【摘要】 “机器智能”正处于从对人类解决问题能力形式上的模仿，到一种深层次的模拟的过渡阶段。当前机器智能研究不同进路所依托的三种哲学范式(符号主义、联结主义和行为主义)存在局限性。维特根斯坦语言哲学思想为机器智能领域“语义进路的联结主义”的新范式提供了支撑，给我们以重要的哲学启思：对“智能”概念理解上的纠正，对语义的强调以及对“关系”问题的洞见。维氏的“关系”思想将为困扰机器智能领域的“框架问题”提供某种问题解决方案。更多还原

【Abstract】 “Machine Intelligence” is in the transition stage from superficial imitation of human problem-solving ability to deep simulation.There are limitations in the three philosophical paradigms(symbolism, connectionism and behaviorism) on which different approaches to machine intelligence research are based.Wittgenstein′s philosophy of language supports the new paradigm of “semantic connectionism”,and provides us with important philosophical reflections: the correction of the concept of “intelligence”,the emphasis on semantics and the insight into the problem of “relation”.Wittgenstein′s thought on “relation” will provide a solution to the “Frame Problem” that plagues the field of machine intelligence.更多还原