Hi, I'm just starting a course at school called an EPQ, and its basically where you get to do a project/essay on anything you want, and I'm planning to do mine on artificial intelligence. The options are to do a 5000 word essay, or a 1000 word essay and a project. I would rather do a project, so I was wondering if anyone here had any ideas on a AI project, ideally for a complete beginner, as I only have 2 years to plan, produce and review everything, plus do a presentation. Also, if anyone has any other ideas related to computer science, they would be gladly appreciated. Some ideas I have currently thought of are: A program that finds the shortest path between two points, using pre-defined paths (kind of like a sat-nav I suppose) A robot, probably in the form of a car, that can manoeuvre around obstacles. A spelling/grammar checker, that tries to guess what you were trying to say Thanks for any Help
I'm no expert on AI or on in-depth software, for that matter; however, the projects you are considering are sophisticated but well developed, so you will be able to find detailed blueprints on the web. Why don't you want to go for the essay? Personally, I find it easier to describe how things work instead of making them. This way you can engage in critical analysis, which is university level => top marks at school ;-) and I'm sure many of us here would be more than happy to help you with it along th way
When you decide on what you want to do, you can always e-mail professionals, who are engaged in this tasks on everyday basis, for advise and in-depth explanations. I do that for my postgrad law degree and I'm still amazed how tolerant and patient those people are
Thanks for the suggestion, Im not totally against doing the essay, but I often find I'm better at doing things, than writing about them. My teacher is married to someone who did a computer science degree (and did a bit of ai) and she said she would pose some of my questions to him. If I can't find a suitable AI project, I'll probably do the essay on AI+philosophy such as what is intelligence or something like that anyway.
I'd say that your route optimisation project is the least 'AI' option of your three. It's more of a pure mathematical problem than one that requires intelligence or judgement. It's part traveling salesman part vehicle routing and part optimisation problem (Depending on the direction you take - pun intended). It all depends on how 'AI' you want your AI. Does a mathematical problem solver suit you and the project requirements? Does a 'robotic driver' or 'smart assistant / context analyser' fit better? If it was me (and a big part of me wishes it was) I'd go with a 'humanity simulator' that simulates several facets of human behaviour and can run different scenarios with this simulated cross section of population. EG: Take a "What people will do in times of civil emergency" simulation. You could research and develop it in phases. Start with the 'How many people run away?' 'How many freeze?' 'How many run to help?' aspects, and then by adding on additional layers 'Where do they live?' 'Where do they work?' 'How much food do they have?' 'Where can get to safety?' you can move up to a limited but 'full city' simulation.
You could expand your pathfinding idea to become a collective AI between multiple objects - for example a swarm of objects aligning themselves as a group to fit through smaller gaps, or spreading out and encircling a target on command. This can in turn be extended to use a concept of cover, where the objects would find the best route between cover to get to a destination. The real life example of this being soldiers in a war environment.
I just finished my MA degree in Philosophy and International Relations, so if you decide to go with the second option, I'll be more than happy to read it over for you. What subject is this project for? Sent from my iPad using Tapatalk HD
Isn't AI just a really complex mathematics? :-/ I'm just not sure how easy even the basic level of it would be in practical terms...
Its a qualification on its own, done over two years and worth two as levels. Uni's are meant to like it because it shows independent research and suchlike. I actually quite like the idea of going down the philosophy route, but my English skills aren't as good as my practical/maths skills, and it makes the project a lot easier if you actually answer a question/hypothesis which can be difficult to do with philosophical questions, as there often isn't a right answer
I do like the sound of this, but the project doesn't have to be too complicated, as its only worth about 10% of the marks anyway, and I think my teacher who's assessing this will probably think whatever I do is good enough as she isn't that technically minded
I done an AI module last year, I could forward the notes to you. It looks at topics like fuzzy logic (How does words like heavy and big translate into hard values), Use of the PROLOGUE language, pattern recognition, Language and speech recognition. Speech recognition was very interesting, possible you could do something with it on android. Alot of the APIs are already developed so just find a novel way of using it.
I did think about speech recognition but didn't know how to go about it, but it does seem an interesting idea? If I PM you my email, it would be great to have a look over some notes.