informative bood bia nazet konam :)
Wow, the best video ever on introduction to evolutionary computation. I am so happy to come across this.
Your gentle, paced and plain-language introduction is a real help for beginners looking to get into this fascinating field :)
This one was really helpful for my thesis. Thank you!
You're very welcome!
it was very easy for me to understand the concepts. best explanation I've seen so far, thank you!!
Professor, I appreciate your extremely clear explanation. This will be highly useful in my research.
This video is absolutely incredible. I appreciate the simple way you approached the complex topic! Keep up the incredible videos.
This video was really helpful. You should upload more videos.
Very Impressive and interesting topic very helpful,Thanks
Very good introduction on evolutionary algorithms. I am also a BIG fan of evolutionary algorithms (i.e. genetic algorithm). Many thanks Dr.
This video is really good for understanding the concept of evolutionary algorithm. Thank you
east to understand, thanks for making this
I see the whole system frontloaded with specified complexity and creative intelligence and then evolution acting on that existing information. My point is that it still starts with creative intelligence with evolutionary algorithms programmed in to the system. Very facinating, smart and advanced work, but to me that testifies of an external intelligence encoding the system to run as it should but over time, when the computer system is left to itself, then still transrciption errors through virusses might creep in and causes the overall system to malfunction.
well, sure, but in a computer simulation everything of information is made by intelligence from the start, and still we can not always see what the answer would be, and also, in a biological world, the information would be there, since matter, energy, and chemistry exist
Great job! Thanks for putting this together. If my paper gets accept to a conference, I will make sure I include you in the acknowledgements.
I’m glad it was useful! Let me know how it works out, I would love to see it
Kindly make a video on adjoint based computations .
perfect description (y)
but views......
you saved my hours of research on this topic.
thanks alot.
Thank you. I'm glad you found it helpful
absolutely brilliant explanation...what a fascinating field, I will endeavour to know more :).
Excellent and to the point!
If possible please give such classic explanation on Differential Evaluation
Great illustration. Given the wide variety of applicability of GA at times it’s become very difficult to encode the population. I would like you to give a similar example in terms of social network influence maximizations.
Perfect explanation thank you.
Excellent presentation...It is so explanatory and useful for someone who want to get into that field! The introduction that everyone would find useful for evolutionary algortithms :) Keep up the good work!
Glad you found it useful!
Move 37 brought me here!! Great Explanation
Looking forward for more great content from you specifically Neural architecture estimation and QC
Interesting demonstration
great explanation , simple and easy to follow thanks a lot
Excellent video.
Great work sir....sir can you help me to understand the any algorithm matlab coding
Excellent introduction. Subscribed, looking forward to more follow-up videos. I love that you added further reading resources if you make more videos please keep that. As for the follow along part, maybe you could ask those people to pause the video and work it out, or alternatively include an annotation for those that want to skip it.
Thank you for the feedback and the idea. I've included an annotation which skips the 15 second activity time.
Shouldn't we call this Micro-evolution Algorithm or Adaptation Algorithm? It allows living forms to change for survival, it's called adaptation. This has been proven by experiments both in the lab (fruit flies) and found in nature (finches). What about macro-evolution?
Issues: Where did the initial population come from? Who or what decided their original characteristics? How is the characteristic information pass down? Who or what decides when a species is completed or finished?
Really great introduction!
Glad you liked it!
@@ShahinRostamiI'm currently in my final year of my computer science undergrad and my final year project is on 'Evolutionary Reinforcement Learning', neither of which I knew much about before starting. This video really helped me get started on understanding the evolutionary algorithms side of things!
Excellent explanation. Thank you.
This was really helpful! You ignited a curiosity in me to explore further and that's exactly what a good introduction is meant to do.
I'm glad it was useful - good luck on your exploration
Thanks Dr. Rostami. I'm going through the book on Reinforcement Learning by Sutton and a little knowledge of evolutionary methods is highly resourceful in its reading
This was very helpful.
what an amazing video. Thanks a lot.waiting for new videos about Evolutionary Algorithms.
much love!
Thank you for the amazing videos
Glad you like them!
Amazing video for beginners, thx very much!!
This is an excellent video. I'm working on an evolutionary algorithm to produce a process model that reflects a "real process".
Big help to me so far. Any chance you could put up perhaps a walk-through of some actual code for intermediate-level coders?
Hello Stan - thank you for your interest in the video and I'm glad to see you're working within the area. Talking through some code may be difficult because there are so many evolutionary algorithms out there, not to mention the multiple implementations of these algorithms in different languages. I can recommend some frameworks which have many of these implemented - i.e. jMetal (Java) or PlatEMO (MATLAB). I hope this helps
Thanks for this wonderful video! Got to learn a lot!
