TL;DR
A group with persons interested in Artificial Intelligence and Machine Learning. New members (beginners and experts) are welcome to join. See requirements below.
Thank you.


Hi,

As a follow up to the survey on this thread https://www.nairaland.com/4209747/survey-people-interested-artificial-intelligence, this thread has been created since we had more than 10 persons interested in this group.

One point raised in the survey thread was about the potential of using whatsapp as the group's platform. We can discuss about this here (there are several platforms to consider; nairaland, whatsapp, google group, slack etc).

Lastly, this post on this thread kick starts the group, therefore, anyone with any discussions/questions/projects can jump right in.

Thank you all.



QUICK REMINDER

Requirements
To ensure you know what you are getting into, below is a list of requirements for interested persons
a. Interested in discussions/developments/research about modern AI systems only (see note below for description about what counts as mordern AI).
b. Serious minded
c. That's all about requirement. :-)

Important Points to Note
a. I'm not here to teach you about AI/ML. No, it's not a one sided thing. We all contribute and learn from each other about the current AI/ML projects/research that you're are working on. The aim is to foster a community of AI developers/researchers in Nigeria.

b. Modern AI (as opposed to classical AI which includes subjects like travelling salesman problem, breadth first search, simple puzzle solver etc.) focus on Machine/Deep Learning techniques (e.g. neural networks) being applied to areas such as computer vision tasks (e.g. image classification/localization), natural language processing tasks (e.g. language translation), speech recognition tasks (e.g. speech-to-text), control/planning tasks etc.

c. I know there exist AI/ML related forums online that bring together people from different countries (such as this reddit group [3]). However, I feel we need to do something more local to foster an AI community here in Nigeria.

d. I'll try to keep this thread in the front page of the programming section daily by so as to get enough viewers as possible.



cc:
Dmayor7
osarenomaspecial
unphilaz
shams040
Samcent
islamics
lum1
kelvine95
4kings
rayval
omokoladejames
Ibroh24
Chukwudaalu
BunchObute
NaHiim
EbonyMira
themonk7
Unmarshalled
Omoadeola
lydia4211

First to comment.

Can anyone with knowledge of NLP build a question answering system with just part of speech tagging?
Please if you can or have insight on how to go about it, discuss with the group.

4kings:
First to comment.

Can anyone with knowledge of NLP build a question answering system with just part of speech tagging?
Please if you can or have insight on how to go about it, discuss with the group.

would the questions be closed domain or they would be open domain questions

Also what platform do you want it to be developed on

raymod170:


would the questions be closed domain or they would be open domain questions

Any.

Also what platform do you want it to be developed on

Anyone you're comfortable with but if you can preferably Python, Matlab, Octave or R.

4kings:

Any.


Anyone you're comfortable with but if you can preferably Python, Matlab, Octave or R.

if am to work on the project I would pick closed domain questions its easier to create POST with

as for the listed platforms I am sorry am a Microsoft person lol but Microsoft AI platform isn't bad at all

I'm late

Unmarshalled:
I'm late

No you aren't

raymod170:


if am to work on the project I would pick closed domain questions its easier to create POST with

as for the listed platforms I am sorry am a Microsoft person lol but Microsoft AI platform isn't bad at all

Nah, i don't need the usual "intent/entity" approach. Correct me if i'm wrong, I've not used Microsoft AI before but i think this is how they roll too.

What i need is for the system to be able to identify a question and the subject of the question based on the part of speech structure of the sentence?

Edit:

if am to work on the project I would pick closed domain questions its easier to create POST with

How easy do you mean?

OK so where should I start, I have basic knowledge of python

Unmarshalled:
OK so where should I start, I have basic knowledge of python

Hmm, SoftEng over to you.

Check this web post for an overview of ML and some interseting ML algorithms.

Pick anyone and learn and let's work on a project here.

4kings:

Nah, i don't need the usual "intent/entity" approach. Correct me if i'm wrong, I've not used Microsoft AI before but i think this is how they roll too.

What i need is for the system to be able to identify a question and the subject of the question based on the part of speech structure of the sentence?

