luyanaa 发布的文章

MuCulloch&Pitts的论文极为有趣,然似乎不应称为“哲学”。留待对逻辑等问题有所了解以后再读。
Churchland的论文其实与MuCulloch&Pitts的有异曲同工之妙,两者都从结构出发,考虑计算的模式问题。前者出于一种纯粹抽象的模式,试图建立一个和神经网络结构类似的模型,证明其上可以实现基本的逻辑计算。后者则出于现实,解释了其实现了的一类变换。
通过材料质别来批评计算机不可能产生意识,借用 Wiener 的话,不禁让人感觉是活力论者的空虚的辩护,类似当年的有机-无机之辩。
作者批评计算是没有意向性的,因而计算不可能产生意识。但代码自身在编写的过程中已经被人赋予了意向性。形式化的道路是否一开始就是走不通的,类似数学中确定性逐步丧失的这一过程。或许我们一开始应该制造出第一推动力,这种第一推动力是否就是作者的意向性?并等待其自我进化,因为人对世界的认识,人作为人的能力,都是在成长过程中逐步习得的。
题外的一点想法,如果肯定计算可以产生意识,从材料与计算的关系出发,等于说材质的介质影响了计算模型的性质。这似乎并不是完全不可能。这一方面的问题不知道有多深刻的研究。之前扫豆瓣的时候粗浅的感觉,似乎这本 Classical and Quantum Computation 会有一些想法?不知道计算理论那边会怎么叙述这一问题。毕竟计算模型及其能力能不能进行量化的比较,它们会不会构成一个连续统这一类的问题,已经超出我的能力之外了。等有时间应该回头反思一下。
中文屋在现实生活中或许是个伪命题?毕竟中文实际上显然是一门自然语言,运用一门自然语言的过程显然是不可能仅靠语法规则的。如果从语言习得的角度,更复杂的问题是“中文”的思考,但是众所周知,即使是真实的人,在初期掌握一门外语时也是不可能完成这一任务的。给予足够长的时间,则这种命题似乎还需再推敲。
Dreyfus的想法:只有大小和回路结构近乎完全相等的联结论系统才能模拟大脑。 - Is that real? Why? 如果成立的话,那么计算模型与计算模型间的差异将成为一个更加重要的问题。为什么是这个结果?

Lecture 1 Introduction (2020.09.17&09.24)

  • Human Visual Systems (HVS)

    • Optical Receptors: Image Information
    • Visual Pathway: Encoding, Representation
    • Primary Visual Cortex: Interpretation

Vision as a source of semantic information (top-down/down-top, difference between man and computer)

  • External Challenges in Computer Vision

    • Viewpoint variation: Machine's lack of robustness to rotation.
    • Illumination variation: Diffrence
    • Scale variation
    • Deformation
    • Background Clutter: subtle difference
    • Motion
    • Object intra-class variation
    • $\dots$
  • Prerequistes

    • Linear Algebra
    • Calculus
    • Matlab Programming
    • C++ Programming

Lecture 2 Person Re-Identification (2020.10.10)

https://arxiv.org/pdf/1610.02984.pdf ?

It was one of the most hopeless period when I hiked to Nanhui twice. I failed to pass almost all the mocks in September, 2018. There were countless assignments and tests to be finished, crowded tutorial classes to be taken, and endless scoldings to be suffered from. National Day in 2018 thus was destined to be a hard one.
However, I was fortunately gifted with several short breaks at that time, one of which was spent in Pudong Airport.
Though located at the seaside, uninhabited and remote, Pudong Airport is among the largest airport in China Mainland. Highways are thronged with vehicles. Green Carriages of Metro Line 2 and the famous Maglevs in Shanghai occasionally rush past.
7.JPG
I have sincere affection for Pudong Airport. My 'first love' left for America for a short-term exchange program years ago. I couldn't see her off personally in Pudong Airport. Later in my senior school, I daydreamed for a long time that I can jump into a plane and escape from myself in the airport.
Night had felt. The giant-like streetlights had been lit. Its dazzling light went through the mist, adding illusion to the scene. Night in October in Shanghai shouldn't have been that cold. Trembling, I left the metro station, and arriving at the huge airport.
People were busy catching their own flights. I ordered a cup of coffee and sat there, watching the passengers. It was in a sudden that loneliness rose in my heart. I had nothing to chase, but daily routine to follow. I arrived at the airport, only to find that there was no proper flight for me. Maybe my situation was much worsen than Nora Helmer.

Reference Books of Sichuan University's Sanskrit Course

  1. 阅读梵文佛典/Reading Buddhist Sanskrit Texts (English Version)
  2. 梵文基础读本/Elementarbuch Der Sanskrit-Sprache (Chinese Version)
  3. 梵语佛经读本 (Chinese Version)

Ancient Greek

  1. From Alpha to Omega
  2. 《古希腊语入门教程》

Latin

  1. Wheelock's Latin
  2. Workbook for Wheelock's Latin
  3. Ecclesiastical, Medieval and Neo-Latin Sentences
  4. Wheelock's Latin Reader, 2e
  5. Lingva Latina
  6. Collins Latin Dictionary & Grammar

引用者注:猜测原作者写作本文大致写作于 2002 - 2008 年间(北京工业大学实验学院在 2004 年以后迁出东方大学城,Windows XP在 2001 年以后发布,文章中提到的《面向计算机科学的数理逻辑》出版于2002年1月。),读者应该注意其中的很多推荐教材已经开始过时。关于作者身份,猜测是北京工业大学的计算机科学与技术系的毕业生。引用者根据目前的实际情况,适当进行整理与勘误,调换了一些章节的顺序。

- 阅读剩余部分 -