Intelligence on tap: AI as a new design material http://interactions.acm.org/arc

Question

Intelligence on tap: AI as a new design material

http://interactions.acm.org/archive/view/july-august-2017/intelligence-on-tap

Read this article and write an one-page reflection on one of the following questions:

1. What are some practical applications AI currently can not solve? What problems are too difficult that you believe AI can never solve?

2. Think of a problem that you want to solve using Neural Network. Explain why? What kind of benefits can be provided? What kind of issues and concerns that may arise?

Explanation / Answer

I would like to answer to question 2 .

2.PROBLEM THAT YOU WANT TO SOLVE USING NEURAL NETWORK.
   The problem that i want to solve using the neural network is POWER MANAGEMENT OF COLLING SYSTEM IN DATA CENTER
REASON:
    Because of generation of lots of data(due to moderen application,social networks(facebook,whatsapp),IOT(Internet of things),etc)
Data center size and numbers has been increasing exponentially .Which indicates lots of power used to store and retrive data from data center.
Lots of power means lots of head generated.
We known that storage devices and processor cannot perform well at high temperature .Even high temperature leads to malfunctioning of STORAGE DEVICE AND PROCESSOR.
We handle this problem by installing a cooling system in data centers.AGAIN like a butterfly effect this cooling system installation consumes more amount of energy
leads to INCREASES THE COST IN THE MAINTAINANCE OF THE DATA CENTER .If there is increase in maintainance of DATA CENTER ,product that relies on Data Center cost
also increases.TO OVERCOME THIS PROBLEM WE ARE GOING TO USE NEURAL NETWORS ON COOLING SYSTEM OF DATA CENTER.
BENEFITS:
   Which gives as significant improvement in cost of maintainance.Consider the following example:Internet giant google, while facing the above problem they decided
to use NEURAL NETWORK to overcome power management of cooling system which helped them in saving MILLION OF DOLLERS.
ISSUES:
   NEURAL NETWORS highly dependent on the data that they have access.If the data is lacking in quality or quantity or reliability,
this will greatly increase the risk of the system making poor decisions.FOR EXAMPLE:consider the neural network which adjust the temperature according to data
provided by temperature sensor.what happens if that temperature sensor fails.IT LEADS TO THE PREDICTON OF WRONG RESULT AND THAT LEADS TO SETTING OF WRONG TEMPERATURE.
which lead to faliure of the disk and processor.TO OVERCOME THIS PROBELM EXTRA SENSORS NEED TO BE ADDED LET THE SYSTEM TO DEPEND UPON MULTIPLE SENSOR INSTEAD OF
SINGLE ONE.which increase the reliablity of decision.

Related Questions

Navigate

• Browse (All)
• Browse I
• Subjects

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