Edit:
How easy do you mean?

closed domain questions focuses on a single domain eg Hotel, where you ask the system questions about hotels and they system responds with answers from a data source which must be created to suite the close ended questions

if you just understand what i just typed you would agree with me that it focuses on descriptive results.

i implemented this model along with other models on my own personal assistant
https://www.nairaland.com/4125501/vee-personal-assistant-new-update

which as i said earlier Microsoft platform can also be used to achieve

best regards

raymod170:


closed domain questions focuses on a single domain eg Hotel, where you ask the system questions about hotels and they system responds with answers from a data source which must be created to suite the close ended questions

if you just understand what i just typed you would agree with me that it focuses on descriptive results.

i implemented this model along with other models on my own personal assistant
https://www.nairaland.com/4125501/vee-personal-assistant-new-update

which as i said earlier Microsoft platform can also be used to achieve

best regards

Great project.
I understand your explanation of closed domain questions.
But Microsoft AI platform should be similar to wit.ai and IBM Watson api, were you are to feed in datasets for training and possibly using naive classifier and decision trees.

Or did you use pos taggers in your approach(without using bag of words model)?

4kings:

Great project.
I understand your explanation of closed domain questions.
But Microsoft AI platform should be similar to wit.ai and IBM Watson api, were you are to feed in datasets for training and possibly using naive classifier and decision trees.

Or did you use pos taggers in your approach(without using bag of words model)?

You correct tho i used bag of words model and feed in datasets for training and naive classifier and decision trees.

but since you in need some differnt from Microsoft ai platform am sorry i cant contribute no more as my knowledge is limited to Microsoft ai platform among other things #Proudly.Net LOL

Best regards

raymod170:


You correct tho i used bag of words model and feed in datasets for training and naive classifier and decision trees.

but since you in need some differnt from Microsoft ai platform am sorry i cant contribute no more as my knowledge is limited to Microsoft ai platform among other things #Proudly.Net LOL

Best regards

Thanks.

I suggest we have a slack channel for the group. I believe it is better structured for collaborations and knowledge sharing.

Unmarshalled:
OK so where should I start, I have basic knowledge of python

4kings thanks for your input. Suggestions from others are also welcomed.

My Suggestion
In addition to 4kings suggestion, the links below are short blog posts that gives some perspective.
1. https://www.digitalocean.com/community/tutorials/an-introduction-to-machine-learning
(gives a very high level introduction to Machine Learning)

2. https://www.toptal.com/machine-learning/machine-learning-theory-an-introductory-primer
(similar to the first link, but goes a little bit deeper, with very little amount of math to describe the algorithm that powers some methods)


Lastly, if you more interested in getting a well rounded knowledge (not just high level overview), then any
1. If you are interested in first getting a rounded knowledge about Machine Learning, then link below is a good introduction
https://www.coursera.org/learn/machine-learning.
The course is video based, so you will need a fair amount of internet data.

2. If you can't do video (because of data), the website below is also nice, but you will need to browse through it (especially the blog section) to mine useful information from it.

3. The link below (machine learning cheat sheet with much emphasis on neural networks) may also be nice.
https://ml-cheatsheet.readthedocs.io/en/latest/index.html

omokoladejames:
I suggest we have a slack channel for the group. I believe it is better structured for collaborations and knowledge sharing.

Yeah, we all need to discuss about the platform issue. We need to decide whether we continue here on Nairaland or else where.

Thank you for your suggestion. Still waiting for other suggestions.

SoftEng please can you help me out with my question.
Cc: osarenomaspecial whiztim umaryusuf

SoftEng:


4kings thanks for your input. Suggestions from others are also welcomed.

My Suggestion
In addition to 4kings suggestion, the links below are short blog posts that gives some perspective.
1. https://www.digitalocean.com/community/tutorials/an-introduction-to-machine-learning
(gives a very high level introduction to Machine Learning)

2. https://www.toptal.com/machine-learning/machine-learning-theory-an-introductory-primer
(similar to the first link, but goes a little bit deeper, with very little amount of math to describe the algorithm that powers some methods)


Lastly, if you more interested in getting a well rounded knowledge (not just high level overview), then any
1. If you are interested in first getting a rounded knowledge about Machine Learning, then link below is a good introduction
https://www.coursera.org/learn/machine-learning.
The course is video based, so you will need a fair amount of internet data.

2. If you can't do video (because of data), the website below is also nice, but you will need to browse through it (especially the blog section) to mine useful information from it.

3. The link below (machine learning cheat sheet with much emphasis on neural networks) may also be nice.
https://ml-cheatsheet.readthedocs.io/en/latest/index.html

Tanx

4kings:
First to comment.

