課程資訊

Computational Cognitive Neuroscience

106-1

EE5156

921 U4370

3.0

Ceiba 課程網頁
http://ceiba.ntu.edu.tw/1061EE5156_

.注意:本課程將需主動投入甚多時間搜尋閱讀可能超過課程進度或範圍之大量且結論可能互相矛盾的參考書籍, 文獻, 及網路資源; 且必須撰寫作業報告及分組完成期末專題實作, 兩者均需完成認知行為之模型實作, 並能討論其成果. 課業繁重, 學習較被動, 無強烈決心及毅力之同學請慎重考慮.

The BRAIN Initiative (http://www.nih.gov/science/brain/,
http://en.wikipedia.org/wiki/BRAIN_Initiative);

http://en.wikipedia.org/wiki/Human_Brain_Project);

The Blue Brain Project (http://bluebrain.epfl.ch/, http://en.wikipedia.org/wiki/Blue_Brain_Project );

SpiNNaker(http://apt.cs.manchester.ac.uk/projects/SpiNNaker/project/, S. B. Furber, et. al.
"The SpiNNaker Project," IEEE Proceedings, vol.102, no.5, 2014, pp 652-665; S. B. Furber et. al.,
"Overview of the SpiNNaker System Architecture," IEEE Transactions on Computers, vol. 62, no. 12, 2013, pp. 2454 - 2467)等.

1. 計算機程式設計: 本課程絕大部分的的範例, 作業, 期末專題等, 均使用Python程式語言. 修課同學不一定要學過Python, 課程中會有簡介, 但同學最好開學前就能對以下一般程式語言觀念有相當的了解: 變數, if...else,

D. Banas, Python Programming (Learn Python in One Video,約45分鐘), https://www.youtube.com/watch?v=N4mEzFDjqtA
.

2. 機率與統計: 例如probability density function, 期望值, mean, variance, Gaussian distribution, Poisson distribution, random variables,
conditional probability, Bayes定理等.

3. 認知神經科學: 作為計算模型須符合的限制(constrains), 但不是本課程唯一重點. 上課時會由基礎開始介紹, 所以沒學過也行.

1. 基礎教材: 盡量詳細講解, 並於上課時間內上完的內容
2. 參考教材: 同學自修以及準備作業和期末專題時可參考的內容
3. 延伸學習: 同學深入瞭解之途徑

Peter Harris, Designing and Reporting Experiments in Psychology, 3rd ed., In Open Guides to Psychology.Maidenhead : McGraw-Hill Education. 2008.

Robot Operating System 內的工具, 設計簡化的機器人軟體模型及其模擬環境. 隨後團隊即持續合作, 以Nengo及Python

Build a creature to seek food and return to a home nest
Build a serial working memory model that allows for loading, reset, etc.
Build a robot controller for a robot that you present instructions to, that it memorizes and then executes.
Build an adaptive controller for a 3-link arm using Slotine's adaptation methods.
Build a model that does any one of the tasks in Spaun.

Office Hours

[Computational Cognitive Neuroscience and Computational Neuroscience]
R. C. O'Reilly, Y. Munakata, T. Hazy, and M. J. Frank, Computational
Cognitive Neuroscience, 2nd edition, 2014.

title=CCNBook/Main

P. W. Glimcher and E. Fehr ed., Neuroeconomics: Decision Making and the
Brain, 2nd ed., Oxford, UK: Elsevier Inc. ( in association with the Society
for Neuroeconomics—www.neuroeconomics.org), 2014.

http://www.sciencedirect.com/science/book/9780124160088

B. Anderson, Computational Neuroscience and Cognitive Modelling: A Student's
Introduction to Methods and Procedures, London: SAGE, 2014. (所用數學及程式語

T. P. Trappenberg, Fundamentals of Computational Neuroscience, Oxford
University Press, 2010. 另附參考資源網址:
http://www.cs.dal.ca/~tt/fundamentals

D. Sterratt, B. Graham, A. Gillies, D. Willshaw,
Principles of Computational Modeling in Neuroscience
Cambridge University Press, 2011.

[Cognitive Neuroscience]
M S. Gazzaniga, R. B. Ivry, G. R. Mangun, Cognitive Neuroscience: The
Biology of the mind, 4th ed., New York: Norton, 2013.

D. Purves, R. Cabeza, S. A. Huettel, K. S. LaBar, M. L.
Platt, M. G. Woldorff, Principles of Cognitive Neuroscience, Sunderland,
MA, USA: Sinauer Associates, Inc.., 2013. 另附參考資源網址:
http://sites.sinauer.com/cogneuro2e/index.html

[Neural Science]
E. R Kandel ed., Principles of neural science, 5th ed., New York : McGraw-
Hill Medical, 2013.

