3
$\begingroup$

I'm looking for good examples of successful AI projects and theories that had a relatively good impact on society, economics and military field.

So many years have passed after the first AI researches; hence I'm wondering if it has really increased the quality of our lives.

$\endgroup$
  • $\begingroup$ You are supposing that there is a consensus on the definition of AI. But I am not aware of such a consensus. The "AI" locution means a lot of different things and has several incompatible meanings. $\endgroup$ – Basile Starynkevitch Jan 22 '17 at 21:15
  • $\begingroup$ @BasileStarynkevitch Could you suggest me a better term so I can edit the question? With AI I didn't ment something specific but anything (software, abstract models, ...) that has a good heuristical outcome related to its environment. $\endgroup$ – Lovecraft Jan 22 '17 at 21:38
  • $\begingroup$ No, I cannot. AI means really a lot of very different things, even from the point of view of recognized AI experts. IMHO, that is ok. $\endgroup$ – Basile Starynkevitch Jan 23 '17 at 6:07
  • $\begingroup$ AI means really a lot of very different things, even from the point of view of recognized AI experts. I've got a PhD in AI (1990, symbolic AI). I went a few days ago at an interesting public audition on AI at the French Senate with leading French AI experts (from Google, to some CTO of startups, and our top-class AI scientists from universities, CNRS, INRIA) and there is a consensus that AI does not have any precise meaning today (also in recent US reports) $\endgroup$ – Basile Starynkevitch Jan 23 '17 at 6:13
  • 1
    $\begingroup$ @BasileStarynkevitch I absolutely agree with you. There are different ways to define intelligence and that fact implies different definitions of AI. $\endgroup$ – Lovecraft Jan 23 '17 at 12:16
10
$\begingroup$

I'd say the most successful are the ones so commonly used that we don't even notice them:

  • The mail systems that automatically decipher handwritten addresses on your packages, they use machine vision and have probably been doing it since mid-90s.

  • Algorithmic trading bots on stock markets - they handle something like 85% of all trades.

  • Many modern CPUs use AI techniques, including neural networks, to guess what your program is going to do next and optimize branch prediction and memory fetches.

  • Most good modern fraud and spam detectors use some combination of AI techniques (clustering, decision trees, SVMs, even some machine vision to check out attached pictures) - and, in the opposing BlackHat camp, the latest automatic CAPTCHA breakers use all the latest advancements in deep learning too).

  • And of course there's Google, Facebook and US DoD who try to put AI into anything they can think of.

$\endgroup$
  • 1
    $\begingroup$ Are you so sure of the third point (OS using AI techniques)? AFAIK, branch prediction & memory fetches are mostly done by the processor (and indeed several processors & their makers claim to use AI techniques). For example, do you know, for some Linux distribution, an OS level component (in the kernel or elsewhere) using neural networks or machine learning techniques? Can you name it and give a reference? $\endgroup$ – Basile Starynkevitch Jan 24 '17 at 8:26
  • $\begingroup$ I meant CPUs, of course, I had in mind things like AMD Ryzen and Exynos M1 Not sure why I typed OS... $\endgroup$ – user7427549 Jan 27 '17 at 10:56
3
$\begingroup$

There are lots of great projects in AI.

  1. Self-driving cars: This types of cars use AI to learn the pattern of the roads, speed of car, motion of car, braking power and lots of different features and after sufficient learning, they are capable of driving the car autonomously. The best example of this type of cars is Tesla's self driving car.

  2. Games: Games also use AI to learn the game with the aim of winning the game when played against a human or an AI player. You must have played lots of games on mobile and PC like Chess, Tic-Tac-Toe, etc. You play against the computer and according to the difficulty value set, the computer plays its moves. This difficulty value is nothing but the ability of the AI engine to predict the next moves by the opponent.

  3. Chatbots: There have been lots of development and improvements in Chatbots, so that humans can communicate with them as if they are talking to other human. There are many chatbots designed which answer any question asked by us (of course it is dependent on how much intelligence the bot holds). Some examples are ALICE bot, IBM Watson (which has been the most advanced bot till now).

  4. Expert Systems: Expert systems are those systems which focus on one specific domain and can solve any query related to that domain which is given to it. For example, an expert system can be designed to solve any mathematical equation queried to it. An expert system, in such case, will give the solution of the equation along with the steps (providing steps is important because it is an important component in expert system which is called inference engine).

  5. Prediction systems: There are lots of prediction systems which use AI and Machine Learning to predict something based on some past data. Examples are Weather Forecast system, Stock Market prediction system, Recommendation system (usually available in e-commerce websites like Amazon), etc.

$\endgroup$
-1
$\begingroup$

I believe which most successful AI theory is machine learning, in entire web have machine learning algorithms running, learning by what you do, watch, search, even by the photos you take.

succesful AI projects:

TensorFlow;

scikit-learn;

$\endgroup$
  • $\begingroup$ Does your answer really fit the question? Or we just want to full-fill what SE community is all about Q&A $\endgroup$ – quintumnia Jan 27 '17 at 14:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.