I will understand 10% of this today. After I finish taking this Machine Learning class I am currently taking, I will understand 30% (we haven't covered deep learning yet). After I implement some more advanced machine learning etc myself I may understand 60%. Then I will come back to this. See, I have a plan.
kareem sakr I am taking one with my school. I didn't finish Andrew Ng's machine learning but took a bunch of other ML courses from Johns Hopkins and wash U. These are on Coursera.
#Siraj has this incredible way of blending Advanced Technical knowledge with humor. Awesome #Siraj. U r helping me understand and learn AI easily . Thanks a lot.
siraj, have you considered doing more videos where you focus more on the machine learning itself, with just numpy? I love these practical application videos so much but it wouldn't be a bad idea to go into a more detailed level.
I think that all Q&A problems are NLP tasks. NLP and linguistics are the foundations from which different levels of abstractions constructs more hard things like disambiguation (to answer factoid questions), translation (to understand or reformulate questions), etc... Maybe, there are equivalent. But I don't think so. Great video and code btw! :):)
Qs after viewing (Notes about video shown below) episodic module: what changes on each pass? if sent thru the same wouldn't output be same? Is the code he was showing in the video available to see? I didn't see it on the github page Notes on video Dynamic Memory Network types of memory; semantic (input txt), episodic (addtl info) Gru cell replaces lstm cell; simplifies by only using two gates & no memory unit; update & reset Create sequence (Glove) vectors from input text Separate training data from testing data Feed to input method that creates hidden states after each sentence hidden states known as facts ?? matrix multiplication, biased term. gets into how to use the gru cell with updates and resets trying to find if the current fact is relevant to the answer output from input method feeds question module which processes input word by word Q mod outputs vector using same gru and weights as input module model created by replicating the hippa campus function in human brain ?? Attention function assigns 0 or 1 to each fact multiple episodes are created; processes all facts x times helps determine what info is relevant but what changes on each pass? if sent thru the same wouldn't output be same? loss model cascading loops entropy. Discratic gradient ascent implementation (rmsprop)
At first I thought you ran out of ideas and were building yet another chatbot, then you slam some state-of-the-art shit. Your title is good for new viewers but I almost didn't watch it thinking it was nothing new. Dank vid, anyways! Keep it up.
Hello Siraj, and community, anyone know what would I have to change or tweak in this method for build a Q&A System in another language, other than English as shown? any ideas of pre-trained language models for Portuguese? Keep up the nice work Siraj, congrats! Cheers to all.
Hi Siraj, i had a question regarding the chatbot models......How do i combine a generative model with a retrieval based model??Generative model will be for training the network to learn the language and the retrieval model is for the purpose of retrieving domain knowledge. Can you suggest me something?
Sorry for my dumb question, I am just a newbie. How comes did i type " python execute.py", the screen showed " no module named six.moves". What should I do?
I know this is an old video but I got a question, this kind of AI is useful if you give her long articles you find of the web, mosto of them are not written like the text you provide to your chatbot here. So can you for example, provide a normal article from Vanity Fair about life of Jennifer lawrence and ask to the bot "where is jennifer lawrence born?" will it be able to answer?
Hi siraj how I can build chatbots for health care , which provide some health care solution. Is there any resource to start this kind of chatbot??? or public available dataset ????
Hey Siraj Is it possible to find the similarity of two sentences using Tensorflow? I mean, I would want the result to be a non binary value. Is there a link from where I could find more information about this? I have already looked at Denny Brtiz's code for text classification. Thank you!
siraj please let us know if we can use , one shot learning or mann for fraud or loan defaulter prediction ??? . if so please make a comprehensive video ?
I love the diversity. I do most of my stuff in Keras right now, but there are some problems which are just a bit more complicated and it makes more sense to run them straight in Tensorflow. Keras is great for the 5 min vids, and I think the depth of Tensorflow lends itself to the hour long vids and live chats.
I like to think of Keras as the "bunny slope" of deep learning. It makes it crazy easy to get started. But that does not mean you can't do impressive stuff with it. With the backend, you can do some really low level stuff directly with tensors, such at dot product, matrix multiplication, etc. keras.io/backend/ I'm actually not sure what TF can do that Keras can't.
Siraj I downloaded your program from Github and wanted to try it but it gives me this error. "TypeError: Expected int32, got list containing Tensors of type '_Message' instead." The place where i get this error is on the first instance of "answer = concatenate([response, question_encoded])". Could you please help me with the same ? Thank you.