[Cognitive Psychology and Psychological Science]
M. S. Gazzaniga, T. F. Heatherton, D. F. Halpern, Psychological
science, 4th ed., New York : W. W. Norton, 2013. 另附參考資源網

[Cognitive Science]
J. L. Bermudez, Cognitive science : an introduction to the science of
the mind, 2nd ed., Cambridge : Cambridge University Press, 2014.

John R. Anderson, How can the human mind occur in the physical universe?
New York : Oxford University Press, 2007.

[Philosophy]
Stuart M. Shieber, The Turing Test:Verbal Behavior as the Hallmark of Intelligence,
MA: MIT Press, 2004.

http://ieeexplore.ieee.org/xpl/bkabstractplus.jsp?bkn=6267336

W. R. Uttal, Mind and Brain: A critical appraisal of cognitive neuroscience,
Cambridge, MA: MIT Press, 2011.

[ROS]
M. Quigley, B. Gerkey, and W. D. Smart, Programming Robots with ROS:
A practical introduction to the Robot Operating System, Sebastopol, CA:
O'Reilly Media, Inc., 2015.

[推薦科普讀物]
R. Carter, Mapping the Mind, University of California Press, 2000.

C. Eliasmith, How to Build a Brain: A Neural Architecture for Biological
Cognition (Oxford Series on Cognitive Models and Architectures), Oxford
University Press, 2013.

http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780199794546.001.0
001/acprof-9780199794546

python 程式語言整合開發環境 PyCharm: https://www.jetbrains.com/pycharm/

Nengo (Neural ENGineering Objects) 認知神經科學模型開發工具: http://www.nengo.ca/

Nengo手冊: http://pythonhosted.org/nengo/
Nengo 範例: http://pythonhosted.org/nengo/examples.html

ROS (Robot Operating System) 機器人作業系統: http://www.ros.org/,

(僅供參考)

 No. 項目 百分比 說明 1. 兩次作業 60% 每次作業30%. 作業由任課教師批改, 評分要點為報告內容是否適當回應主題, 內容安排是否有創意且系統化, 自創想法是否有新意 且有證據支持, 邏輯論述是否清楚合理易懂, 相關文獻引用是否適當, Python/Matlab程式測試數值實驗安排及結果. 討論是否周延合理, 學習心得是否言之有物等. 請注意本課程與人工智慧等之理論或研究成果領域如Neural Networks, Deep Learning, Machine Learning等有所差距, 本課程作業需結合本課程內涵, 不宜將他處課程之作業原封不動交來. 同學若自修或向其他師長先進, 學長或同學學得超過課程進度之內容,並應用於自行完成之作業, 需敘明自修或討論學習過程於學習心得部份. 作業亦可利用其他既有之相關開放程式碼或無版權顧慮之軟硬體工具及網路上可公用之 數據, 影音, 圖片等資料, 直接或修改後採用, 以完成同學構想之主題, 但需註明出處及修改應用之處. 作業討論時間若抽到的修課同學未事先請假, 又不在教室內, 作業成績先九五折, 隨後由抽到之時間至該同學出現或下課, 每十分鐘將該份作業成績乘以0.95一次. 作業遲交時,成績先打九折,隨後自預定繳交日第一節開始上課時間算起,每逾一日(24小時),該次作業成績乘以0.85 一次. 若因另有要事無法出席作業討論者, 仍須於作業討論日上課第一節前上傳作業壓縮檔. 並須與授課老師商量, 於適當時間進行作業簡報. 2. 期末專題口頭報告及展演 20% 於期末專題發表日(期末考日)原上課第一節時間開始, 輪流抽籤決定報告團隊. 每個團隊都會上台, 簡報加系統展示(必要時得播放預錄影片)10分鐘, 持續到所有團隊報告完畢. 如果團隊數過多, 無法於4小時內結束, 任課老師可能 重新安排專題發表時間, 將其延長至一到兩整天. 簡報時其他聽講同學為報告團隊以1至5分評分, 並註記優缺點. 自己團隊報告時, 應迴避不予評分. 所評分數1至5分各分數所占人數應照報告當日宣佈之規定比例分配, 不可全部 打相同或只打少數特定分數. 各團隊之口頭報告及展演成績為其他同學評分截頭去尾後平均(x),換算得分為 70+5x, 缺席為0分. 因不可抗力原因如參加遊學團或其他課程之野外實習等, 須盡早通知任課教師請假. 每團隊於報告日當天, 至少要有一人出席簡報成果. 若因如上述原因, 致報告時間無人能參加簡報者, 須及早通知任課老師, 可安排於最後上課日 最後一節下課前簡報. 但要求之pdf檔與PowerPoint檔仍應於規定時間之前上傳.發表會後, 若確因不可抗力原因致團隊中無人能出席 發表會者, 須附證明, 並於報告日當晚12:00前以email通知任課老師, 安排第二日適當時間向任課老師簡報, 並由任課教師評分. 如仍無法簡報者, 由團隊與任課老師商討變通方法. 無論何種情形, 專題報告之Power Point 檔均應於規定時間前上傳. 遲交者不論原因, 成績打八折. 3. 期末專題書面報告 20% 由任課老師主觀依照書面報告之整體表現, 以 A+ = 95, A = 90, A- = 85, B+ = 80, B = 75, B- = 70 等評分, 缺交為0分. 不論原因, 遲交者成績打八折,期末專題發表日開始算起兩天(48小時)後,不再收補交之書面報告,該項成績為零分 請注意本課程與人工智慧等之理論或研究成果領域如Neural Networks, Deep Learning, Machine Learning等有所差距, 本課程期末實作需結合本課程內涵, 不宜將他處課程之term project原封不動交來.

