‫ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬

‫ﺑﺨﺶ اول‪ :‬ﻣﻌﺮﻓﯽ درس و ﻣﻘﺪﻣﻪ‬

‫ﻣﻨﯿﺮه ﻋﺒﺪوس‬
‫‪[email protected]‬‬

‫‪S E P. 2 0 2 4‬‬
‫ﻓﻬﺮﺳﺖ ﻣﻄﺎﻟﺐ‬
‫‪ ‬ﻣﻌﺮﻓﯽ درس‬
‫‪‬ﻣﻌﺮﻓﯽ ﻣﺮﺟﻊ درس‬
‫‪‬ﻧﺤﻮه ارزﯾﺎﺑﯽ‬
‫‪‬ﺳﺮﻓﺼﻞ ﻣﻄﺎﻟﺐ‬
‫‪‬ﻣﻘﺪﻣﻪ‬

‫‪2‬‬
 ‫ﻫﺪف درس‬
‫‪‬درك ﻣﻔﺎﻫﯿﻢ ﻫﻮش ﻣﺼﻨﻮﻋﯽ‪ ،‬اﻫﺪاف‪ ،‬زﻣﯿﻨﻪ ﻫﺎ‪ ،‬ﻗﺎﺑﻠﯿﺘﻬﺎ و ﮐﺎرﺑﺮدﻫﺎي‬
‫آن و ﻣﻌﺮﻓﯽ ﻣﺴﺎﺋﻞ ﺑﺎز‬
‫‪‬ﯾﺎدﮔﯿﺮي ﻣﻔﺎﻫﯿﻢ ﭘﺎﯾﻪ در ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬
‫‪‬آﺷﻨﺎﯾﯽ ﺑﺎ ﺑﺮﺧﯽ اﺑﺰارﻫﺎي ﭘﺎﯾﻪ در ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬
‫‪‬ﯾﺎدﮔﯿﺮي ﺑﺮﺧﯽ ﮐﺎرﺑﺮدﻫﺎي ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬

‫‪3‬‬
 ‫ﭼﺮا ﻫﻮش ﻣﺼﻨﻮﻋﯽ؟‬
‫‪‬ﯾﮑﯽ از ﺟﺪﯾﺪﺗﺮﯾﻦ ﺣﻮزه ﻫﺎ در ﻋﻠﻢ )از ﺳﺎل ‪ (1956‬ﻣﺤﺴﻮب ﻣﯽ ﺷﻮد‪.‬‬
‫‪‬ﻫﻨﻮز ﻣﺴﺎﺋﻞ ﺑﺎز ﺑﺴﯿﺎر زﯾﺎدي در اﯾﻦ ﺣﻮزه وﺟﻮد دارد‪.‬‬
‫‪‬زﯾﺮﺷﺎﺧﻪ ﻫﺎي ﺑﺴﯿﺎر وﺳﯿﻌﯽ دارد‪.‬‬
‫‪‬ﮐﺎرﺑﺮدﻫﺎي ﻣﻬﻢ و ﮔﺴﺘﺮده اي دارد‪.‬‬

‫‪4‬‬
 ‫ﻣﺮﺟﻊ درس‬
Artificial Intelligence:A Modern Approach
by Stuart Russell and Peter Norvig
4rd Edition, 2021

5
‫ﻧﺤﻮه ارزﯾﺎﺑﯽ‬
‫ﻣﯿﺎن ﺗﺮم‪%30 :‬‬
‫ﭘﺎﯾﺎن ﺗﺮم‪%35 :‬‬
‫ﺗﻤﺮﯾﻦ‪%20-25 :‬‬
‫ﭘﺮوژه‪%15 :‬‬
‫ﮐﻮﯾﯿﺰ‪%5 :‬‬