You mean after the deadline on April 6 or Nemanja's code on this vid? Nemanja's was usng TF and saver which I am used to. Actually, I have managed to saved the model into .json and .h5 not sure is that the correct or the only way? Anyway, Thanks for the vid.
I ran out of time and have to do some (paying) work. Fitting working, but sirajbot is a bit rude right now. github.com/JacobChrist/jacob.christ.sirajbot
Konstantin H It was trained for only 1000 steps (40 min.). For any good results it has to be trained for at least 10000 steps or like 6-7 hours (on my machine though). See details in the notebook.
imagine a chatbot that outputs Bash comands , comand line programs would get an interface more natural than GUI , without having to modify them ( note : non graphical Linux apps work on Windows ) also could i submit that for the coding challenge ?
![[1hr Talk] Intro to Large Language Models](http://i.ytimg.com/vi/zjkBMFhNj_g/mqdefault.jpg)
One day, one day I will understand all of this.
LF Tang... that day is now
I will understand 10% of this today. After I finish taking this Machine Learning class I am currently taking, I will understand 30% (we haven't covered deep learning yet). After I implement some more advanced machine learning etc myself I may understand 60%. Then I will come back to this.
See, I have a plan.
Am in the same boat , most videos on this channel overwhelm me within the first minute , which ML course are you working on ?
kareem sakr I am taking one with my school. I didn't finish Andrew Ng's machine learning but took a bunch of other ML courses from Johns Hopkins and wash U. These are on Coursera.
same here also
#Siraj has this incredible way of blending Advanced Technical knowledge with humor. Awesome #Siraj. U r helping me understand and learn AI easily . Thanks a lot.
siraj, have you considered doing more videos where you focus more on the machine learning itself, with just numpy? I love these practical application videos so much but it wouldn't be a bad idea to go into a more detailed level.
yes Spark that is on my radar, youre reading my mind thanks
come to hyderabad , eat biryani and help us!! we have machines just need the learning part
i should
You could take workshops inhere
I wish I could understand what the hell you're saying. Good lord dude. Is this a tutorial for chemists
I'm a beginner student of deep learning. So your videos help me a lot to understand this universe. Thanks.
np
You'll be the big brother of the AI, the infinite right hands!
indeed
Oh my god. Sirajology is my new religion. Let's change the world with this technology!
lets do this german!
April 1st is a bad day to publish complex instructions about an advanced technology.
lol true
Wow this was way more complicated than the typical video I've seen on your channel. I would need an hour long version of this to understand anything.
I think that all Q&A problems are NLP tasks. NLP and linguistics are the foundations from which different levels of abstractions constructs more hard things like disambiguation (to answer factoid questions), translation (to understand or reformulate questions), etc...
Maybe, there are equivalent. But I don't think so.
Great video and code btw! :):)
good thoughts and thanks!
"Update Reset" is stuck in my head. Thanks.
That update reset part killed me! xD
lol thx
I didn't know we can have explicit memory in an ANN. Awesome video! just as always it is!
thanks!
Qs after viewing (Notes about video shown below)
episodic module: what changes on each pass? if sent thru the same wouldn't output be same?
Is the code he was showing in the video available to see? I didn't see it on the github page
Notes on video
Dynamic Memory Network
types of memory; semantic (input txt), episodic (addtl info)
Gru cell replaces lstm cell; simplifies by only using two gates & no memory unit; update & reset
Create sequence (Glove) vectors from input text
Separate training data from testing data
Feed to input method that creates hidden states after each sentence
hidden states known as facts
?? matrix multiplication, biased term. gets into how to use the gru cell with updates and resets
trying to find if the current fact is relevant to the answer
output from input method feeds question module which processes input word by word
Q mod outputs vector using same gru and weights as input module
model created by replicating the hippa campus function in human brain
??
Attention function assigns 0 or 1 to each fact
multiple episodes are created; processes all facts x times
helps determine what info is relevant
but what changes on each pass? if sent thru the same wouldn't output be same?
loss model cascading loops entropy. Discratic gradient ascent implementation (rmsprop)
Man, I you are awesome! Thank you so much. Stay this much motivated everytime!
At first I thought you ran out of ideas and were building yet another chatbot, then you slam some state-of-the-art shit. Your title is good for new viewers but I almost didn't watch it thinking it was nothing new. Dank vid, anyways! Keep it up.
yooo you know it. Yea chatbot is more eye grabbing, couldve called it Q&A systems though. im gonna keep it fresh
knew I should have waited, was going to use myown algorithm that was way more complex.
yo siraj i have been a fan for awhile now and i love this video.
