March 2017 - Pharo-users - lists.pharo.org

Re: [Pharo-users] Class versions

by Dale Henrichs

On 03/21/2017 02:06 AM, Norbert Hartl wrote: > Hi Dale, > >> Am 20.03.2017 um 19:30 schrieb Dale Henrichs <dale.henrichs(a)gemtalksystems.com>: >> >> Norbert. >> >> Are you thinking of GemStone-style class versions for Pharo? >> > I was thinking of something less elaborate. The GemStone implementation is not very elaborate ... the `class history` exists pretty much to provide #isKindof: sanity for "two versions of the same class". > The first thing that should be possible is that two classes with different shape of the same name can co-exist in the image.That might be provided by #environment. The next step would be to have something like #migrateTo: in GemStone. Meaning that in case of a class/shape change the old instance is called with the new shape as parameter so I can intercept the process of object migration. > > The #history case of GemStone would be nice of it would be modular and could be loaded. I think it can be done. But the points above need to be done first. > > Norbert > >> Dale >> >> On 03/19/2017 06:16 AM, Norbert Hartl wrote: >>> How far are we to support multiple versions of the same class in an image? Are there existing approaches? >>> >>> Norbert >>> >>> >>> >>> >>> Von meinem iPad gesendet >>> >> >

March 21, 2017

Re: [Pharo-users] Magritte extension

by Esteban A. Maringolo

2017-03-21 7:41 GMT-03:00 Otto Behrens <otto(a)finworks.biz>: >> do you have any documentation that summarizes the changes you did. With all the new mails about magritte and problems it might be time to form a new group caring about a new implementation of a meta-object layer. I think now that we have slots the implementation can be made much more powerful than magritte is today. > > +1 >>> We ran into the instance side problem too, and solved that with >>> prototypes. You need them anyway to get better memento's. >>> >> Same here. There are times where it seems cumbersome to use the instance side descriptions. But the benefits outweigh that problem by far (for me at least). But I do like the idea that instance _and_ class side descriptions are taken into account. Feels right at first. > > We needed to retain the class side descriptions and hacked the new > magritte to still do the old class side thing using the > <magritteDescription> pragma. (ulgy) > > This is because we use MAReport with MADescribedColumn a lot. This > often have to be on the class side because the column applies to all > the rows in the report. How do you guys to table-like lists of things? > > We have an extension to use the descriptions to edit the cells in the > table as well, for the case where you need to edit various columns in > a list. The standard MAReport is read only, as I understand it. Do you > have the requirement to edit tables? And how do you do that? I don't use it in MAReports, but I define queries based on the descriptions for instances yet unknown, so the use case is similar. Currently what I did is an approach similar to how Dolphin Smalltalk handles its "Aspects", e.g. Model>>mementoClass ^self class mementoClassOfInstances Model class>>mementoClassOfInstances ^MACheckedMemento And this can be done for almost everything, incluiding building the description container, ask the instance first, and then the class. Also, I like the pragmas, but only for "tagging" the description methods, I prefer lookup and builds based on simple message sends. Regards, Esteban A. Maringolo

March 21, 2017

Re: [Pharo-users] Neural Networks in Pharo

by Offray Vladimir Luna CÃ¡rdenas

Nice to see this development. On the examples issues suggested by Alexandre, maybe Grafoscopio[1] could be useful to combine prose with code. I have write a new user manual [2] and I'm going to work on it this summer of code. [1] http://mutabit.com/grafoscopio/index.en.html [2] http://mutabit.com/repos.fossil/grafoscopio/doc/tip/Docs/En/Books/Manual/ma… [3] http://gsoc.pharo.org/#topic-grafoscopio-literate-computing-and-reproducibl… Cheers, Offray On 21/03/17 08:38, Alexandre Bergel wrote: > Hi Oleksandr! > > I had a look at your code a couple of weeks ago, as I also got some > interest in Neural networks and genetic algorithm (I will start a > lecture here at my university on this topic). > > I think that your code needs examples. Maybe you can add some simple > examples, such as learning boolean expressions, or having a more > complete example on recognizing handing writing. There is in python > here that does exactly that: > http://neuralnetworksanddeeplearning.com/chap1.html > > My code is available here: > http://smalltalkhub.com/#!/~abergel/NeuralNetworks > <http://smalltalkhub.com/#%21/%7Eabergel/NeuralNetworks> > I wrote this code to support some aspect of my lecture. Do not feel > this as an absolute answer. Having concrete and relevant example is > important and my code do not have such example. > > Push push your code! We need it! > > Cheers, > Alexandre > > -- > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > Alexandre Bergel http://www.bergel.eu > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > > > >> On Mar 21, 2017, at 6:09 AM, Oleksandr Zaytsev <olk.zaytsev(a)gmail.com >> <mailto:olk.zaytsev@gmail.com>> wrote: >> >> Hello. >> >> I'm implementing Neural Networks in Pharo as part of my thesis >> (object-oriented approaches to neural networks implementation). It >> would be really nice to receive some feedback. So if you have any >> ideas, recommendations, or critique, please, write me. What are the >> existing projects or common pitfalls that I should consider? Maybe, >> you can recommend me some nice book or paper on a related topic. I >> would be grateful for any kind of feedback. >> >> Here is the repository: >> http://smalltalkhub.com/#!/~Oleks/NeuralNetwork >> <http://smalltalkhub.com/#%21/%7EOleks/NeuralNetwork>. It's not much, >> but I'm working on it. >> >> Yours sincerely, >> Oleksandr >