‫‪6‬‬
 ‫ﺳﺮﻓﺼﻞ ﻣﻄﺎﻟﺐ‬
‫‪‬ﻣﻘﺪﻣﻪ‬
‫‪‬ﻋﺎﻣﻠﻬﺎي ﻫﻮﺷﻤﻨﺪ‬
‫‪‬ﺣﻞ ﻣﺴﺎﺋﻞ ﺑﺎ ﺟﺴﺘﺠﻮ‬
‫‪‬روﺷﻬﺎي ﺟﺴﺘﺠﻮي ﻣﺤﻠﯽ‬
‫‪‬ﺟﺴﺘﺠﻮي ﺧﺼﻤﺎﻧﻪ‬
‫‪‬ﺣﻞ ﻣﺴﺎﺋﻞ ﺑﺎ ﻣﺤﺪودﯾﺖ‬
‫‪‬ﻋﺎﻣﻠﻬﺎي ﻣﻨﻄﻘﯽ‬
‫‪‬ﻣﻨﻄﻖ ﻣﺮﺗﺒﻪ اول و اﺳﺘﻨﺘﺎج در آن‬
‫‪‬ﻋﺪم ﻗﻄﻌﯿﺖ و ﺳﯿﺴﺘﻤﻬﺎي ﻣﺒﺘﻨﯽ ﺑﺮ اﺳﺘﻨﺘﺎج اﺣﺘﻤﺎﻟﯽ‬
‫‪‬ﯾﺎدﮔﯿﺮي ﺗﻘﻮﯾﺘﯽ‬

‫‪7‬‬
 ‫ﻫﻮش ﻣﺼﻨﻮﻋﯽ ﭼﯿﺴﺖ؟‬
‫‪‬ﻫﻮش ﻣﺼﻨﻮﻋﯽ ﭼﯿﺴﺖ؟‬
‫‪‬ﻫﻮش ﭼﯿﺴﺖ؟‬
‫‪‬ﭼﻪ وﯾﮋﮔﯿﻬﺎﯾﯽ ﺑﺎﻋﺚ ﻣﯽ ﺷﻮﻧﺪ ﯾﮏ اﻧﺴﺎن )ﯾﺎ ﻫﺮ ﻣﻮﺟﻮدﯾﺖ دﯾﮕﺮ( ﻫﻮﺷﻤﻨﺪ ﺗﻠﻘﯽ‬
‫ﺷﻮد؟‬

‫‪8‬‬
 «‫ﺗﻌﺮﯾﻒ »ﻫﻮش‬
The ability to carry out abstract thinking (Terman, 1921)
The capacity for knowledge, and knowledge possessed (Henmon,
1921)
The capacity to learn or to profit by experience (Dearborn, 1921)
Intelligence is what is measured by intelligence tests (Boring, 1923)
A global concept that involves an individual's ability to act
purposefully, think rationally, and deal effectively with the environment
(Wechsler, 1958)
A general factor that runs through all types of performance (Jensen)

9
 «‫ﺗﻌﺮﯾﻒ »ﻫﻮش‬
Intelligent activity consists of grasping the essentials in a given
situation and responding appropriately to them (Heim 1970)
A person possesses intelligence insofar as he had learned, or can
learn, to adjust himself to his environment (Colvin 1982)
Intelligence is the ability to use optimally limited resources -
including time to achieve goals. (Kurzweil, 1999)

