Extending my knowlege


One of the thing I am doing to better myself is learn more about AI and a great place is in IBM for the Great AI Debate.


Here is the panel. Starting from the left
Igor Portugal
Catalyst IT

David Bloch

Isuru Fernando

Peter Harrison

Priti Ambani
The next billion

Richard Wilde

Sonya Crosby

Stu Christie
Ai Forum

and moderated by Guy Burgess
Clendon Lawyers

Personally, I take a positive outlook when it comes to AI. People who like AI gave positive answers and not apocalyptic improving us and looking at the history through the industrial revolution, the life of horses after the cars came. For horses, it wasn’t good, but England had a 30% of farmlands dedicated to feeding horses and that 30 % was opened up.

Again I am biased to this question but none of the other people convinced me to be afraid. All they were saying is that we should fear the unknown and one person was talking about that we are only a little bit better than the rest of the animal and should fear something better than us. One said that if we created an AI smarter than us then are they smarter than us?

The point is people are looking at the worst of humanity that we cant teach it to be good. Is it inadequacy that makes you look at the flaws of humanity and not the marvel we have created?

We have constructed beast of steel that breaths steam and exhale black fogs that travels much faster than we have ever before.

We have constructed weapons of mass destruction that were truly feared to ignite the atmosphere if it goes off.

At the end the day it is intent. It is so far off from a realised self-aware AI.

Fear of the unknown is something that stagnates us.

people who take risk are the ones who are remembered

the one who challenges


Signing off


Spark Event Day


Spark event,

During class time, we had the chance to visit Spark for an Ideation session and tour of the building with Amelia.


We were given the chance to take part in an ideation session on the spot to propose a product that Spark could produce in the future.

We proposed an IOT device assuming the world reached the ability that most of the world appliance and objects are connected.

If everything is connected then there is data, if there is data there is predictability. If we have so much data that is all around us waiting to be caught and analysed there will be patterns to be decerned again going back to predictability.

The example we used was between the fridge and the oven/stove top all being controlled by a phone to determine which objects should connect on a data level.

It will track the contents of the fridge, the cooking time of the food in the oven and on the stove.

The beauty of the thing we came up with is that it isn’t limited to a fridge and cooking options but how much water we use in the shower and we represent in milk bottles to show how much we are using which is an important visualisation when it comes to drought countries or what does your sleeping position say about when you are about to wake up.

It is personalised, not two people cook the same and not two people sleep the same way or use the shower the same way. There ware data here that is waiting to be mined and examined.

Personally I would extend this project using AI particularly IBM Watson because I know more about it, and the fact that we can use AI specifically for pattern recognition.

These data will be relevant to us on an individual level.


Shigenobu Kobayashi’s Colour Theory

As my work is heavily depending on the colour of the particular topic and relating it to emotions, it can be daunting. Especially one of the main focus of many psychologists is to make their environment to be as neutral as possible. Following the advice of my lecturer, I have looked at Shigenobu’s Kobayashi’s Colour theory.

Above is what Kobayashi considered colours that work well and in a particular genre or type. I am trying to achieve to visually convey emotions using shapes and colours and using the colours to convey positive and negative emotions. Do I though?

There are 5 areas that I want to convey are:

  • Cognitive Behaviour
  • Dialectal behaviour
  • Interpersonal
  • Psychodynamic and Psychoanalysis
  • Acceptance and Commitment

with the following meanings:

  • Cognitive behavioral therapy (CBT), which helps you identify unhealthy, negative beliefs and behaviors and replace them with healthy, positive ones
  • Dialectical behavior therapy, a type of CBT that teaches behavioral skills to help you handle stress, manage your emotions and improve your relationships with others
  • Acceptance and commitment therapy, which helps you become aware of and accept your thoughts and feelings and commit to making changes, increasing your ability to cope with and adjust to situations
  • Psychodynamic and psychoanalysis therapies, which focus on increasing your awareness of unconscious thoughts and behaviors, developing new insights into your motivations, and resolving conflicts
  • Interpersonal psychotherapy, which focuses on addressing problems with your current relationships with other people to improve your interpersonal skills — how you relate to others, such as family, friends and colleagues
  • Supportive psychotherapy, which reinforces your ability to cope with stress and difficult situations

In terms of colours, the background should be romantic and sits in dreamy.

The topic names should be divided into two categories,  positive and negative and it shifts between these two types of colour to show the type of emotions.

Clear, neat and clean are good starting points for the positive emotions,

Dandy, dapper and chic, stylish are good for the negative emotions

Staff, M. C. (2016, March). Psychotherapy. Retrieved from http://www.mayoclinic.org/tests-procedures/psychotherapy/details/what-you-can-expect/rec-20197200

5 things to consider for ITS

construct Organigramme a flow chart describing what your artifact will do

You are in charge of your world, play with it and make sure everything fits your world

Make an artifact that best explains your vision which in turn generate more question.

Simulate the perfect scenario where your idea will exist in

Prepare a mock presentation?

Work in progress.

It has been an exciting time for me and my project Conv-O. I was able to present Conv-O at AUT live and both got to meet the Dean of Design and the head lecturer from computer science and relations and they both like the idea of Conv-O.