Hey Siraj
you're the best.
Excellent!
Makes Numa Numa looks new😂😂😂...ohh man....that's some comedy
Great video.. Thanks for Making
Like in previous videos, Thank you very much!
thanks!
Once I trained It how can I use model.predict method?
Oh great, you're now having over 85,000 subscribers.
nice video
thanks!
out[1]: how fix it when i run
Code + Challenge for this video:
github.com/llSourcell/How_to_.. chatbot.
dope t shirt Siraj!
thanks Sabri!
The only person who makes me proud of me being indian
awesome
thx
lol i need to go to episode 1 i think ..
np
n = np....... ( ͡° ͜ʖ ͡°)
Good idear
still awesome， still interesting。
do not forget rap......
thanks siraj
thanks tony
What are the prerequisites for understanding your content? Python programming would certainly be one. Calculus? Other branches of math?
Barnard Rabenold besides programming basically lin algebra, calc and stats, yeah.
yup
Great! Thanks for this, it's right what I was looking for, I'll give it a try
np
"I've gotta go ask the right questions" lol
Hello Siraj, and community, anyone know what would I have to change or tweak in this method for build a Q&A System in another language, other than English as shown? any ideas of pre-trained language models for Portuguese? Keep up the nice work Siraj, congrats! Cheers to all.
thanks! use portugese dataset
Hi Siraj, i had a question regarding the chatbot models......How do i combine a generative model with a retrieval based model??Generative model will be for training the network to learn the language and the retrieval model is for the purpose of retrieving domain knowledge. Can you suggest me something?
hey Siraj great work man like always, just one noob question is that keras code complete?
and what changes should be made to run it on google colab?
awesome!
nice one siraj.Can you try a video on creating chatbot using watson conversation service
My entry, a chatbot to order Ice Cream 🍦 - github.com/Eibriel/ice-cream-truck - (with instructions to deploy on Telegram) Enjoy! - Eibriel
so close to winning Gabriel, but had to give it to Vishal's last minute entry. I put you in the desc. amazing work
Thanks for choosing my code!! :D
I'll be probably participating on more challenges from now on, is a interesting way of getting practice
NIce video tho
thank you for chatbot video
np
Sorry for my dumb question, I am just a newbie. How comes did i type " python execute.py", the screen showed " no module named six.moves". What should I do?
i want to see play lists rather than a single video on topics they are awesome but play lists will be more awesome
we planned to do major project on chat bot..but we dont know how to start and which languages we use.can u suggest me how to do the project?
Siraj, for domain specific chat bot does neural network suit?
Thanks Siraj
np
Hi, which one is the following video? I'd like to see who wins the contest of this video . Many thanks.
Hey did you remove those python tutorials for machine learning? I can't find them anymore
still there in desc
got it! thx
How do we test it once we've run the training?
plz someone reply to this
Please anyone help us
Hi, I want to know the basic hardware and software infrastructure of a chatbot. Could you pls help me with that?
hi siraj u are just awesome ...I love the way u explain the complex thing in funny manner ..... :)..I want to talk to you live..
lol
sir i want to make Chatbot system which tools we used.who language we requried to make this project.
I know this is an old video but I got a question, this kind of AI is useful if you give her long articles you find of the web, mosto of them are not written like the text you provide to your chatbot here. So can you for example, provide a normal article from Vanity Fair about life of Jennifer lawrence and ask to the bot "where is jennifer lawrence born?" will it be able to answer?
Hi siraj how I can build chatbots for health care , which provide some health care solution. Is there any resource to start this kind of chatbot??? or public available dataset ????
@SirajRaval how would you build a phrase or sentence classifier?
The link to the web app is broken.
Hey Siraj
Is it possible to find the similarity of two sentences using Tensorflow? I mean, I would want the result to be a non binary value. Is there a link from where I could find more information about this? I have already looked at Denny Brtiz's code for text classification. Thank you!
yes, vectorize then compute the cosine distance
Cool thanks!
Vectorize using Sequence autoencoders. Then find cosine similarity.
Wouldn't tf-idf be better? I haven't worked with sequence autoencoders.
tf-idf is good when you have a huge corpus.
Siraj I still don't get how episodic and semantic memory is used in the system...
siraj please let us know if we can use , one shot learning or mann for fraud or loan defaulter prediction ??? . if so please make a comprehensive video ?
Things are happening so fast that he had to cover three different architectures in order to catch people up
1:22 Oh man, i remember Numa Numa......yeah.
memories...
How do I access the webapp while the learning code is executing?