10
 ‫ﺗﻌﺮﯾﻒ »ﻫﻮش«‬

‫ﻋﻘﻼﯾﯽ ﻓﮑﺮ ﮐﺮدن‬ ‫ﻣﺎﻧﻨﺪ اﻧﺴﺎن ﻓﮑﺮ ﮐﺮدن‬

‫ﻋﻘﻼﯾﯽ ﻋﻤﻞ ﮐﺮدن‬ ‫ﻣﺎﻧﻨﺪ اﻧﺴﺎن ﻋﻤﻞ ﮐﺮدن‬

‫‪11‬‬
 ‫ﺗﻌﺮﯾﻒ »ﻫﻮش«‬
‫ﺗﻌﺮﯾﻒ ﻫﻮش در ﭼﻬﺎر دﺳﺘﻪ ﻗﺮار ﻣﯽ ﮔﯿﺮد‪:‬‬

‫ﻋﻘﻼﯾﯽ‬ ‫ﻫﻮش اﻧﺴﺎن‬

‫ﮐﺮدن‬
‫ﻓﮑﺮ ﮐردن‬
‫ﻋﻘﻼﯾﯽ ﻓﮑر‬
‫ﻋﻘﻼﻧﯽ‬ ‫ﮐﺮدن‬
‫ﻓﮑﺮ ﮐردن‬
‫اﻧﺴﺎن ﻓﮑر‬
‫ﻣﺎﻧﻨﺪ اﻧﺳﺎن‬
‫ﻣﺎﻧﻧد‬ ‫ﻓﮑﺮ‬
‫ﻋﻣل ﮐﺮدن‬
‫ﮐردن‬ ‫ﻋﻘﻼﻧﯽ ﻋﻤﻞ‬
‫ﻋﻘﻼﯾﯽ‬ ‫ﻋﻣل ﮐﺮدن‬
‫ﮐردن‬ ‫اﻧﺳﺎن ﻋﻤﻞ‬
‫ﻣﺎﻧﻨﺪ اﻧﺴﺎن‬
‫ﻣﺎﻧﻧد‬ ‫رﻓﺘﺎر‬

‫ﻣﻨﻄﻘﯽ ﺑﻮدن ﯾﻌﻨﯽ ﯾﮏ ﺳﯿﺴﺘﻢ ﺑﺮ اﺳﺎس داﻧﺶ ﺧﻮد ﺑﻬﺘﺮﯾﻦ ﮐﺎر ﻣﻤﮑﻦ را در‬
‫ﯾﮏ ﻟﺤﻈﻪ اﻧﺠﺎم دﻫﺪ در ﺣﺎﻟﯽ ﮐﻪ اﻧﺴﺎن در ﻫﺮ ﻟﺤﻈﻪ ﻧﻤﯽ ﺗﻮاﻧﺪ ﺑﻬﺘﺮﯾﻦ ﮐﺎر‬
‫ﻣﻤﮑﻦ را اﻧﺠﺎم دﻫﺪ ﭼﻮن ﻫﻤﻪ اﻧﺴﺎن ﻫﺎ ﮐﺎﻣﻞ ﻧﯿﺴﺘﻨﺪ‪.‬‬

‫ﻋﻘﻼﯾﯽ‪ Rational :‬ﻋﻘﻼﻧﯿﺖ‪Rationality :‬‬ ‫‪12‬‬

 ‫ﻣﺎﻧﻨﺪ اﻧﺴﺎن ﻓﮑﺮ ﮐﺮدن)ﻣﺪل ﺳﺎزي ﺷﻨﺎﺧﺘﯽ(‬
‫‪‬درك ﭼﮕﻮﻧﮕﯽ ﺗﻔﮑﺮ اﻧﺴﺎن و ﻋﻤﻠﮑﺮد ﻣﻐﺰ‬
‫◦ درون ﮔﺮاﯾﯽ‬
‫‪ ‬اﯾﺠﺎد ﺗﺌﻮري دﻗﯿﻖ درﺑﺎره ﻋﻤﻠﮑﺮد ذﻫﻦ اﻧﺴﺎن و ﺗﺒﺪﯾﻞ ﺑﻪ ﯾﮏ ﺑﺮﻧﺎﻣﻪ ﮐﺎﻣﭙﯿﻮﺗﺮي‬
‫‪ :(General Problem Solver) GPS ‬ﺳﯿﻤﻮن و ﻧﯿﻮول ‪1961‬‬
‫• ﺑﺮﻧﺎﻣﻪ اي ﮐﻪ ﻓﮑﺮ اﻧﺴﺎن را ﺷﺒﯿﻪ ﺳﺎزي ﻣﯽ ﮐﻨﺪ‪.‬‬
‫• ﻣﺮاﺣﻞ اﺳﺘﺪﻻل ﺑﺮﻧﺎﻣﻪ و ﻣﻘﺎﯾﺴﻪ آن ﺑﺎ ﻣﺮاﺣﻞ ﺣﻞ ﺗﻮﺳﻂ اﻧﺴﺎن‬
‫‪‬ﺗﺠﺮﺑﻪ ﻫﺎي رواﻧﺸﻨﺎﺳﯽ‬
‫◦ ﺗﺮﮐﯿﺐ ﻣﺪل ﻫﺎي ﮐﺎﻣﭙﯿﻮﺗﺮي‬
‫و ﻓﻨﻮن ﺗﺠﺮﺑﯽ رواﻧﺸﻨﺎﺳﯽ‬

