designing artificial minds
DESCRIPTION
Designing Artificial Minds. Harri Valpola. Computational neuroscience group Laboratory of computational engineering Helsinki University of Technology http://www.lce.hut.fi/~harri/. “The great end of life is not knowledge but action” — Thomas Henry Huxley (1825-1895). - PowerPoint PPT PresentationTRANSCRIPT
![Page 1: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/1.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Computational neuroscience group
Laboratory of computational engineering
Helsinki University of Technology
http://www.lce.hut.fi/~harri/
Harri Valpola
Designing Artificial Minds
![Page 2: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/2.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
“The great end of life is not knowledge but action”
—Thomas Henry Huxley (1825-1895)
![Page 3: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/3.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Sea Squirts — Our Distant Cousins
• Sea-squirts are common marine animals
• Two stages of development: larva and adult
• Larvae look very much like tadpoles
Picture:
http://www.jgi.doe.gov/News/ciona_4panel.jpg
![Page 4: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/4.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
No Movement No Brain
Picture:
http://www.gulfspecimen.org/images/LeatherySeaSquirt.jpg
![Page 5: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/5.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Smooth, Elegant, Skillful
Picture:
http://www.africanbushsafaris.com/fotos%20touren/Oryx.jpg
Picture:
http://www.ebroadcast.com.au/blahdocs/uploads/tiger_running_sml_2784.jpg
![Page 6: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/6.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Evolution of the Motor System
• Spinal cord: “simple” reflexes and rhythmic movement
• Brain stem: more complex reflexes
• Cerebellum / midbrain structures: complex motor coordination
• Basal ganglia: action selection
• Hippocampal formation: navigation
• Neocortex: integration and planning
![Page 7: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/7.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Methods• Synthetic approach: learning by building• Neural network simulations• Real and simulated robots
![Page 8: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/8.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Early Motor System
• Spinal cord: “simple” reflexes and rhythmic movement
• Brain stem: more complex reflexes
• Cerebellum / midbrain structures: complex motor coordination
• Basal ganglia: action selection
• Hippocampal formation: navigation
• Neocortex: integration and planning
![Page 9: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/9.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Prediction and Anticipation
![Page 10: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/10.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Adaptive Motor Control Based on a Cerebellar Model
![Page 11: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/11.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Prediction and Anticipation
![Page 12: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/12.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Self-Supervised Learning in Control
• Corrections are made by a large number of “reflexes” (spinal cord, brain stem, cortex / basal ganglia).
• Cerebellar system learns to control using the reflexes as teaching signals.
Picture of cerebellar system
![Page 13: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/13.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Reflexes
Opto-kinetic reflex
Stretch reflex
Picture:
http://www.cs.stir.ac.uk/courses/31YF/Notes/musstr.jpg
Picture:http://www.inma.ucl.ac.be/EYELAB/neurophysio/perception_action/vestibular_optokinetic_reflex_fichiers/image004.jpg
![Page 14: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/14.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Robot “Reflex”
![Page 15: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/15.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
The Cerebellar System
Picture of cerebellar cortex
![Page 16: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/16.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Long-Term Depression (LTD) Guided by Climbing Fibres
Picture of LTD in Purkije cells
![Page 17: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/17.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Vestibulo-Ocular Reflex
Picture:
http://www.uq.edu.au/nuq/jack/VOR.jpg
![Page 18: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/18.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
System-Level Computational Neuroscience
Questions to be answered:
• What kind of components are needed for a cognitive architecture?
• What are different algorithms good for and how they can be combined?
The brain is a good solution to these questions Try to
understand its algorithms on system level (level of behaviour)
![Page 19: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/19.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
“Without knowledge action is useless and action without knowledge is futile”
—Abu Bakr (c. 573-634)
![Page 20: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/20.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Components for a Cognitive System
Basal ganglia: selection, reinforcement learning (trial-and-error learning)
Hippocampal formation: one-shot learning, navigation, episodic memory
Neocortex:• Represents the state of the world — including oneself• Invariant representations, concepts• Attention / selection (both sensory and motor)• Simulation of potential worlds = planning and thinking• Relations and other structured representations (akin to
symbolic AI)
![Page 21: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/21.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Neocortex
• A hierarchy of feature maps: increasing levels of abstraction
• Bottom-up and top-down/lateral inputs treated differently
• Local competition• Long-range reciprocal
excitatory connections
Picture:http://www.pigeon.psy.tufts.edu/avc/husband/images/Isocrtx.gif
![Page 22: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/22.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Representations for Natural Images by Independent Component Analysis (ICA)
http://www.cis.hut.fi/projects/ica/imageica/
• ICA is an example of ..unsupervised learning.
• Can learn something like .. V1 simple cells.
![Page 23: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/23.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Invariant features• Group simple features into complex in a hierarchical
model.
Picture:
http://cs.felk.cvut.cz/~neurony/neocog/en/images/figure3-1.gif
![Page 24: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/24.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
”Complex Cells” from Images
![Page 25: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/25.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Abstractions and Meaning
![Page 26: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/26.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Abstractions and Meaning
• Once we have motor output, we can learn which information is important and meaningful
Left camera Right camera
![Page 27: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/27.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Relevance to Human Enhancement
How about mind prostheses? New senses (like web-sense)?
Knowing how the brain works would certainly be useful for prostheses, but for healthy persons…
• Input to the brain is easiest to deliver through existing senses — content matters, not the channel
• Output from the brain through motor system is more limited implanted electrodes might surpass this capacity
• I expect intelligent tools to be far more common than prosthetic devices for a long time, but this doesn’t mean their societal impact would be any smaller
![Page 28: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/28.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
Brain is a good solution for an engineering problem
Neuroscience
Technology
Picture:
http://britton.disted.camosun.bc.ca/escher/drawing_hands.jpg
![Page 29: Designing Artificial Minds](https://reader036.vdocuments.us/reader036/viewer/2022062409/56814dcd550346895dbb24f0/html5/thumbnails/29.jpg)
August 19th, 2006 Computational Neuroscience Group, LCE Helsinki University of Technology
KIITOS! THANK YOU!