Artificial Intelligence and related fields
This is a modal window.
Das Video konnte nicht geladen werden, da entweder ein Server- oder Netzwerkfehler auftrat oder das Format nicht unterstützt wird.
Formale Metadaten
| Titel |
| |
| Serientitel | ||
| Teil | 2 | |
| Anzahl der Teile | 10 | |
| Autor | ||
| Lizenz | CC-Namensnennung 3.0 Deutschland: Sie dürfen das Werk bzw. den Inhalt zu jedem legalen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen. | |
| Identifikatoren | 10.5446/63163 (DOI) | |
| Herausgeber | ||
| Erscheinungsjahr | ||
| Sprache |
Inhaltliche Metadaten
| Fachgebiet | |
| Genre |
9
10
00:00
Notepad-ComputerNeuroinformatikKünstliche IntelligenzDatenfeldMathematische LogikMustererkennungLogischer SchlussSelbstrepräsentationGenetische ProgrammierungAutomatische HandlungsplanungHeuristikDatenverarbeitungssystemMinkowski-MetrikNeuronales NetzKontrollstrukturFolge <Mathematik>AlgorithmusZeitbereichGewicht <Ausgleichsrechnung>GruppenoperationGlobale OptimierungBeschreibungskomplexitätVollständigkeitEin-AusgabeZeichenketteZahlenbereichAggregatzustandWurzel <Mathematik>TopologieVerschlingungGraphROM <Informatik>KnotenmengeTopologieMusterspracheHalbleiterspeicherBildverstehenProgrammierungLuenberger-BeobachterDomain <Netzwerk>Minkowski-MetrikOrdnung <Mathematik>AggregatzustandRauschenDemoszene <Programmierung>MustererkennungSelbstrepräsentationDatenfeldFolge <Mathematik>Genetischer AlgorithmusNatürliche SpracheMathematische LogikGraphNatürliche ZahlZeichenketteNeuronales NetzSuchverfahrenVerschlingungSuchbaumMultiplikationsoperatorCodeFeasibility-StudieCodierung <Programmierung>Kartesische KoordinatenZeichenvorratElement <Gruppentheorie>Sigma-AlgebraASCIIEndlichkeitArray <Informatik>QuantenzustandDateiverwaltungDatentypFunktion <Mathematik>DatenstrukturFormale SpracheSoft ComputingMereologieUmwandlungsenthalpieInformationFunktionalFlächeninhaltVarietät <Mathematik>Web-SeitePunktInformationsspeicherungVollständigkeitGlobale OptimierungQuellcodeObjekt <Kategorie>Gewicht <Ausgleichsrechnung>Ein-AusgabeWurzel <Mathematik>GruppenoperationZählenOrtsoperatorSpieltheorieSummierbarkeitZahlenbereichSchlüsselverwaltungComputerschachDifferenteKomplex <Algebra>AlgorithmusLokales MinimumTheoremKlasse <Mathematik>TVD-VerfahrenRichtungTaskPrädikat <Logik>Patch <Software>TermSoftwaretestHeuristikKategorie <Mathematik>Physikalisches SystemOntologie <Wissensverarbeitung>BeobachtungsstudieMultigraphQuadratzahlRechenwerkDesign by ContractSchnittmengeStrategisches SpielResultanteATMRoutingGenerator <Informatik>RandomisierungProzess <Informatik>TeilbarkeitMapping <Computergraphik>Automatische HandlungsplanungEinfach zusammenhängender RaumProdukt <Mathematik>SchlussregelHash-AlgorithmusAutorisierungMatchingDefaultAutomatische IndexierungMathematikSchaltwerkMessage-PassingEreignishorizontBildschirmmaskeTypentheorieDatenbankGesetz <Physik>Lesen <Datenverarbeitung>Physikalische Theoriep-BlockRelativitätstheorieArithmetische FolgeBinärcodeLogarithmusLogischer SchlussZusammenhängender GraphSoundverarbeitungBefehl <Informatik>IndexberechnungOrdnungsreduktionWahlfreier ZugriffFormale GrammatikElektronische PublikationMAPMailing-ListeHydrostatikGamecontrollerKonfigurationsraumCASE <Informatik>Ganze ZahlAlgebraische K-TheorieDickeExpandierender GraphTrennschärfe <Statistik>HomepageSichtenkonzeptComputeranimation
Transkript: Englisch(automatisch erzeugt)
00:00
Hello and welcome! In this lesson we will consider the issues of artificial intelligence and related areas. Artificial intelligence and related fields Logical AI What a program knows about the world in general of the facts of specific situations in which it must act in its goals are all represented by sentences of some mathematical logical language.