 課程進度
 週次 日期 單元主題 第1週 9/15 1. Science of Cognition. Text: Sections 1.1 ~ 1.2 - The Science of Cognition: A brief history of cognitive science, Central features of the main approaches. Sections 1.3 ~ 1.4 - Challenges for cognitive science: Overview of the approach in this course. Extension: Cognitive Psychology and Python. 第2週 9/22 2. Neurons. Text: Sections 2.2 - An introduction to basic neurophysiology and anatomy: Extension: Spiking Models, Physiology and computational models of neurons. 第3週 9/29 3. Brains. Text: Section 2.3 - Levels of description in the behavioral sciences. Extension: Brain, Connectivity. 第4週 10/06 4. NEF Principle 1: Representation. Text: Sections 2.2.1 - Principle 1 of the NEF - representation. 第5週 10/13 5..NEF Principles 2: Computation Text: Sections 2.2.2 Principle 2 of the NEF - computation. 第6週 10/20 6. NEF Principle 3: Dynamics. Text: Section 2.2.3 - Principle 3 of the NEF - dynamics. 第7週 10/27 7. Perception and Semantic Pointer Hypothesis Text: Sections 3.1 - 3.4 Overview of the semantic pointer hypothesis, Distributed neural semantics. Text: Sections 3.5 - 3.7 An introduction to visual semantics, An introduction to motor semantics. Extension: Sensation, Visual system, Auditory system 繳交第一次作業. 第8週 11/03 8. Autoencoder. Extension: Deep Belief Network and Autoencoder. 第9週 11/10 9. Vector Symbolic Architectures and Structured Concepts Text: Sections 4.1 - 4.7 Syntactic representations, Vector symbolic architectures, Implementations of VSAs in neurons. Learning syntactic manipulations, Modeling fluid intelligence, Syntax and semantics for structured concepts. 第10週 11/17 10. Basal Ganglia, Thalamus, and Cortex: Physiology. Text: Sections 5.1-5.3 Basal ganglia anatomy and physiology, Basal ganglia function. 第11週 11/24 11. Basal Ganglia, Thalamus, and Cortex: Models. Text: Sections 5.4, 5.6-5.8 Basal ganglia use for flexible action selection, Example uses of the basal ganglia model in the SPA. Extension: Decision Making. 繳交期末專題實作分組名單及實作主題簡述. 第12週 12/01 12. Memories. Text: Sections 6.1-6.3 Introduction to cognition through time, Working memory and serial working memory. Extension: Short-time and long-time memory. 繳交第二次作業. 第13週 12/08 13. Learning in General Text: Sections 6.4 Spike-timing dependent plasticity (STDP), Reinforcement learning, Learning transformations with the hPES rule. Extension: Learning in General. 第14週 12/15 14. Reinforcement Learning Text: Sections 6.5-6.6 Learning New Actions Learning New Syntactic Manipulations Extension: Reinforcement Learning 第15週 12/22 15. Spaun Model. Text: Sections 7.1-7.3 A review and overview of the SPA, The Spaun model and tasks. 第16週 12/29 16. Emotions Neirophysiological and psychological views Computation Models Extension: Emotions 第17週 1/05 17. Languages Neirophysiological and psychological views Computation Models Extension: Languages 第18週 1/12 (期末考週) 期末專題實作成果發表會