March 21, 2017

Re: [Pharo-users] Neural Networks in Pharo

by Serge Stinckwich

On Tue, Mar 21, 2017 at 2:38 PM, Alexandre Bergel <alexandre.bergel(a)me.com> wrote: > Hi Oleksandr! Hi all, > I had a look at your code a couple of weeks ago, as I also got some interest > in Neural networks and genetic algorithm (I will start a lecture here at my > university on this topic). This is great ! > I think that your code needs examples. Maybe you can add some simple > examples, such as learning boolean expressions, or having a more complete > example on recognizing handing writing. There is in python here that does > exactly that: http://neuralnetworksanddeeplearning.com/chap1.html > > My code is available here: > http://smalltalkhub.com/#!/~abergel/NeuralNetworks > I wrote this code to support some aspect of my lecture. Do not feel this as > an absolute answer. Having concrete and relevant example is important and my > code do not have such example. Would be nice to have something similar to Scikit-learn in Python. http://scikit-learn.org/stable/ > Push push your code! We need it! Yes we definitively need something like that ! Have a look to PolyMath that already provide a lot of math libraries: https://github.com/PolyMathOrg/PolyMath I just release v0.85 of PolyMath. Regards, -- Serge Stinckwich UCN & UMI UMMISCO 209 (IRD/UPMC) Every DSL ends up being Smalltalk http://www.doesnotunderstand.org/

March 21, 2017

Re: [Pharo-users] Neural Networks in Pharo

by Alexandre Bergel

Hi Oleksandr! I had a look at your code a couple of weeks ago, as I also got some interest in Neural networks and genetic algorithm (I will start a lecture here at my university on this topic). I think that your code needs examples. Maybe you can add some simple examples, such as learning boolean expressions, or having a more complete example on recognizing handing writing. There is in python here that does exactly that: http://neuralnetworksanddeeplearning.com/chap1.html <http://neuralnetworksanddeeplearning.com/chap1.html> My code is available here: http://smalltalkhub.com/#!/~abergel/NeuralNetworks <http://smalltalkhub.com/#!/~abergel/NeuralNetworks> I wrote this code to support some aspect of my lecture. Do not feel this as an absolute answer. Having concrete and relevant example is important and my code do not have such example. Push push your code! We need it! Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > On Mar 21, 2017, at 6:09 AM, Oleksandr Zaytsev <olk.zaytsev(a)gmail.com> wrote: > > Hello. > > I'm implementing Neural Networks in Pharo as part of my thesis (object-oriented approaches to neural networks implementation). It would be really nice to receive some feedback. So if you have any ideas, recommendations, or critique, please, write me. What are the existing projects or common pitfalls that I should consider? Maybe, you can recommend me some nice book or paper on a related topic. I would be grateful for any kind of feedback. > > Here is the repository: http://smalltalkhub.com/#!/~Oleks/NeuralNetwork <http://smalltalkhub.com/#!/~Oleks/NeuralNetwork>. It's not much, but I'm working on it. > > Yours sincerely, > Oleksandr

March 21, 2017

Re: [Pharo-users] Neural Networks in Pharo

by Guillermo Polito

Hi Oleksandr, I'm working half-time on a team doing simulations of spiking neural networks (on scala). Since the topic is "new" to me, I started following a MOOC on traditional machine learning and putting some of my code in here: https://github.com/guillep/neural-experiments Also, I wanted to experiment on hand-written characters with the Mnist dataset (yann.lecun.com/exdb/mnist/) so I wrote a reader for the IDX format to load the dataset. https://github.com/guillep/idx-reader If you want I can take a look and we can discuss further :) Guille On Tue, Mar 21, 2017 at 11:30 AM, Oleksandr Zaytsev <olk.zaytsev(a)gmail.com> wrote: > I started by implementing some simple threshold neurons. The current goal > is a multilayer perceptron (similar to the one in scikit-learn), and maybe > other kinds of networks, such as self-organizing maps or radial basis > networks. > > I could try to implement a deep learning algorithm, but the big issue with > them is time complexity. Probably, it would require the use of GPU, or some > advanced "tricks", so I should start with something smaller. > > Also, I want to try different kinds of design approaches, including those > that are not based on highly optimized vector algebra (I know that it might > not be the best idea, but I want try it and see what happens). For example, > a network, where each neuron is an object (normally the whole network is > represented as a collection of weight matrices). It might turn out to be > very slow, but more object-friendly. For now it's just an idea, but to try > something like that I would need a small network with 1-100 neurons. > > Yours sincerely, > Oleksandr >