00:27
The program decides what to do by inferring that certain actions are appropriate for achieving its goals. Search AI programs often examine a large number of possibilities, for example, moves in a chess
00:42
game or inferring by a theorem proving the program. Discoveries are continually made about how to do this more efficiently in various domains. Pattern recognition When a program makes observance of some kind, it is often programmed to compare what it sees with a pattern.
01:00
For example, a vision program may try to match a pattern of eyes and noise in a scene in order to find a face. More complex patterns, for example, in a natural language text, in chess positions or in the history of some events are also studied. Representation Facts about the world have to be represented in some way.
01:24
Usually, languages of mathematical logic are used. Inference From some facts, others can be inferred. Mathematical logic deduction is adequate for some purposes, but new methods of non-monotonic
01:40
inference have been added to logic since 1917. The simplest kind of non-monotonic reasoning is default reasoning in which a conclusion is to be inferred by default, but the conclusion can be withdrawn if there is an evidence to the contrary.
02:01
For example, when we have a bird, we may infer that it can fly, but this conclusion can be reversed when we heard that it is a penguin. It is the possibility that a conclusion may have to be withdrawn that constitutes the non-monotonic character of the reasoning.
02:20
Ordinary logical reasoning is monotonic in that the set of conclusions that can be drawn from a set of premises in a monotonic increases function of the premises. Common sense knowledge and reasoning This is the area in which AI is fuzzers from a human level. In spite of the fact that this has been an active research area since the 1950s, while
02:46
there has been considered progress, for example in developing system of non-monotonic reasoning and theories of action, yet more new ideas are needed. Learning from experience
03:04
Programs do that. The approaches on AI based on connectinism and neural nets specialize in that. There is also learning of laws expressed in logic. Programs can only learn what facts or behavior they formalize can represent, and unfortunately
03:24
learning systems are almost all based on very limited abilities to represent information. Planning Your program starts with general facts about the world, especially facts about the effects
03:40
of actions, facts about the particular solutions and a statement of a goal. From this, they generate a strategy for achieving the goal. In the most common cases, the strategy is just a sequence of actions.
04:03
Epistemology This is a study of kinds of knowledge that are required for solving problems in the world. Ontology Ontology is a study of kinds of things that exist.
04:23
In AI, the programs and sentences deal with various kinds of objects, and we study what these kinds are and what their basic properties are. Emphasis on ontology begins in the 1990s.
04:43
Heuristics A heuristic is a way of trying to discover something or an idea embedded in a program. The term is used variously in AI. Heuristic functions are used in some appropriates to search to measure how to a node in a search
05:05
tree seems to be from a goal. Heuristic predicates that compare two nodes in a search tree to see if one is better than the other. For example, constitutes a advance towards the goal, maybe more useful.
05:25
Genetic programming Genetic programming is a technique for getting programs to solve a task by meeting random list programs and selecting features in millions of generations. Search and control strategies
05:43
Problem solving is an important aspect of artificial intelligence. A problem can be considered to consist of a goal and a set of actions that can be taken to lead to the goal. At any given time, we consider the state of the search space to represent where we have
06:06
reached as a result of actions we have applied so far. For example, consider the problem of looking for a contract length on a football field. The initial state is how we stay out, which is to say we know that the length is somewhere
06:26
on the field, but we don't know where. If we use the representation where we examine the field in units of the square foot, then out the first action may it be examine the square in the top left corner of the field.
06:44
If we do not find the length there, we could consider the state now to be that we have examined the top left square and have not found the length. After a number of actions, the state may be that we have examined 500 squares, and we have now just found the length in the last square we examined.
07:05
This is a goal state, because it satisfies the goal that we had of finding a contact length. Search is a method that can be used by computers to examine a problem space like this in order to find a goal.
07:21
Often, we want to find the goal as quickly as possible or without using too many resources. A problem space can also be considered to be a source space, because in order to solve the problem, we will search the space for a goal state. We will continue to use the term source space to describe this concept.
07:43
In this chapter, we will look at number of methods for examining a source space. These methods are called search methods. The importance of search in AI It has already become clear that many of the
08:03
tasks underlying AI can be phrased in the terms of search for the solution to the problem at hand. Many goal-based agents are essentially problem-solving agents which must decide what to do by searching for a sequence for actions that lend to their solution.
08:23
For production systems, we have seen the need to search for a sequence of rule applications that lend to the required facts or action. For neural network systems, we need to search for the set of connection weights that will
08:40
result in the required input to output mapping. Which search algorithm one should use will generally depend on the problem domain. There are four important factors to consider. Completeness. Is a solution guaranteed to be found if at least one solution exists?
