学术月刊
學術月刊
학술월간
Academic Monthly
2010年
6期
43~50
,共null页
意向性 人工智能 解释 表征 计算
意嚮性 人工智能 解釋 錶徵 計算
의향성 인공지능 해석 표정 계산
intentionality, artificial intelligence, interpretation, representation, computation
“塞尔问题”即机器智能的意向性缺失难题,是心灵哲学、人工智能哲学和认知科学中的焦点问题之一。要造出真正能超出人类智能的智能,下一步就是要探讨意向性现实出现的机制与条件以及让它在机器中得以实现的方法和途径。沿着这一思路,经典计算主义在辩护基础上开始研究句法加工如何表现语义性,新生代计算主义正在崛起,联结主义作了新的回应,进化论语义学倡导研究大自然的“意向性建筑术”,解释语义学也是这一探索中所出现的一种具有个性的理论。卡明斯强调:表征并不是意向性的充分条件,而只是必要条件。要让人工智能表现出意向性,一是要让它有表征能力,二是要有社会、文化等约束因素,三是要让计算机有内在的整体结构,四是要在人工智能的研究中探讨如何让机器有内在的解释机制。解释语义学无疑为人工智能如何真正接近或超越自然智能指出了一条可能的出路。人的意向性中那种隐秘的、重要的关联能力似乎隐约地掀开了它的神秘面纱:借解释可部分实现这种关联。
“塞爾問題”即機器智能的意嚮性缺失難題,是心靈哲學、人工智能哲學和認知科學中的焦點問題之一。要造齣真正能超齣人類智能的智能,下一步就是要探討意嚮性現實齣現的機製與條件以及讓它在機器中得以實現的方法和途徑。沿著這一思路,經典計算主義在辯護基礎上開始研究句法加工如何錶現語義性,新生代計算主義正在崛起,聯結主義作瞭新的迴應,進化論語義學倡導研究大自然的“意嚮性建築術”,解釋語義學也是這一探索中所齣現的一種具有箇性的理論。卡明斯彊調:錶徵併不是意嚮性的充分條件,而隻是必要條件。要讓人工智能錶現齣意嚮性,一是要讓它有錶徵能力,二是要有社會、文化等約束因素,三是要讓計算機有內在的整體結構,四是要在人工智能的研究中探討如何讓機器有內在的解釋機製。解釋語義學無疑為人工智能如何真正接近或超越自然智能指齣瞭一條可能的齣路。人的意嚮性中那種隱祕的、重要的關聯能力似乎隱約地掀開瞭它的神祕麵紗:藉解釋可部分實現這種關聯。
“새이문제”즉궤기지능적의향성결실난제,시심령철학、인공지능철학화인지과학중적초점문제지일。요조출진정능초출인류지능적지능,하일보취시요탐토의향성현실출현적궤제여조건이급양타재궤기중득이실현적방법화도경。연착저일사로,경전계산주의재변호기출상개시연구구법가공여하표현어의성,신생대계산주의정재굴기,련결주의작료신적회응,진화론어의학창도연구대자연적“의향성건축술”,해석어의학야시저일탐색중소출현적일충구유개성적이론。잡명사강조:표정병불시의향성적충분조건,이지시필요조건。요양인공지능표현출의향성,일시요양타유표정능력,이시요유사회、문화등약속인소,삼시요양계산궤유내재적정체결구,사시요재인공지능적연구중탐토여하양궤기유내재적해석궤제。해석어의학무의위인공지능여하진정접근혹초월자연지능지출료일조가능적출로。인적의향성중나충은비적、중요적관련능력사호은약지흔개료타적신비면사:차해석가부분실현저충관련。
"Searl problem" that artificial intelligence (AI) has no intentionality is the front and focus problem in the philosophy of mind, philosophy of AI and cognitive science. It is generally recognized that to make AI really exceed human intelligence, the next step is to research conditions of intentionality actual occurrence and ways of how to achieve in machinery. Along the train of thought, classical computationalism start to study syntax how to represent semantics based on advocacy; new computationalism is rising; connectionism made a new response; evolution semantics promotes to research "art of intentionality architecture'of the nature; interpretational semantics also is a kind of theories with its own personality. R. Cummins emphasizes; representation is not sufficient condition but necessary condition for intentionality. To make AI present intentionality four conditions are necessary. First, it must have ability of representation. Second, it must have social and cultural constraints. Third, let the computer has the overall structure of the inner. Fourth, we must study how to make machinery have internal explanation facility. Undoubtedly interpretational semantics points out a possible way to AI how to really approach or exceed the natural intelligence. That hidden and important association in human intelligence has opened the veil of mystery faintly. We can partially achieve the association by interpretation.