‫‪13‬‬
 ‫ﻋﻘﻼﯾﯽ)ﺧﺮدﻣﻨﺪاﻧﻪ( ﻓﮑﺮ ﮐﺮدن‬
‫‪‬ارﺳﻄﻮ‪» :‬ﻓﺮآﯾﻨﺪ اﺳﺘﺪﻻل‪ /‬ﺗﻔﮑﺮ درﺳﺖ ﭼﯿﺴﺖ؟«‬
‫‪‬ﭘﺎﯾﻪ رﯾﺰي »ﻣﻨﻄﻖ«‬
‫‪‬ﺑﺮﻧﺎﻣﻪ ﻫﺎﯾﯽ ﺑﺮ اﺳﺎس ﻗﻮاﻧﯿﻦ ﺗﻔﮑﺮ ﺑﺮاي اﯾﺠﺎد ﺳﯿﺴﺘﻤﻬﺎي ﻫﻮﺷﻤﻨﺪ‬
‫ﻣﻮاﻧﻊ‪:‬‬
‫◦ درﯾﺎﻓﺖ داﻧﺶ ﻏﯿﺮ رﺳﻤﯽ و ﺗﺒﺪﯾﻞ آن ﺑﻪ داﻧﺶ رﺳﻤﯽ‬
‫◦ ﺗﻔﺎوت ﻣﯿﺎن »ﻗﺎﺑﻠﯿﺖ ﺣﻞ ﻣﺴﺌﻠﻪ« در ﺗﺌﻮري و ﻋﻤﻞ )ﺑﻦ ﺑﺴﺖ‬
‫ﻣﺤﺎﺳﺒﺎﺗﯽ(‬

‫‪14‬‬
 ‫ﻣﺎﻧﻨﺪ اﻧﺴﺎن ﻋﻤﻞ ﮐﺮدن‬
‫‪‬اﮔﺮ ﺳﯿﺴﺘﻤﯽ ﺑﺨﻮاﻫﺪ ﻣﺎﻧﻨﺪ اﻧﺴﺎن ﻋﻤﻞ ﮐﻨﺪ ﺑﺎﯾﺪ از ﻃﺮﯾﻖ ﺗﺴﺖ ﺗﻮرﯾﻨﮓ آزﻣﻮده‬
‫ﺷﻮد‪.‬‬
‫‪‬ﺗﻮرﯾﻨﮓ ﺑﺮاي اوﻟﯿﻦ ﺑﺎر ﺗﻌﺮﯾﻒ ﻋﻠﻤﯽ رﺿﺎﯾﺖ ﺑﺨﺸﯽ از ﻫﻮش ﻣﺼﻨﻮﻋﯽ اراﺋﻪ ﮐﺮد‪.‬‬

