CS101 Introduction to Computing Lecture 34 Intelligent Systems

2688 days ago, 931 views
PowerPoint PPT Presentation
2. Amid the last address

Presentation Transcript

Slide 1

CS101 Introduction to Computing Lecture 34 Intelligent Systems

Slide 2

During the last address … (Graphics & Animation) We got comfortable with the part that illustrations and livelinesss play in registering We examined how representation & activity are shown We additionally took a gander at a few organizations utilized for putting away design and activity

Slide 3

Computer Graphics Images made with the assistance of PCs 2-D and 3-D (showed on a 2-D screen however in a manner that they give a fantasy of profundity ) Used for logical research, masterful expression, or for modern applications Graphics have made the PC interfaces more instinctive by expelling the need to remember charges

Slide 4

Displaying Images Most all PC shows comprise of a matrix of minor pixels orchestrated in a consistent network of lines and sections Images are shown by allotting diverse hues to the pixels situated in the sought segment of the PC show Let's talk about the pixel more …

Slide 5

Pixel The littlest picture framing component on a PC show The PC show is comprised of a normal lattice of these pixels The PC has the ability of doling out any shading to any of the individual pixels on the show Let's currently perceive how the PC shows a square

Slide 6

Pixel Colors (1) The shade of every pixel is by and large spoken to in the shape a triplet In a prevalent plan – the RGB plot – each piece of the triplet speaks to the power of one of out of three essential hues : red, green, blue Often, the force of each shading is spoken to with a byte , bringing about 256x256x256 ( 16+ million ) novel shading blends

Slide 7

Color Mapping (1) Instead of giving every pixel a chance to expect one out of 16 million conceivable hues, just a set number of hues – called the platelet – are took into account case, the platelet might be confined to 256 hues (requiring 1 byte/pixel rather than 3)

Slide 8

Dithering In this plan, pixels of rotating hues are utilized to reproduce a shading that is not present in the platelet For instance, red and green pixels can be exchanged to give the impression of brilliant yellow The nature of the showed picture is poorer

Slide 9

Aliasing The PC screen comprises of square-ish pixels masterminded in a settled framework on occasion, when a corner to corner line is drawn on this matrix, it looks more like a staircase , rather than a straight line This impact – called associating – can be overseen by lessening the measure of pixels

Slide 10

Anti-Aliasing (1) Anti-associating is another strategy utilized for dealing with the "staircase" impact Let's say that we have to draw a white straight-line to such an extent that it covers 60% with one pixel, and 40% with another at first , and close to the end, 58%, 41%, and 1%, separately, with three pixels

Slide 11

Vector or Object-Oriented Graphics Treats everything that is drawn as a question Objects hold their personality after they are drawn These articles can later be effectively moved, extended, copied, erased, and so forth Are determination autonomous Relatively little document estimate Examples: swf , svg , wmf , ps

Slide 12

Bit-Mapped or Raster Graphics Treats everything that is drawn as a bit-delineate a protest is drawn on top of another , it is hard to move only one of them while leaving the other untouched Changing the determination frequently requires significant touch-up work Relatively vast record measure Examples: gif , jpg , bmp

Slide 13

3-D Graphics (1) Flat pictures improved to confer the hallucination of profundity We see the world and the items in it in 3-D - expansiveness, width, profundity - in spite of the fact that the pictures framed on the retinas of our eyes are 2-D The mystery of 3-D discernment: stereo vision

Slide 14

3-D Rendering The way toward changing over data around 3-D objects into a bit-guide that can be shown on a 2-D PC show Computationally, extremely costly ! Steps: Draw the wire-outline (skeleton, made with thin lines) Fill with hues, surfaces, designs Add lighting impacts (reflections, shadows)

Slide 15

Animation Graphics in movement , e.g. kid's shows Illusion of movement is made by demonstrating the watcher a succession of still pictures , quickly Drawing those pictures - each marginally not the same as the past one - used to be very dull work Computers have helped in chopping down a portion of the dullness

Slide 16

Tweening (2) This procedure of making these in the middle of pictures from key pictures is brought in-betweening (or tweening for short) The least difficult calculation for tweening ascertains the position of a specific fragment of a picture by figuring the normal of the places of that same picture section having a place with nearby key pictures