Now I am preparing for the presentation of my project in a researchers lunch for the 28th of August. This event is an internal AUT event and after, it gives us the opportunity to expand upon this initial idea and I have a diary full of notes where we can take Conv-O

Speeches part 2

Today we had to speak in front of the class and define what ITS means to us.

ITS to me is bits and pieces of technology coming together as a coherent object with the creator showing an explicit meaning in how it is suppose to be interacted so there is no room for ambiguity.

Thats it.

My interest focused on future and modern technology, particularly in the fields of AI, cloud computing, data analysis and big data as I believe from all the talks I have been going to is heading. the beauty of it is that it looks like everything is very rudimentary clunky and new and everyone is saying it. But people are saying that this is the next revolution how machine learning is going to affect the mountain of unsorted and undocumented data. To me, this is what the people back in the 70s-80s about the Internet, with slow data speeds and major uncertainty and push back and some people going as far to calling it as fad. But the Ai space is a bit different, there are push back due to major uncertainty, with American Hollywood propagating a scary and misinformed information about the coming of Ai.

It isn’t skynet and one day will come and take over, that is incredibly unlikely to happen to the point of improbability.

A good example of Ai is a Bot or more specifically a chat bot. At one of the talk, an energy company here in New Zealand called Genesis Energy released a chatbot for their website to do basic things. Check balance, what services are available and on sale etc. All of these things that dont require a human. But if something is beyond the capabilities of the chatbot, it can then pass it off to a human that can do more specific things.

So instead of 10 people in a call center monitoring all the calls, we get most of the chat bots do the rest problems and then two or three do the specific things.

If this is concerning to you then I ask you this, here in New Zealand, companies are literally investing in expensive self checkout machines. Why? Isn’t the point of people shopping and then to go to a manned checkout to be more convenient than a self checkout?

Turns out people don’t want to talk to people and people are more likely to rely on checking their own items out without the interaction of a human. Shopping centers only need one person to look after a row of self check out machines instead of 5 that do the normal checkout but they are not fired, they are shifted around the place to do other things.

Another machine, although not Ai, the ATM was feared to be the killer of bank teller roles, but guess what, it just shifted the bank tellers to a more specialised role without needing to fire them

It is fear that is bred from ignorance and not knowing what it is and panicking already.


Waterfall methodology

This is a software development cycle called waterfall model. It is an obsolete method of development. All the planning is done right up the beginning and that is it.

If I am thinking about my program, it exist to help people see their conversation which goes back all the way from the beginning.

In terms of manufacturing, where this methodology came from, it was Ok in dealing with it, but as it was applied to software development methodology, problems began to surface quickly.

If we have an 18-24 month project and the client get their final product at the end of the development cycle and the vision of the developers and the client was lost along the way, then everyones times was wasted. Leading to wasted sales etc.

This is where agile comes in.

Agile is a development cycle that promotes rapid and flexible response to change and change is going to happen withing a development cycle.

It is true that both of these methods are more closely linked to developments but I think Conv-O can potentially play a role here one day.

At mornign stand up and Sprint Conv-O can provide invaluable data, What happened last sprint, last stand up meetings. Conv-O is simple and convinient and potentially provide more data than users can come up on their own with neural network and make connection where there was just not enough time to figure out things.

Meet IBM Watson

This is IBM Watson, this is IBM’s Ai and is currently hosted on the cloud and people with access to IBM Bluemix have access to the services or capabilities of IBM Watson which I got.

One evening 3 months ago, the inaugural event at IBM building at the viaduct called Ai Happy hour happened.

I found the event randomly as I was scrolling through facebook randomly one afternoon and decided to participate in it. And I did, a person called Fernando “Issy” Issuru spoke that evening. That was it. It was that moment I knew I wanted to do pretty much after University and probably the rest of my personal career, AI.

I decided to make my Semester two project in line with this and began using IBM Watson and trying to incorporate to my projects, which I have done so far.

I am currently using java to write my program and I am using the speech to text API to incorporate to my program, which have been going pretty good.


So my semester two project is an extension of my semester one project with the working title of Convo-O.

It is a visualization of a conversation at its core and what we submitted is essentially a proof of concept which worked but very basic.

It creates bubbles with topics written on them, and the bubble increases in size  as people talk about the topic this works by  taking the average noise in the room and if something goes over that the radius of the bubble increase.

It visualizes the program by seeing how the topic was talked about and how it grows.

Now I have the goal of taking it out of the proof of concept and into a more product ready stage (hopefully).

I want to make three things essentially. I have to convert my program into an apple ready code, because it was written in processing, which is in java which only works for android, so that is one.

I want to use speech to text functions ( IBM Watson and Bluemix) and grab the words from people talking and if certain words are heard then it is only then only increase the bubble. Specific events

Last one using the intent function of IBM Watson so when it hears Add new topic or any variant, it can determine that someone is asking for a new bubble to be created removing the need for a keyboard and mouse/Using another feature of IBM Watson to keep or store the data. This one is more challenging because it is all about learning about databases which I don’t know more about but this is the thing that is most talked about my project to anyone I talk to is how I capture data, which I don’t.