‫ﺗﺴﺖ ﺗﻮرﯾﻨﮓ‬

‫‪15‬‬
 ‫ﻋﻘﻼﯾﯽ ﻋﻤﻞ ﮐﺮدن‬
‫‪‬رﻓﺘﺎر ﻣﻨﻄﻘﯽ‪ :‬اﻧﺠﺎم ﻋﻤﻞ درﺳﺖ‬
‫‪‬ﻋﻤﻞ درﺳﺖ‪ :‬ﻋﻤﻠﯽ ﮐﻪ ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ اﻃﻼﻋﺎت ﻣﻮﺟﻮد‪ ،‬اﻧﺘﻈﺎر ﻣﯽ رود‬
‫ﺷﺎﻧﺲ رﺳﯿﺪن ﺑﻪ ﻫﺪف را ﺑﻪ ﺣﺪاﮐﺜﺮ ﺑﺮﺳﺎﻧﺪ‪.‬‬
‫‪‬ﻟﺰوﻣﺎ ﺷﺎﻣﻞ ﺗﻔﮑﺮ ﻧﯿﺴﺖ‪ :‬ﭘﻠﮏ زدن‬

‫‪16‬‬
 ‫ﺗﺎرﯾﺨﭽﻪ ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬
‫‪1940 -1950 ‬‬
‫‪ :1943‬ﻣﮏ ﮐﺎﻟﺦ و ﭘﯿﺘﺰ )‪ : (McCulloch & Pitts‬ﻣﺪل ﻣﺪاري ﺑﻮﻟﯽ از‬
‫ذﻫﻦ‬
‫‪ :1950‬ﺗﻮرﯾﻨﮓ‪» :‬ﻣﺎﺷﯿﻦ ﻫﺎي ﻣﺤﺎﺳﺒﺎﺗﯽ و ﻫﻮﺷﻤﻨﺪي«‬
‫‪1950-1970‬‬
‫‪ :1956‬ﻧﺸﺴﺖ در دارﺗﻤﻮث‪ :‬ﭘﯿﺪاﯾﺶ »ﻫﻮش ﻣﺼﻨﻮﻋﯽ«‬
‫‪‬دﻫﻪ ‪ :50‬اوﻟﯿﻦ ﺑﺮﻧﺎﻣﻪ ﻫﺎي ﻫﻮش ﻣﺼﻨﻮﻋﯽ ﺷﺎﻣﻞ‪ :‬ﺑﺮﻧﺎﻣﻪ ﺳﺎﻣﻮﺋﻞ ﺑﺮاي اﻧﺠﺎم‬
‫ﺑﺎزي ﭼﮑﺮ‪ ،‬ﺑﺮﻧﺎﻣﻪ ‪ Logic Theorist‬ﺳﯿﻤﻮن و ﻧﯿﻮول‬
‫‪ :1958‬ﻣﻌﺮﻓﯽ زﺑﺎن ‪LISP‬‬
‫‪ :1965‬اﻟﮕﻮرﯾﺘﻢ ﮐﺎﻣﻞ راﺑﯿﻨﺴﻮن ﺑﺮاي اﺳﺘﺪﻻل ﻣﻨﻄﻘﯽ‬

‫‪17‬‬
 ‫ﺗﺎرﯾﺨﭽﻪ ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬
‫‪ :1966-1973‬ﭘﯿﺪاﯾﺶ ﻧﻈﺮﯾﻪ ﭘﯿﭽﯿﺪﮔﯽ ﻣﺤﺎﺳﺒﺎﺗﯽ‪ ،‬ﺗﻮﻗﻒ ﺗﺤﻘﯿﻘﺎت ﺑﺮ روي‬
‫ﺷﺒﮑﻪ ﻫﺎي ﻋﺼﺒﯽ‬
‫‪ :1969-79 ‬ﭘﯿﺪاﯾﺶ ﺳﯿﺴﺘﻢ ﻫﺎي اوﻟﯿﻪ ﻣﺒﺘﻨﯽ ﺑﺮ داﻧﺶ‬
‫‪ :1980‬ورود ﻫﻮش ﻣﺼﻨﻮﻋﯽ در ﻋﺮﺻﻪ ﻫﺎي ﺻﻨﻌﺘﯽ‬
‫‪ :1986‬ﻣﺤﺒﻮﺑﯿﺖ ﻣﺠﺪد ﺷﺒﮑﻪ ﻫﺎي ﻋﺼﺒﯽ‬
‫‪ :1987‬ﺗﺒﺪﯾﻞ ﺷﺪن ﻫﻮش ﻣﺼﻨﻮﻋﯽ ﺑﻪ ﯾﮏ ﻋﻠﻢ‬
‫‪ :1995‬ﻇﻬﻮر ﻋﺎﻣﻞ ﻫﺎي ﻫﻮﺷﻤﻨﺪ‬