Slide 17

Today's Goals: (Intelligent Systems) To get comfortable with the recognizing elements of clever frameworks as for other programming frameworks To wind up distinctly ready to welcome the part of shrewd frameworks in logical, business and shopper applications To take a gander at a few methods for planning astute frameworks

Slide 18

(Artificial) Intelligent Systems SW projects or SW/HW frameworks intended to perform complex errands utilizing techniques that copy some part of human thought One can face off regarding unendingly about whether a specific framework is smart or not But rather to my brain, the key paradigm is advancement : it is keen on the off chance that it can learn (regardless of the possibility that exclusive a constrained sense) and improve with time

Slide 19

Not a Suitable Hammer for All Nails! in the event that the way of calculations required in an errand is not surely knew or there are an excessive number of exemptions to the rules or known calculations are excessively mind boggling or inefficient then AI has the capability of offering an adequate arrangement

Slide 20

Selected Applications Games: Chess, SimCity Image acknowledgment Medical analysis Robots Business knowledge

Slide 21

Sub-Categories of AI Expert frameworks Systems that, in some restricted sense, can supplant a specialist Robotics Natural dialect handling Teaching PCs to comprehend human dialect, talked and additionally composed Computer vision

Slide 22

Selected Techniques Artificial neural systems Genetic calculations Rule-based frameworks Fuzzy rationale Many circumstances, any of them can take care of the current issue, however at others, just the correct one will do. Along these lines, it is vital to have some energy about them all

Slide 23

Neural Networks (1) Original motivation was the human mind ; accentuation now on handiness as a computational device Many valuable NN ideal models, yet extent of today's examination constrained to the sustain forward system , the most famous worldview

Slide 24

Neural Networks (2) Feed-forward Network: It is a layered structure comprising of various homogeneous and straightforward (yet nonlinear ) handling components All preparing is neighborhood to a preparing component and is offbeat During preparing the FN is compelled to conform its parameters so that its reaction to information turns out to be nearer to the fancied reaction

Slide 25

Genetic Algorithms (1) Based on Darwin's transformative guideline of " survival of the fittest' GAs require the capacity to perceive a decent arrangement, however not how to get to that arrangement

Slide 26

Genetic Algorithms (2) The strategy: An underlying arrangement of arbitrary arrangements is positioned as far as capacity to take care of the issue best case scenario arrangements are then cross­bred and changed to frame another set The positioning and development of new arrangements is proceeded until a sufficient arrangement is found or …

Slide 27

Rule­based Systems (1) Based on the standards of the legitimate thinking capacity of people Components of a RBS: Rule­base Working memory Rule translator

Slide 28

Rule­based Systems (2) The plan procedure: A RBS design meets the master to secure the complete arrangement of heuristics that covers the circumstances that may happen in a given space This set is then encoded as IF-THEN structures to shape the required RBS

Slide 29

Fuzzy Logic (1) Based on the standards of the rough thinking staff that people utilize when confronted with phonetic vagueness The data sources and yields of a fluffy framework are exact , just the thinking is surmised

Slide 30

Fuzzy Logic (2) Parts of the knowledge­base of a fluffy framework: Fuzzy principles Fuzzy sets The yield of a fluffy framework is registered by utilizing: The MIN-MAX system for joining fluffy tenets The centroid technique for defuzzification

Slide 31

Now we think around a couple of methods Let's currently consider the circumstance when we are given a specific issue and made a request to discover an AI answer for that issue. How would we decide the correct procedure for that specific issue?

Slide 32

Selection of an Appropriate AI Technique A given issue can be settled in a few ways Even if 2 procedures deliver arrangements of a comparative quality , coordinating the correct system to an issue can save money on time & assets Characteristics of an ideal strategy : The arrangement contains the greater part of the required data The arrangement meets all other essential criteria The arrangement utilizes the majority of the accessible (valuable) learning

Slide 33

How would we decide the appropriateness of a specific AI method for a given assignment We take a gander at the undertaking's prerequisites and after that see which method satisfies those necessities all the more totally – the one which does, is the one we utilize! Here are a couple of parts of the assignment and the procedures that we should know off …

Slide 34

Accuracy Explainability Response speed Scalability Compactness Flexibility Embedability Ease of utilization Learning bend Tolerance for unpredictability Tolerance for commotion in information Tolerance for scanty information Independe