Why is test data necessary in current artificial intelligence research - can som

Question

Why is test data necessary in current artificial intelligence research - can someone give me a brief and some pros/cons on this . Thanks

Here are some examples if you do not understood my question - e.g. why the artificial intelligence researchers need it ? what is lacking in the test data? why the test data is important to artificial intelligence ? why is the test data necessary? why the current AI research require test data? could talk about the self driving cars? Hope that sum up my question.

Explanation / Answer

Artificial Intelligence is the science and engineering of making intelligent machines (intelligent computing systems) and intelligent programs. Artificial Intelligence (AI) is that ability of computer software and hardware to do the things with intelligent behavior aspects.
These may include activities such as:
1.Searching for a required material provided only limited direction, especially from a large volumes of available data.
2.Surmounting constraints: estimating possible ways in real time that something will fit into a confined space, taking apart or building a complex object, or moving through a difficult maze.
3.Pattern Recognization : finding items with similar characteristics dynamically
4.Drawing dynamic conclusions based on understood reasoning and making logical inferences

AI has gained lot of prominence in multiple specialized areas like Neural Networks, Expert Systems, Genetic
Algorithms, Robotics, Logic Programming, and Fuzzy Logic and much more.

Testing plays a vital role in software development life cycle. Software testing is a important element in
software quality assurance and represents the ultimate review of design and coding for the given specifications. It is in totaly a laborious,
costly and time consuming.

The systematic design and appropriate selection of test cases and test data is highly essential for a well quality test. Test cases and test data generation is a key problem in software testing and its automation improves the efficiency and effectiveness and reduces the high cost of software testing. The application of Artificial Intelligence techniques in Software Testing is an emerging highly proactive area
that brings about the comibination of ideas across the domains. Artificial Intelligence techniques of searching are used to automate test data and test cases.

Generally for testing AI alogrithms, in simplest methodology there will be two datasets:
1.A large"training set" which will be used to build and tune the algorithm and
2.A separate smaller "probe set" that is used to evaluate the performance also reffered as 'evaluation set'

For tuning your AI algorithms, probably you need 3 datasets, each one for learning, tuning, and for evaluation. What defines tuning depends on your algorithm, but a typical example is a model where you have a few hyper-parameters. In short you train the algorithm to make a probabilistic decision if a particular item is part of a valid group or simply a junk. So test data is highly crucial ingredient in AI.

So the testing is usually actually how your A.I. is being programmed. Most models are just probabilistic representations of what is probably good and probably bad. Calculating every possible state is easier for computers because they can focus on one task for very long periods of time and how much they remember is exactly how much RAM is provided.

Related Questions

Navigate

• Browse (All)
• Browse W
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.