Can anyone with knowledge of NLP build a question answering system with just part of speech tagging?
Please if you can or have insight on how to go about it, discuss with the group.

I didn't want to comment on this because of my limited knowledge in NLP. The only area(task) that I have dabbled into in the past is Language Modeling. However, from what I remember Question Answering Systems (QAS) are still very much limited as researchers have not been able to get it to practically work well on open ended domain. This does not mean that it doesn't work, but it just very hard right now. It's does not work practically well compared to Language Translation task which works well.

Here's something that maybe useful to you. Amazon hosts an Alexa Prize competition (https://developer.amazon.com/alexaprize) for universities around the world to build a conversational agent (which is atleast similar to QAS, if not thesame). You can check out the 2017 prize winner and read through their paper or blog post (if available) to see the techniques used. Who knows, if they decide to open it to candidates outside universities, you can form a group here that can participate as well.

Below are 2 additional links (not related to the competition) that you might find useful
https://aws.amazon.com/blogs/ai/creating-a-question-and-answer-bot-with-amazon-lex-and-amazon-alexa/
https://aws.amazon.com/blogs/aws/announcing-the-aws-chatbot-challenge-create-conversational-intelligent-chatbots-using-amazon-lex-and-aws-lambda/

[url]Question[/url]
Do you have any dataset in mind that you want to use to train your QAS?

Final Note
I once watched a video where a top NLP scientist mentioned that QAS are like crown jewel of NLP because a fully functional QAS should be able to perform other NLP tasks (e.g. you can ask a QAS to translate a text from one language to another, or you can ask a QAS to summarize text)

But guys I'm learning JavaScript currently, should I drop it.

before you think of dropping a language ask yourself if the aim of you picking it up to learn at first is in sight or not.

I remembered when I picked up python to learn and then realized basically it's the the same oop just the syntax thats different

I decided to improve on the four languages I know already and get to stage where I can confidently Say am not just a professional but an expert in these languages

I'm learning js because I like the idea of PWAs and I want to be able this create them

Unmarshalled:
But guys I'm learning JavaScript currently, should I drop it.

I won't advice you drop js. You can use any language to start learning Machine Learning (ML). There's even a deep learning (a subset of machine learning) library for js called https://deeplearnjs.org/. However, it is easier to use the language used by the resource (video lectures or blogposts) from which you are learning.

Lastly, you mentioned in your last post that you were learning js to enable you create PWAs. I assume you mean Progressive Web Applications and if that's the case, then I'll advice you stick to one thing (PWAs or Machine Learning) for now and learn it to level where you can at least work on a simple project and then you can afterwards move on to the next thing. Don't try to learn both PWAs and ML at the same time. It's just my opinion, nothing more. You can still choose to do both (PWAs and ML).

Thanks.

Yes I mean Progressive Web Apps. Since I've started it already I'll stick with it. I'll learn ML and AI after that Though I'll like to be a member of this group.

Unmarshalled:
Yes I mean Progressive Web Apps. Since I've started it already I'll stick with it. I'll learn ML and AI after that Though I'll like to be a member of this group.

OK. You are welcomed to continue to be part of the group.

Thanks a lot

So good to be here. I suggest we make use of a channel where we can easily share articles, videos, PDF files, pictures, etc. Seasons greetings.

Unmarshalled:
Yes I mean Progressive Web Apps. Since I've started it already I'll stick with it. I'll learn ML and AI after that Though I'll like to be a member of this group.

Aah! We are in the same boat ai. Me am learning Android right now. I just do like @SoftEng suggested, I will build my Android skill up, then delve into ML.

I once participated in a challenge and my task was to build a bot using NLP, I wanted using DialogFlow but I was told to build a custom NLP for the nature of my task, na so I give up o.

In natural language processing, research is still going on to create a machine that can read and understand and even compose large text just as human do. maluuba.com they mainly focuses on NLP, they have some nice research papers but if you don't understand those technical jargons, just go to Github and browse the code of that paper and get the code and study it. I believe you will get the intuition.

osarenomaspecial:
In natural language processing, research is still going on to create a machine that can read and understand and even compose large text just as human do. maluuba.com they mainly focuses on NLP, they have some nice research papers but if you don't understand those technical jargons, just go to Github and browse the code of that paper and get the code and study it. I believe you will get the intuition.

thanks for the information would be helpful to me too

raymod170:

thanks for the information would be helpful to me too

you are welcome.