NewStats: 3,265,085 , 8,185,631 topics. Date: Friday, 13 June 2025 at 12:24 PM 2x5n2d

NaijaTroops:
hello all, i also have interest in AI but all books i have seen do not show a way to translate the idea in to code, i personally know a bit of java and hoping to find something that implements the theory to practical. please any one know a good book to consult

Hi @ Mr. NaijaTroops,

I don't really know your areas of research in Artificial Intelligence, so I hope this helps. First look at the quora discussion below :
https://www.quora.com/What-is-the-difference-between-AI-Machine-Learning-NLP-and-Deep-Learning

Most researchers in AI focus on Machine Learning nowadays (as opposed to classic Artificial intelligence). In the last discussion with the early researchers in the field, I was told no useful research results has been produced in classic AI except for a renowned professor in MIT who died some years ago.
So that ,put simply, you will want to focus on Machine learning, Neural Nets and reinforcements learning ( and newly adversarial network).

Most Probably, you won't use JAVA. Among the Artificial Intelligence communities , JAVA is not that common, so far as I know--I might be wrong here. For example the recent NEON (https://www.nervanasys.com/technology/neon/) platform is pythonic. Others like Tensorflow,Caffe, Theano are definitely pythonic. Though there are Torch and CNTK for Microsoft which are not pythonic.

Therefore you need to refresh Python programming. A pdf file was provided above for python. You can follow this to refresh python.

There are other 2 pdf research papers attached in previous posst (kindly check the last 2 posts). You should definitely try to read those - even if you find them to be a heavy read, never mind just browse through to get the concepts.

Since you have a programming background,and you want to get some practice ,I humbly invite you to go through this github link and follow the materials:
https://github.com/rasbt/python-machine-learning-book.

I also attach the book that comes with the materials in the dropbox link just below.

You might also want to read previous post by @ Mr. Neahyo (just above).

dropbox link: https://www.dropbox.com/s/460joemh39xw7kf/Sebastian%20Raschka-Python%20Machine%20Learning-Packt%20Publishing%20%282015%29.pdf?dl=0

Hope this helps. Thanks sir.

Hi @Mr. Neaho,

Thanks sir for the feed back.

(1)

I have attached here a pdf file (MCCExternalKeyncept.pdf) from Google Machine learning crash course conducted some months ago. In the file you have the required python programming paradigms and the links to the python doc explaining those. Though, I never looked at it then, it is about the same thing I would recommend-- And I guess I will also keep it as reference too. It also contains links to important python data science libraries.

I don't really have a particular recommended book on python. And maybe sir (and I might be wrong for this) it might be good to use the docs and stackoverflow and try working on small project.[/center]

For Pandas (after one completes the fundamentals), I also find these videos on youtube very useful:
(www.youtube.com/watch?v=eRpFC2CKvao&list=PLyBBc46Y6aAz54aOUgKXXyTcEmpMisAq3)

But whenever one is in doubt it is good to use the doc always.

(2)

The other 2 Pdf files (as dropbox link) are on statistical Machine learning and Neural network. Since sir, you have a strong background in Statistics, I found both papers, (from the early contributors to the field) to be almost perfect for you. Moreover, both papers are far better than most textbooks on the subjects in of clarity and explanation. You will love them!
From these papers one can then start implementing those ideas using scikit or tensorflow.



(3)

For d3js you dont need to be super proficient in JavaScript to use it. Most times you will be reusing hundreds of examples provided by others.
Check this link (www.youtube.com/watch?v=n5NcCoa9dDU&list=PL6il2r9i3BqH9PmbOf5wA5E1wOG3FT22p) to get started. And in addition sir, after getting through the background concept, feel free to use the examples provided by others: (github.com/d3/d3/wiki/Gallery). At a point you will need a local web server to load your pages while using d3js --especially when you have data in a seperate JSON file . I found google chrome extension "webserver for chrome , 200 ok!" to be useful for this sir.

I hope this makes sense. Thanks sir.


