Artificial Intelligence
Advertisement

< Main Page

upcoming articles[]

AGI[]

category Artificial general intelligence‏‎

  • Journal of Artificial General Intelligence[1]
  • is AGI "np complete"?[2]
  • AGI-Related Projects: symbolic, artificial life, program search, integrative
  • see also: Artificial general intelligence

Bottom up[]

category robots

  • model free = bottom up robotics
  • Subsumption Architecture = no world model
  • opposite of path planning is a reactive architecture
  • "artificial general intelligence" "bottom up" reactive subsumption

Bottom up AI is a reactive subsumption architecture which leads to Artificial General Intelligence. Instead of planning something like in a path planning algorithm, the idea is that the system evolves by itself without human intervention. A subsumption architecture is sometimes called a “no world model”.

Chatbot[]

category bots

  • AIML = chatbot description
  • create chatbot from text with nltk
  • AIML describes a chatbot corpus
  • paper “Richárd Csáky: Deep Learning Based Chatbot Models, 2017”
  • amelia chatbot consists of semantic memory, episodic memory
  • chatbot languages: ChatScript (2011) very powerful but huge size to download, AIML (2001) outdated
  • see also: List Of Chat Bots

Chess[]

category games

  • evaluation in chess https://www.chessprogramming.org/Evaluation
  • chess evaluation function is determined with reinforcement learning, paper “Learning to play chess using TD(λ)-learning with database games”
  • Temporal Difference: chess evaluation function
  • film about chess computer Andrew Bujalski: Computer Chess 2013
  • first functional chess computer on IBM 704 by Alex Bernstein in 1958


Informed search[]

1 Original notes

  • Informed Search = heuristics, macro search = subgoals are available
  • reward function is an example for informed search
  • heuristic search with hill climbing
  • in the russel/norvig AIMA books there is an entire chapter about informed search

2 Application

  • pathfinding in videogames Kapi, Azyan Yusra, Mohd Shahrizal Sunar, and Muhamad Najib Zamri. "A review on informed search algorithms for video games pathfinding." International Journal 9.3 (2020).
  • A* is an informated search algorithm, Lavin, Alexander. "A Pareto front-based multiobjective path planning algorithm." arXiv preprint arXiv:1505.05947 (2015).

3 A* heuristic function

  • BFS, DFS are example for uninformed search algorithm,
  • heuristic function determines how close a state is to the goal
  • Pathak, Maharshi J., Ronit L. Patel, and Sonal P. Rami. "Comparative analysis of search algorithms." International Journal of Computer Applications 179.50 (2018): 40-43.

4 Prose text A heuristic function determines how close a state is to the goal. In contrast to uninformed search algorithms like breadth-first search, the A* informed search algorithm [3] is more efficient.[4]

Possible applications are pathfinding problems in Videogames. [5]


NP hard[]

  • possible framework for research AI: np hard problems. Show that a pathplanner won't solve the problem, show that direct control can't stabilize a pendulum
  • "evaluation function" tetris. Knowledge is converted into an evaluation function, game is np hard
  • algorithm for solving np hard problems
  • example for np-hard is traveling salesman, a polynomnial algorithm isn't known

SHRDLU[]

  • shrdlu is an example for human-machine-interaction
  • text to animation shrdlu
  • see also: voice command
  • natural language keyframes

In the late 1960s an early project in man-machine interaction was created, called SHRDLU. The idea was that a human operator types in command in natural language and then the software is executing the command. This principle is sometimes called a text to animation interface.

A possible attempt in creating longer motion trajectories is by using keyframes.[6]



Virtual referee[]

There is a software available called goal line technology.[7] It helps the human referee to take decisions and works by video recording in real time.

In a game of soccer the Video assistant referee is the more general term to describe computer aided scoring of the current game situation. The system is judging about the current situation and is located outside of the game in the environment.[8]



Register[]

A

B

C

E-R

S-Z

Related wikis[]


New section[]

References[]

  1. ISSN 1946-0163It is an open access journal.
  2. Demasi, Pedro, Jayme L. Szwarcfiter, and Adriano JO Cruz. "A theoretical framework to formalize AGI-Hard problems." Proceedings of the 3rd Conference on Artificial General Intelligence (AGI’10). 2010.
  3. Lavin, Alexander. "A Pareto front-based multiobjective path planning algorithm." arXiv preprint arXiv:1505.05947 (2015).
  4. Pathak, Maharshi J., Ronit L. Patel, and Sonal P. Rami. "Comparative analysis of search algorithms." International Journal of Computer Applications 179.50 (2018): 40-43.
  5. Kapi, Azyan Yusra, Mohd Shahrizal Sunar, and Muhamad Najib Zamri. "A review on informed search algorithms for video games pathfinding." International Journal 9.3 (2020).
  6. Oshita, Masaki. "Generating animation from natural language texts and framework of motion database." 2009 International Conference on CyberWorlds. IEEE, 2009.
  7. Ugondo, Peter Iorper, and Maggai Tsokwa. "Interpreting video assistant referee and goal-line technology communication: The pitch-based referees perspectives." International Journal of Trend in Scientific Research and Development 3.4 (2019): 1058-62.
  8. Arenas, Matías, et al. "A robot referee for robot soccer." Robot Soccer World Cup. Springer, Berlin, Heidelberg, 2008.
Advertisement