Brain Computer Interface
Since the beginning of time people have been interested in finding out how the brain works -
see slides on Ray Kurzweil's book How to Create a Mind.
Our understanding of the mind has improved through studying artificial intelligence, cognitive science, neuroscience, and robotics.
This project involves understanding of the existing digital human brain models created so far based of computer science perspective.
This project involves the following steps:
This semester's extension of the project will involve:
- Write a clear, concise, and reasonably complete description of the human brain (about 2.5 double-column pages).
- Generate a table enumerating the major simulations (true simulations,
not artificial neural networks) of the human brain with a brief description of each (about 2.5 double-column pages).
- Generate a table enumerating the major deep learning neural network models that come close to simulating the functionality
of the human brain with a brief description of each (about 2 double-column pages).
- Compare and contrast the simulations of step 2 with the models of step 3 (about 0.5 double-column pages).
- Based on the knowledge gained from the above, describe three research studies you think would be interesting to further explore this area (about 0.5 double-column pages).
The outcome of this project should provide an excellent background and
literature review for existing brain models, deep learning or other studies in this area.
- Investigate and describe the relationship between deep learning and brain architecture, see
- More detail about how the human brain works
- Investigate and understand different human brain models
This project is basically a literature review project and involves no programming.
The students selecting this project must have high proficiency in the English language, and excellent library research and writing capabilities.
Below are some links of some existing Brain/Mind Models:
Below are articles and papers on Brain/Mind Models and long short-term memory(LSTM)
- Blue Brain project: http://bluebrain.epfl.ch/
- Brain Mind Institute (BMI) https://sv.epfl.ch/BMI-home
- IBM - The Computer Brain: http://www.research.ibm.com/cognitive-computing/neurosynaptic-chips.shtml#fbid=um6spNAW15a
- Google DeepMind: https://deepmind.com
- Facebook F8 https://www.facebook.com/careers/teams/building8/
- China Brain Project: https://en.wikipedia.org/wiki/China_Brain_Project
- Hierarchical Temporal Memory (HTM): https://en.wikipedia.org/wiki/Hierarchical_temporal_memory
- Ray Kurzweil's Pattern Recognition Theory of The Mind (PRTM): https://en.wikipedia.org/wiki/How_to_Create_a_Mind
Additional references (previous Research Day Conference papers):
- A Working Brain Model https://www.technologyreview.com/s/409107/a-working-brain-model/
- A Motivational System For Mind Model CAM (Consciousness And Memory Model): https://www.aaai.org/ocs/index.php/FSS/FSS13/paper/viewFile/7476/7537
- A Machine Consciousness Approach to Autonomous Mobile Robotics https://www.aaai.org/Papers/Workshops/2006/WS-06-03/WS06-03-017.pdf
- Getting Up to Speed on Vehicle Intelligence: https://aaai.org/ocs/index.php/SSS/SSS17/paper/viewFile/15322/14599
- How to Build a Digital Brain:
- Convolutional, long short-term memory, fully connected deep neural networks: https://research.google.com/pubs/archive/43455.pdf
- Long short-term memory based recurrent neural network architectures for large vocabulary speech recognition: https://arxiv.org/pdf/1402.1128
- Unidirectional long short-term memory recurrent neural network with recurrent output layer for low-latency speech synthesis: https://research.google.com/pubs/archive/43266.pdf
- Sequence discriminative distributed training of long short-term memory recurrent neural networks: http://188.8.131.52/~czap/letoltes/IS14/IS2014/PDF/AUTHOR/IS141312.PDF
2017 Research Day Conference, Paper c6
2017 Research Day Conference, Paper c9
We recently submitted a Brain-AAAI Conference Paper to
2017 Symposium on ĎA Standard Model of the Mindí
which was prepared hurriedly at the last minute and therefore rejected.
For those who are interested in coding, we recommend coding the Rosenblatt's Brain Memory Model
(Dr. Frank Rosenblatt was the inventor of perceptrons).