NB:
dropbox links to papers:

https://www.dropbox.com/s/3h2ejable3qes9j/neural_networks_titterington.pdf?dl=0
https://www.dropbox.com/s/vwj8dzf1uy88gu0/Statistical_pattern_recognition_a_review.pdf?dl=0

Hi guys,

This is a nice job @neayho ---you are the big boss. If you permit me to humbly make few suggestions:

1 ) It is highly recommended to use Jupyter notebook and Anaconda: (to provide neat code and explanations)

http://jupyter.readthedocs.io/en/latest/install.html

One can install Anaconda as the package manager :
>> https://www.continuum.io/s

With the above installations one can use both python packages and R.
The R packages can be installed ib the conda environment from the link >>https://conda.io/docs/r-with-conda.html

But if you dont like python we can use RStudio the equivalent of jupyter notebook:
>> http://rmarkdown.rstudio.com/r_notebooks.html

2) Provided here is a repository of examples of jupyter notebook: (link below) from online communities
>> http://nb.bianp.net/sort/views/


3) Why R is very good and the work perfect , one can also try out python libraries:
>> Numpy http://www.numpy.org/
>> Pandas http://pandas.pydata.org/
>> Matplotlib http://matplotlib.org/
and some other ones.


4) Why not consider these libraries for visualisation in your preprocessing (python):
>> Seaborn (https://seaborn.pydata.org/)
>> (http://bokeh.pydata.org/en/latest/)
>> D3JS (the best for interactive visualisation not pythonic https://d3js.org/)


These are just additional standard libraries common in every day job and these are the ones I could --accept my ignorance.

Lastly I ire the treatment of missing data. But missing data is a very heated topic among intellectuals. I humbly invite you to read this paper:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1839993/

and read the book by Roderic and Rubin " Statistical Analysis with missing data"

I hope this helps!

Hi,

This seem an interesting question!
Well it depends on your definition of thinking and what thoughts are. If by thinking , you mean up to human capacity, then I will say no for now. The human thinking mechanism is incredibly complex.

What we can do and we do appreciably well, is to model the human brain and thought process. From this abstraction , we construct a computational model from which we get numbers that we can work with. Check the book by Prof Willian Gerstner "Spiking Neuron Models: Single Neurons, Populations, Plasticity" to get quick insight. Or the "Neural network a comprehensive foundation" by Haykins and Simon or "Theoretical Neuroscience: Computational and Mathematical Modelling of Neural Systems " by Peter Dayan. These are the classics in the field.

As an example you can start with the concept of receptive field from wikipedia . Thats what these models try to do.

Secondly, this question is also related to machine learning. I will say the relation is this:

[ [ (Convex optimization Vs Machine Learning Vs Deep Learning ) ]* Statistical Modelling ]*Algorithmic blackbox = Artificial Intelligence.

This is to illustrate overlapping fields that seeks to build intelligent machines. They share a lot in common but are quite different.!
In particular when you do machine learning , you also do convex optimisation, and when you do deep learning , you are also have to use machine learning. Statistical Modelling wraps everything up--and is the origin of the field. If you ve all these but you don't use the right algorithm then the model is a failure, hence the need for algorithm.

All these fields do seek to make machine make good Decisions-- Thinking. But as far as the *Thinking* concept you are interested in goes, it is the neural networks/deep learning models that are often used -- probably. You can search for Google AlphaGo , an AI machine to confirm this.

In short in the ML world, it is the deep learning/neural models that win. This is a complex topic to discuss here.

Start with Machine Learning then to deep learning, there are still lots to be learned. All these things have existed before but the difference is that today we have enough computational capabilities. In particular these complex models are built on High Performance Computing Machines (with GPU).
I hope this helps!

Straight away.

DONT LEARN ALL LANGUAGES. LEARN ALGORITHM.

Learn to implement algorithm in the language of your choice. You can learn the prog languages in the world but if you dont learn algorithm design you have learnt nothing.

To put it in Language of Leisersson " if you want to be good programmer learn programming lang for ten years or learn algorithm for 2 years"

Which one would you choose. By the way Leiserson is Author of the book "Introduction to Algorithm" which is a classic,

Though I recommend Algorithm design by Tardos as a favorite

I am not a usual here. But this subject seems about education. I think I can add some things.