I'm a geologist, I want to know more about that bot lunar? did a search on google came back with a lot of rover stuff.
hey Siraj, can you explain new google's BERT model and how to implement it?
+Siraj Raval Just a humble question. Why do you keep switching between frameworks, rather than sticking to just one (such as tensorflow)? :)
Regards.
hard to fit full TF code in this video will consider sticking to one thanks
Siraj Raval by the way, does it suffice to simply be good at on framework, or is it smart to have a knowledge of many?
I love the diversity. I do most of my stuff in Keras right now, but there are some problems which are just a bit more complicated and it makes more sense to run them straight in Tensorflow. Keras is great for the 5 min vids, and I think the depth of Tensorflow lends itself to the hour long vids and live chats.
How did you get started with both? As in, which one to choose when, and how to get better in both etc.
I like to think of Keras as the "bunny slope" of deep learning. It makes it crazy easy to get started. But that does not mean you can't do impressive stuff with it. With the backend, you can do some really low level stuff directly with tensors, such at dot product, matrix multiplication, etc. keras.io/backend/ I'm actually not sure what TF can do that Keras can't.
Imagine a real-time chatbot for game engines, used for RPGs.
Siraj I downloaded your program from Github and wanted to try it but it gives me this error. "TypeError: Expected int32, got list containing Tensors of type '_Message' instead." The place where i get this error is on the first instance of "answer = concatenate([response, question_encoded])". Could you please help me with the same ? Thank you.
So after i train the model, how can i test it? Not so familiar with Keras , doesn't see any checkpoints or saved file in the directory
see the challenge winners for this one
You mean after the deadline on April 6 or Nemanja's code on this vid? Nemanja's was usng TF and saver which I am used to. Actually, I have managed to saved the model into .json and .h5 not sure is that the correct or the only way? Anyway, Thanks for the vid.
Keras has a .save method to save a model.
sometimes its good to give the direct link for the resources specially datasets
We have GRU here. Where are the minions? :P
Why sigmoid...why not some other function like....1/(1+7^-x) ??
Bro can you make a video tutorial on implementation of text summarisation using abstraction method😅
How could one apply a Chatbot to a Credit Card Application?
siraj, please make a question /answer chatbot for bank FAQ into csv file to creating a model in python
where is the challenge readme or code for this video?
live webapp not there?
Hi Siraj, Could you pls help me with a video on Testing a chatbot using python? Need help pls..
We need a Siraj chatbot.
someone should do it
Siraj extracted!
github.com/JacobChrist/RUclips-Caption
I ran out of time and have to do some (paying) work. Fitting working, but sirajbot is a bit rude right now.
github.com/JacobChrist/jacob.christ.sirajbot
I used AIML to give life to my OC T^T
you are so speeding when ur speaking, so shall you speed down your speaking way please, by the way, thanks a lot for these courses
Some times it feels like you are going too fast in your video..try to slow down a little bit and let us understand better!
will do
the German translations of the challenge winner are completely bullshit - "I am reading a book" is translated as "I am a breakfast"
yea true. the architecture was dope tho was the best submission
Konstantin H It was trained for only 1000 steps (40 min.). For any good results it has to be trained for at least 10000 steps or like 6-7 hours (on my machine though). See details in the notebook.
can you provide me a source to learn nlp
I've got a question is AIML that bad?
Rather than bad i mean is it obsolete?
Did you leave Tensorflow and became Keras user?
switching back and forth
Very interesting, but please speak more slowly for people who aren't initiated. It's too much information in too little time.
What about Neural Semantic Encoders?
maybe soon thanks
is this still working today?
Please do voice recognition chatbot next time
will do
Siraj Raval waiting
Can you plz make a tut on api.ai chatbot (Python)?
Thanks in advance!
there is one video available for that
. ruclips.net/video/c6R3EjMQ7H0/видео.html
imagine a chatbot that outputs Bash comands , comand line programs would get an interface more natural than GUI , without having to modify them ( note : non graphical Linux apps work on Windows ) also could i submit that for the coding challenge ?
4:40 ... Can someone plz tell me which song it is?
Its Shooting stars meme... Thanks for Alan for helping me out.
I think i watch his videos just to look smart.
Siraj sir are you Indaian?
THIS IS MY CLASS PROJECT! except in Tensorflow...
woot!
Siraj Raval haha yeap. Especially the part about how it can make education available everywhere, that's like half of my proposal.
can you give a tutorial of an ai powered os
Where my man Vishal Batchu at? (;
beast status
What language does he speak ?