5 answers
Asked Viewed 269 times Translate

How does one program code for artificial intelligence, and how long does that take?

I have always wondered how long it takes for a coder to program A.I's like Siri.
#technology #tech #programming #ai #coding

+25 Karma if successful
From: You
To: Friend
Subject: Career question for you
100% of 4 Pros

5 answers

Updated Translate

José’s Answer

The time it will take you to develop an A.I. application will obviously depend on the complexity and size of what you want to do. Siri's application uses natural language processing to answer questions, make recommendations and perform actions by delegating requests to a set of web services that has been increasing over time. Applications with few functionalities can take only a couple of days or just hours to develop, everything is relative. The main languages that are used most frequently to make artificial intelligence projects are: Python, R, Lisp, Prolog y Java. Python is one of the most used due to its simplicity. Lisp and Prolog are the oldest and most suitable languages for A.I. development. Java can also be considered as a good option for the development of Artificial Intelligence. A.I. has a lot to do with search algorithms, artificial neural networks and genetic programming.

José recommends the following next steps:

If you are passionate about the subject of the A.I., research on the internet some study pensums related to a career in Informatics or computer science, for some universities.
Observe which one devotes more subjects to the area of Artificial Intelligence. They can be compulsory or free choice subjects.
Success ..... I hope it helps you!

Updated Translate

Kin’s Answer

Hi Ty, Artificial Intelligence (AI) we have many definitions, no universal agreement, based on different point of view, a different definition. You can think it as a system thinks or acts humanly, rationally. Such as you said who to made a AI likes "Siri", you need to think about how to let the computer as a human of language understanding and problem solving. Solves crossword puzzles even better than humans. Using a large database of past puzzles, and a variety of information sources, etc. Another area is machine translation.

(1) Incapable to solve real-world problems;
(2) Or can only solve problems inefficiently (Too long time or too expensive).

Logic + Human intelligence = AI (Impractical and difficult problems can be possibly solved)

The aim of AI:
(1) to produce smart programs;
(2) difficult problems become easier/faster to solve.

You can try to using "Python" or "Prolog" to process some BIG DATA, try to mainly concerned with PROLOG and implementation of different algorithms.

Kin recommends the following next steps:

Textbook: ARTIFICIAL INTELLIGENCE: A modern approach, 3rd edition
Authors: Stuart RUSSELL and Peter NORVIG

Updated Translate

Srinivasa’s Answer

Artificial Intelligence (AI) comprises of Data, Algorithms and programming languages to apply Algorithms on top the data. Machine learning is one aspect of the AI. With Machine learning, you can take sample data (for example, 100s of cat pictures) and use different Neural Networks algorithms to teach the machine what a cat looks like. After the training data, you can then take a cat picture and do inference to see how the probablity of the picture you are testing is a cat or not.

Learning AI/ML is not that complex. In fact there are number of ways one can pick these up. If you interested i would suggest taking a small course in Udacity and then use a free tool such as the H2O.ai to apply what you learn

Updated Translate

Sucharitha’s Answer

Artificial Intelligence is a superset of Machine Learning, Deep Learning and Data Science. To begin with AI we must understand what AI can do and what it can't. Once we understand how and where it could be used, we can design the requirements of the model we are going to build.

Assuming that we have novice programming skill level, we can start off learning how to code with Python. Why choose Python when we have so many other programming languages? Because a wide spectrum of tasks can be accomplished through Python Codes be it a multiprocessing task , computer vision task, data science task ,etc. or even a network automation task ! To work on a functionality, just import the specific library and begin to code with the library functions.

Once we get comfortable with establishing the environment to code the model we designed earlier, we need to divide the model into smaller individual features that can be coded and integrated. Depending upon the size of your feature, the coding time [development] can span from days to weeks or months.

Key takeaway here is the simpler the feature is the easier and faster it is to build it !

Sucharitha recommends the following next steps:

To read about How Siri became a feature from being an app : https://www.bigspeak.com/cheyer-kittlaus-siri-viv/

Updated Translate

Shrinivas’s Answer

There are multiple ways to code. One of most commonly used language is Python. You could do some quick learning. Once you have understanding of basic structure of python, you can install various packages available. E.g. There is package called NLTK ( https://www.youtube.com/watch?v=X2vAabgKiuM). This package in python would allow computer to understand language and system can then make sense of what is being typed, its context. E.g. applications like Siri or when you use iPhone or Android to type, you see some recommendations. They could be attributed to packages like NLTK. You could start with Python and try doing simpler ML programs.

Shrinivas recommends the following next steps:

You could get a lot of material on You Tube - www.youtube.com, Coursera(https://www.coursera.org/learn/machine-learning) , Udemy (https://www.udemy.com/) to get you started.