First it depends on the level at which you want to learn machine learning. But I assume you want to do interesting things with it---not petty studies

1) Statistics and probability are of course a requirement.
Probability,random variables and stochasic processes. Athanasious Papoulis (up to the chapter on statochastic process)
Statistics for Mathematicians by Victor Panaretos
Principles of functional analysis Rudin Walter
A first look at rigorous probability theory Jeffrey Rosentha (this is if you have the time)

This is a video on " Statistical machine learning from stanford and the books are free" Use these to reinforce your knowledge.
https://lagunita.stanford.edu/courses/HumanitiesSciences/StatLearning/Winter2016/about
http://www-bcf.usc.edu/~gareth/ISL/

These text books cover much of the mathematical requirement. These are not really hard! It takes time, I mean even years to *master all. So dont be harsh on yourself. Nobody is an inherent genius--- you only become genius when you do and think about it for a long time---hence and thats what you what your machines to learn to achieve.
*****(I must say I am bias towards Math and forgive me for this)More importantly you have to solve exercises in the books I have listed.Get the solution manual and solve at least 2 to 3 questions per chapter. If you are getting rusty on your math ---thats normal...just relearn your stuffs and you are good. Start from calculus --see recommendation below--- and go to linear algebra, functional analysis, probability/statistics, and machine learning. For multivariate calculus you dont need to know the whole concept , you will only get to use only some of the theorems.



2) Python Machine learning by Sebastian will give you hands on idea about the topic of machine learning (The book uses python which is very easy to learn).

3) Pattern recognition and machine learning Bishop (this is somehow the bible of machine learning)

Then there are few courses on coursera and edx as mentioned by others (Machine learning specialisation , Machine learning by Andrew Ng, and the last from Caltech)

As I said this is if you want rigorous stuff. Otherwise just read the book (2) and get your hands dirty.

Thanks,
Happy reading.


NB:


If you need to brush up your calculus and linear algebra . Here is what I recommend:

Caculus 1A,B,C from MIT (I should say that this is not about trying to solve differentiation but understanding the subtle genesis of the concept and see the picture clearly in your mind. With this I bet you there is no where you would go that you wont say you understand calculus --at least at a good level)

https://www.edx.org/course/calculus-1a-differentiation-mitx-18-01-1x

For Multivariable calculus . check vector calculus by Marsden. or the OCW MIt course on multivariable calculus


For linear algebra straight away go for Linear Algebra and its application by Gilbert!


For programming : don't worry once you get your hands dirty with the book you will learn a lot of machine learning algorithm. But feel free to to MIT OCW python programming course (I assume you will be using python for start--- or you can change to any language of your choice__Especially any language you already know.





********Lastly after working this hard, ask yourself what can you do with machine learning . How do I monetize my ideas? Of course you can go for computer vision, robotics and data analytics and other applied areas ! May be even machine learning algorithm for Nairaland!

can someone provide the newspaper/vendor that carries the full list?

oyo update 50

good luck to everyone sitting NITDA test today

Pls add 07032384471 to the group chat @

The current affairs questions are basically a combination of what is trending (international and Nigerian news) and a little bit of Nigerian history.
The Verbal and Quantitative reasoning are not as hard as that of GMAT (which is a different entity entirely)
The IT/ICT test basic knowledge and current affairs therein.

here can post questions on current affairs that they deem is worth knowing...

Bobbi Kristina Brown, daughter of Whitney Houston, dies at the age of ___ (22)(today's news).
What is the capital of Mauritius __
Which African country is not a member of the AU ___

Here is also the NITDA act http://www.nitda.gov.ng/documents/NITDA%20act%202007.pdf

The current affairs questions are basically a combination of what is trending (international and Nigerian news) and a little bit of Nigerian history.
The Verbal and Quantitative reasoning are not as hard as that of GMAT (which is a different entity entirely)
The IT/ICT test basic knowledge and current affairs therein.

here can post questions on current affairs that they deem is worth knowing...

Bobbi Kristina Brown, daughter of Whitney Houston, dies at the age of ___ (22)(today's news).
What is the capital of Mauritius __
Which African country is not part of the AU ___

Here is also the NITDA act http://www.nitda.gov.ng/documents/NITDA%20act%202007.pdf

SOLARPOWER1:



Drop your number here if u are interested in the whatsapp group, I have created a group for it.That guys number is not on whatsappp

ok.
Its a welcome idea. But I dont use whatsapp -sorry for this. In any case we can communicate here or through mail.

this site might be useful

http://www.indiabix.com/