maban001

[maban001]

Near Completion

I have completed the report and made the presentation video for the software project. The program itself is still not finished but I outline it in the report and presentation that this project is an ongoing project which will be built upon even after university.

[maban001]

Automation of Writing an essay

I focus a lot on automation which is why this idea came to mind. I would just build a program which reads through existing essays across the internet using colab, a tool used for machine learning and artificial intelligence.

The tool will then identify the structure of an essay and attempt to build said essay with keywords.

However, there is a big part of me saying i should not explore this because people will use the tool out of laziness instead of writing their own essay. I will probable not make this to save humans.

[maban001]

Project Update

I have been thinking about implementing the ability for the user to choose five keywords for the program to write and essay.

This feels like it would take a lot of time to plan and build which is why it may not happen in this project but will definitely be discussed in the report as a future improvement to the tool. 

[maban001]

Project Update

https://github.com/Abanour/Abanour.github.io

The coding aspect is going well however i have run into some problem regarding utilising a text document as reference to reducing the word document. I have discussed the problem in the report.

[maban001]

Online VS Offline #2

I have been planning out this discussion and there are some points i would like to cover.

The points i would like to cover are:

- Loss of Data

- Software Updates

- Pricing

- Virus Risks

- Ease of Access

These point are the most important, in my opinion when trying to weigh the pros and cons of online vs offline

[maban001]

Online VS Offline

I will be looking in to the advantages and disadvantages of both online software’s and standalone offline software’s.

[maban001]

Providing a text editor

This may come as a surprise. Providing a text editor is very important,  companies like Google and Microsoft provide this service but it would require the user to sign in to their email accounts to use the service. In other words, it is not free.

Since this tool revolves around a writer, an editor on hand would be very useful in case they feel like writing something. This editor does not have an automatic save feature because i have not got a server to store documents and the editor will not ask for any credential so the editor would not know who owns which text file. A simple export feature has been added to the editor which allows the writer to export the text to PDF.

I have not got the time to design and build a text editor from scratch so i am using a 3rd party API to provide this service.

Tiny: https://www.tiny.cloud/

[maban001]

Spellcheck

Adding the feature to spellcheck using the Levenshtein distance seems to be an advanced feature which is not really practical in this tool. My reasoning for this is that the writer needs to have the basic ability to proofread the document themselves regardless of a tool. However, as a safety precaution for the writer, i have added in the ability for the document to be spellchecked when loaded on to the site.

In future development of the site, there will automation in spellchecks, so in short, instead of the writer having to go through the document to replace words, the website will do just that.

Automation in everything is a good thing but I feel like it would add on the human race not doing things themselves which would make people more lazy.

[maban001]

Learning and using React.JS

React.js, to my understanding is the library used to allows the user to interact with the interface. I have tried to learn the external library for my code, however, there seems to be not much use for the project. The reason why i have decided to learn this library is because it has many application to ease navigation around the website.

The main reason is because learning the language is good for a programmers C.V. and i wanted to showcase this in my project. I have decided to stick with Bootstrap 4 because it is less tedious to use and i am more familiar with the library.  

[maban001]

The beginning of implementation

I have set up the webpage on gitpages where the summarisation tool will live.

The link to the code is:

https://github.com/Abanour/Abanour.github.io

The link to the actual website is:

https://abanour.github.io

The final year project is under “Final Year Project”

As the project stands, the first objective has been completed where the user is able to upload a file and the website is able to read from the file and load the content into the text area provided. For now, it can only take text and not images or any other media but text.

If any other media is included in the file then the textarea gets filled with jargon and symbols because it cannot read the image. I will have to look into that whilst working on the next objective which is to setup and area where the code will be displayed when the summarisation tool finishes with the summarisation process.

[maban001]

Behaviour Tree and WireFrame

The behaviour tree, first 2 images, are designed to understand how the webpage would work. The images identify key points in the process from start to finish.

The wireframe, image above, identifies how the webpage would look to a user and shows all the containers on a website and brief description as to what the purpose of each container is.

[maban001]

Measurement and Classification of Humans and Bots in Internet Chat

https://www.usenix.org/legacy/events/sec08/tech/full_papers/gianvecchio/gianvecchio.pdf

This scientifical paper aims to investigate the potential danger of chat systems used in different applications. Some of the applications stated in the report vary from games, instant messaging, mail services and more. The problem, which is the centre of this report, is to work out how effective a program can detect a chat bot.

This paper has many similarities to my research, in the sense that producing honest quality work is what employers and teachers expect. Some of the tools developed today can be used to avoid spending anytime writing a report. The user would just write an automation program, as stated in the report “The term bot, short for robot, refers to automated programs, that is, programs that do not require a human operator”.

From experience, chat bots are very monotonous. They have a script with “if this is asked then answer in this way”. So, generally, you will not get the answer you need because the bots have been programmed to respond in a certain...

The rest will be in my draft report

[maban001]

07/12/2020

It has been a month since my last post and i would like to explain the reason why i have not been active doing some research around my project.  Both of my other modules have taken alot of time; with coursework due dates close to each other. However, this isnot me making an excuse.

I have done some research into papers i would like to read and gain a better understanding of how an AI would understand human pattern or more specifically human text. 

i have decided to research these 3 topics to a certain extent to help with my report:

Measurement and Classification of Humans and Bots in Internet Chat

https://www.usenix.org/legacy/events/sec08/tech/full_papers/gianvecchio/gianvecchio.pdf

S Gianvecchio, M Xie, Z Wu, H Wang - USENIX security symposium, 2008 - usenix.org

 Humans and bots in internet chat: measurement, analysis, and automated classification

https://ieeexplore.ieee.org/abstract/document/5755133

S. Gianvecchio, M. Xie, Z. Wu and H. Wang, "Humans and Bots in Internet Chat: Measurement, Analysis, and Automated Classification," in IEEE/ACM Transactions on Networking, vol. 19, no. 5, pp. 1557-1571, Oct. 2011, doi: 10.1109/TNET.2011.2126591.

 A neural attention model for abstractive sentence summarization

https://arxiv.org/abs/1509.00685

AM Rush, S Chopra, J Weston - arXiv preprint arXiv:1509.00685, 2015 - arxiv.org

I will be making a separate Tumblr post for each paper to give a detailed analysis of the research.

[maban001]

02/11/2020 - My First supervisor meeting

Today, 2nd November 2020, I have met with my supervisor. We scheduled this appointment beforehand to clear out some stuff for my software project. The meeting was a very engaging conversation and discussion.

Some of the aspects we discussed today was the need for better clarification of the project; in better terms, what problem i am specifically targeting and how am i going to overcome this problem.

The problem i have decided to tackle is the difficulty of staying within the word count limit when writing an essay, research paper or even journalist report. The chosen solution is to build up a dictionary filled with sentence and their synonyms and write a program to substitute the word in place of the sentence to shorten the word count. We also briefly discussed whether it would relatively achievable if i took the machine learning route or a dictionary in the time frame available to me.

Some of the existing programs i would like to look into more are Microsofts Word API and Grammarly. A separate Tumblr post will be made to discuss my findings. Researching research papers was recommended by Dr Golnaz to understand what experts think of machines learning and bots being utilised in recognising human speech patterns. An example given to me by Dr Golnaz was how humans can recognise sarcasm in tweets or social media but a bot will have difficulty doing so, which i will look into more details and cite my findings.