Honestly, you nailed it. Thank you very much for this information. Greetings from Argentina
Sir this video was very good and really helped me understand this (I'm using the algorithm NEAT that does this with neural networks). Subscribed
Glad it helped!
wow, thank you for this great explanation of evolutionary algorithm
As always, fantastic presentation. Great job and thank you.
Great explanation!
Even the introduction learns
Congratulations and thanks for posting. Very very very good. One small detail, your population of cartoons has not 6 but 7 parameters, as head width and length vary independently.
Thank you kind comment. You are absolutely right - I didn't separate them to keep the example simple, and I wanted to make one of their heads wide to demonstrate an undesirable mutation :)
Thank you.Great Explanation. Looking forward for more videos.
Superb, Good introduction with good examples really helpful for me to understand! Great work
thank you very much
so well explained!
Thank you Jessaya
Hey, I want to make an essay on things like holding a religious faith while accepting genetic algorithms that work towards things like aerodynamic optimization. Because the main problem is that you can’t hold a belief and be a scientist in the sense of applying evolutionary algorithms. I just want a simple reply.
thank you very much, it helped me to understand the concept better :)
This is an amazing and perfect explanation Sir. I would like to watch more videos. Keep it up Sir.
Thanks for the wonderful video. If possible please make a video on Non-dominated Sorting Genetic Algorithm (NSGA)
incredible
nice introduction :) thank you Dr Shahin
Thank you
its awesome! pls make some more on evolutionary computation
Can i ask you about how to link matlab with a hydrological model SWAT?
Thank you ! I am currently studying the subject and it seemed abstract to me. It's now better :D
I'm glad you found it helpful!
Great enjoyable presentation.
Great work!
amazing introduction , helped alot thx u r brilliant
Thanks , great work!
This made my head hurt less than a 2 hour lecture
Happy to hear it!
Ive been struggling on working on my own algorithm. It technically evolves until the desired fitness but the way the mutations occur is too inefficient. What would you recommend looking at for inspiration?
Hello - happy to have a look at your approach and make a recommendation. Send me a description of the problem using the contact form on www.shahinrostami.com and I'll get back to you soon!
awesome explanation
I am very very thankful for that interesting tutorial. It made things easier than I expected. Keep forward :).
1st time learning about GA. I think the speed is about right. Color coding of body parts makes explanation of cross over algorithm clear. it would be nice to have extra text when new terms are mentioned like cross over or mutation although they are common terms. thanks for the video, hope to see more! Subbed.
Thank you for the feedback - I'll look to include text definitions of esoteric words in future videos. I'm glad you found the video useful.
Thanks, great explanation, would love to see a more realistic example.
Does such evalutionary algo's have overlapping use cases with deep learning methods?
Hello - thank you for the kind comment. If you're interested in more realistic examples I would recommend reading up on multi-objective optimisation (I have a recent video on the subject). There are some overlaps between EA's and Deep Learning techniques. EA's have been used to train neural networks for a while, with the benefit of being able to optimise the structure (number of hidden layers/neurons per layer) as well as the weights/biases.
Keep it up, great work!
This is just perfect!! Thanks :)
Great visualization for such an interesting topic, thanks for the help and good luck!
Cheers :)
TQ sir,TQ very much
thank you very much !
My pleasure!
can i get code of DE in Matlab or Python
Thanks Informative vdio
I am not sure how to say this but.. although your computational work is clear and insightful your assumption that these models which demonstrate how the algorithm of evolution works (agreed) also verifies biological evolution of life is incorrect, for several reasons.
Would you or anyone here like to know why this is the case?
Mike Bellamy simple, because you believe in magical sky fairy, and evolution contradicts that believe;)
@@tgstudio85 No biological evolution contradicts the laws of thermodynamics and information theory..
Great video ..thank you 👍
Nice work!
Great explanation!
tnx a lot!
Shouldn't the chromosome crossover Variation stage occur *before the Variation stage as it does in nature through organisms different genders being defined before they are sexualy mature?
Hello Paul. Each Evolutionary Algorithm (EA) uses its own distinctive set of operators for variation. Those which fall under the area of Genetic Algorithms typically include some type of recombination (or crossover) operator as part of that set. I hope this helps and that the video was useful.
During the selection stage, why did you typically pick the top three and not just the top two?
Hello Paul, the selection stage may involve one or more selection operators which all work differently. For the demonstration in the video I chose to keep it simple and went with keeping the best half of the population as parents for the next generation. This actually introduces elitism too, and the size of your selected parent population can introduce different selection pressures.
I would love a video on the benefits of using elitism with Evolutionary Algorithms - Thank you for these uploads, they are a great help.
Nice explanation and good visual example (although the figures do not look too nice 😉)
Thank you - unfortunately stick figures are all I can draw!
Thank u, that`s awesome !
Thank you.
very good work.....