‫‪18‬‬
‫ﮐﺎرﺑﺮد ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬
‫‪ ‬زﺑﺎن ﻃﺒﯿﻌﯽ‬
‫‪‬ﺗﺸﺨﯿﺺ ﮔﻔﺘﺎر‬
‫‪ ‬ﺗﺒﺪﯾﻞ ﻣﺘﻦ ﺑﻪ ﮔﻔﺘﺎر و ﺑﺮﻋﮑﺲ‬
‫‪‬ﺳﯿﺴﺘﻤﻬﺎي ﻣﺤﺎوره اي‬
‫‪‬ﺗﺮﺟﻤﻪ ﻣﺎﺷﯿﻨﯽ‬
‫‪‬ﭘﺮﺳﺶ و ﭘﺎﺳﺦ‬
‫‪‬ﺟﺴﺘﺠﻮي وب‬
‫‪‬ﻃﺒﻘﻪ ﺑﻨﺪي ﻣﺘﻦ‪ ،‬ﺗﺸﺨﯿﺺ ﻫﺮزﻧﺎﻣﻪ‬

‫‪‬ﺑﯿﻨﺎﯾﯽ ﻣﺎﺷﯿﻦ‬
‫‪‬ﺗﺸﺨﯿﺺ اﺷﯿﺎ‪ ،‬ﺗﺸﺨﯿﺺ ﭼﻬﺮه‬
‫‪ ‬ﻧﺎﺣﯿﻪ ﺑﻨﺪي ﺗﺼﺎوﯾﺮ‬
‫‪‬ﻃﺒﻘﻪ ﺑﻨﺪي ﺗﺼﺎوﯾﺮ‬

‫‪19‬‬
 ‫ﮐﺎرﺑﺮد ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬
‫‪‬ﺳﯿﺴﺘﻤﻬﺎي ﻣﻨﻄﻘﯽ‪:‬‬
‫‪‬اﺛﺒﺎت ﮐﻨﻨﺪه ﻫﺎي ﻗﻀﺎﯾﺎ‬
‫‪‬ﺗﺸﺨﯿﺺ ﺧﻄﺎ‬
‫‪‬ﭘﺮﺳﺶ و ﭘﺎﺳﺦ‬

‫‪‬ﺳﯿﺴﺘﻤﻬﺎي ﺗﺼﻤﯿﻢ ﮔﯿﺮﻧﺪه‪:‬‬

‫‪ ‬زﻣﺎن ﺑﻨﺪي ) ﺧﻄﻮط ﻫﻮاﯾﯽ‪ -‬ﮐﺎرﺑﺮدﻫﺎي ﻧﻈﺎﻣﯽ(‬
‫‪‬ﻣﺴﯿﺮﯾﺎﺑﯽ )ﻣﺜﻼ ‪(google-map‬‬
‫‪‬ﺗﺸﺨﯿﺺ ﭘﺰﺷﮑﯽ‬
‫‪‬ﻣﻮﺗﻮرﻫﺎي ﺟﺴﺘﺠﻮي وب‬
‫‪‬ﺳﯿﺴﺘﻤﻬﺎي ﺗﻮﺻﯿﻪ ﮔﺮ‬
‫‪...‬‬

‫‪20‬‬
‫ﮐﺎرﺑﺮد ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬

‫‪21‬‬
 ‫ﮐﺎرﺑﺮد ﻫﻮش ﻣﺼﻨﻮﻋﯽ‬
‫ﻃﺮاﺣﯽ ﻋﺎﻣﻠﻬﺎي ﻫﻮﺷﻤﻨﺪ‬

Environment

Sensors
Percepts
Agent

?
Actuators
Actions

22