March 21, 2017

Re: [Pharo-users] Magritte extension

by Otto Behrens

> do you have any documentation that summarizes the changes you did. With all the new mails about magritte and problems it might be time to form a new group caring about a new implementation of a meta-object layer. I think now that we have slots the implementation can be made much more powerful than magritte is today. +1 >> We ran into the instance side problem too, and solved that with >> prototypes. You need them anyway to get better memento's. >> > Same here. There are times where it seems cumbersome to use the instance side descriptions. But the benefits outweigh that problem by far (for me at least). But I do like the idea that instance _and_ class side descriptions are taken into account. Feels right at first. We needed to retain the class side descriptions and hacked the new magritte to still do the old class side thing using the <magritteDescription> pragma. (ulgy) This is because we use MAReport with MADescribedColumn a lot. This often have to be on the class side because the column applies to all the rows in the report. How do you guys to table-like lists of things? We have an extension to use the descriptions to edit the cells in the table as well, for the case where you need to edit various columns in a list. The standard MAReport is read only, as I understand it. Do you have the requirement to edit tables? And how do you do that?

March 21, 2017

Re: [Pharo-users] Neural Networks in Pharo

by Oleksandr Zaytsev

I started by implementing some simple threshold neurons. The current goal is a multilayer perceptron (similar to the one in scikit-learn), and maybe other kinds of networks, such as self-organizing maps or radial basis networks. I could try to implement a deep learning algorithm, but the big issue with them is time complexity. Probably, it would require the use of GPU, or some advanced "tricks", so I should start with something smaller. Also, I want to try different kinds of design approaches, including those that are not based on highly optimized vector algebra (I know that it might not be the best idea, but I want try it and see what happens). For example, a network, where each neuron is an object (normally the whole network is represented as a collection of weight matrices). It might turn out to be very slow, but more object-friendly. For now it's just an idea, but to try something like that I would need a small network with 1-100 neurons. Yours sincerely, Oleksandr

March 21, 2017

Re: [Pharo-users] Magritte extension

by Norbert Hartl

Hi, > Am 21.03.2017 um 08:04 schrieb Stephan Eggermont <stephan(a)stack.nl>: > > On 19/03/17 14:14, Norbert Hartl wrote: >> I never managed to take a look at QCMagritte but it sounds as an extension. Why was it never integrated? > > It is rather a monolitical extension. And it was mostly in heavy development. We never stopped to make a real release. Somehow the tutorial code and the demo's were not enough to get traction in the community. Probably too much change at a time. I think the design is sound but we never got into discussions/feedback about that. > do you have any documentation that summarizes the changes you did. With all the new mails about magritte and problems it might be time to form a new group caring about a new implementation of a meta-object layer. I think now that we have slots the implementation can be made much more powerful than magritte is today. > We ran into the instance side problem too, and solved that with > prototypes. You need them anyway to get better memento's. > Same here. There are times where it seems cumbersome to use the instance side descriptions. But the benefits outweigh that problem by far (for me at least). But I do like the idea that instance _and_ class side descriptions are taken into account. Feels right at first. Norbert

March 21, 2017

Re: [Pharo-users] Class versions

by Norbert Hartl

Stef, > Am 21.03.2017 um 07:53 schrieb Stephane Ducasse <stepharo.self(a)gmail.com>: > > What are multiple versions of a classes? > it is what I wrote to Dale. Multiple versions of a class are different classes with the same name in the same image. And in a way that one class can access the other of same name. This way the two classes can build a context like old/new version or something else. It would also need to VM support of having a image side interception method while migration is done. > We are working on a module system for Pharo. The prototype is making progress. You will be able to have different classes with the same name in different classes. > I don't understand that sentence. The last work is supposed to be "environments"? How far is that prototype? thanks, Norbert > Stef > > On Sun, Mar 19, 2017 at 2:16 PM, Norbert Hartl <norbert(a)hartl.name <mailto:norbert@hartl.name>> wrote: > How far are we to support multiple versions of the same class in an image? Are there existing approaches? > > Norbert > > > > > Von meinem iPad gesendet > >

March 21, 2017