09:05
Optimality. Is a solution found guaranteed to be the best or low-cost solution if there exists more than one solution? Time complexity. The upper bound on the time required to find a solution is a function of the complexity
09:21
of the problem. Space complexity. The upper bound on the storage page memory required at any point during the search is a function of the complexity of the problem. Preliminary concepts. Two values of space for time algorithms.
09:44
Purpose is the input or its part to store some info to be used later in solving the problem. In computer science, counting source is an algorithm for sorting a collection of objects
10:04
according to case that are small positive integers, that is, it is an integer sorting algorithm. It operates by counting the number of objects that possess distinct k-values and applying prefix sum on those count to determinize the positions of each k-value in the output sequence.
10:25
Its running time is linear in the number of items and the difference between the maximum k-value and the minimum k-value, so it's only suitable for direct use in situations where the variant in the case is not significantly greater than the number of items.
10:44
It is often used as a subroutine in writing sort, another slot in algorithm, which can handle large k's more efficiently. String searching algorithms. String searching algorithms, sometimes called string matching algorithms, are an important
11:02
class of string algorithms and try to find a place where one of several strings, also called patterns, are found within a large string or text. A basic example of string searching is when the patterns and the source text are arrays of elements of an alphabet, finding its set.
11:23
Sigma may be a human language alphabet, for example, that letters A through Z and other representations may choose a binary alphabet or DNA alphabet in view informatics. In practice, the method of feasible string search algorithm may be affected by the string
11:41
encoding. In particular, if a variable with encoding is in use, then it may be slower to find the n-th separator, perhaps requiring time proportional to n. This may significantly slow some search algorithms.
12:01
One of many possible solutions is to search for a sequence of code units instead, but doing so many produce false matches unless the encoding is specifically designed to avoid it. Restructuring preprocesses the input to make existing its elements easier.
12:21
Hashing A hashing algorithm is a mathematical function that garbles data and makes it unreadable. Hashing algorithms are on-way programs, so the text can be unscrubbled and decoded by anyone else. And that's the point.
12:40
Hashing protects data at rest, so even if someone gains access to your server, the items towards this remain unreadable. Hashing can also help you prove that data isn't adjusted or altered after the author is finished with it. And some people use hashing to help them make sense of remains of data.
13:03
Those on different hashing algorithms exist, and they all work a little differently. But in each one, people type in data and the program alters it to a different form. Most hashing algorithms follow in this process.
13:23
Create the message. A user determinates what should be hashed. Choose the type. Dozens of hashing algorithms succeed, and the user might decide which works best for this message. Enter the message. The user taps out the message in a computer running the algorithms.
13:45
B3 is a self-balancing 3-data structure that maintains sorted data and allows searches, sequential access, insertions, and deletions in logarithmic time.
14:02
The B3 generalizes a binary search tree, allowing for nodes with more than two children. Unlike other self-balancing binary search trees, the B3 is well-suited for search systems than read-and-write related large blocks of data, such as databases and file systems.
14:23
B3s were inverted by Rudolph Baret and Edward M. McRae while working on Boeing Research Labs for the purpose of efficiently managing index pages for large number-axis files. The basic assumptions for the indicates would be so voluminous that only small chunks of
14:43
the tree could fit in the main memory. Google developed the C++ B3 reporting at 50-80% reduction in memory, use for small data types and improved performance for large datasets when compared to a red-black tree.
15:04
State space representations. The state space is simply the space for all possible states, or configurations that our systems may be in. Generally, of course, we prevent work with some convenient representations of that search space.
15:20
There are two components to the representation of state spaces. Static states. Transitions between states. State space graphs. If the number of possible states of the systems is smaller now, we can represent all of them, along with the transitions between them, in a state space graph.
15:44
For example, roads throw state space. Our general aim is to search for a road, or sequence of transitions, through the state space graph from an initial state to a goal state.
16:02
Sometimes there will be more than one possible goal state. We define a goal test to determine if a goal state has been achieved. The solution can be represented as a sequence of link labels, or transitions, on the state space graph. Note that the labels depend on the direction moved along the link.
16:28
And we may want to find the optimal, best possible paths. We can define link costs and path costs for measuring the cost of going along their particular paths. For example, the path cost may just equal the number of links, or could be sum of individual
16:44
link costs. For most realistic problems, the state space graph will be too large for us to hold all of it explicitly in memory at one time. Search trees. It is helpful to think of the search process as a built-in hub a search tree of roads throw
17:02
the state space graph. The root of the search tree is the search node corresponding to the initial state. The leaf nodes correspond either to states that have not yet been expanded, or to states that generated to other nodes when expanded. At each step the search algorithm shows a new and expanded leaf node to expand.
17:26
The different search strategies essentially correspond to the different algorithms one can use to select which in the next mode to be expanded at each state. Thank you for your attention, see you again!
Empfehlungen
Serie mit 10 Medien