Collections of papers and articles that I’ve spotted this week that seem interesting. Comments on some of them.
- Sinnot-Armstrong, W. The Disunity of Morality
- There is no fixed region of the brain unique to moral judgement
- Simple explanations for morality (It is for suppressing self-interest, it is for cooperation, it serves X social purpose) are wrong
- Morality is like memory: many components to be individually studied.
- Rowley, D.A. et al. Counter-intuitive moral judgment following traumatic brain injury.
- Some papers say that injury to a certain area of the brain causes more utilitarian judgements because it lowers emotional response.
- Actually, no. It causes counterintuitive judgements. Intuitive utilitarian judgements are reduced in these patients alongside intuitive deontological judgements.
- Caveat N=31, and they do not study an individual damaged brain region
- Lubinski, D. From Terman to Today, a century of findings on intellectual precocity
- Berger, J. Are luxury brand labels and green labels costly signals of social status? An extended replication
- Signaling is not about signaling
- Wiltermuth, S.S. Creativity in unethical behavior attenuates condemnation and breeds social contagion when transressions seem to create little harm.
- Highly creative unethical behaviours are less punished by people
- Shepherd, J. The moral insignificance of self-consciousness
- Skimmed, not convinced
- Phillips, J. et al. True happiness: the role of morality in the folk concept of happiness
- Seth, Anil. The neuroscience of consciousness
- Until recently, psychology and neuroscience ignored consciousness
- The brain basis of consciousness is still a mystery, but an accessible one, as the brain is inside our heads.
- Easy/hard problems of consciousness. Seth introduces a ‘real problem’, uses analogy to vitalism (elan vital). This analogy has always seemed to me quite bad. He proposes to split the issue into pieces. Levels of consciousness, being conscious of things, being conscious os self. It still sounds like a new formulation of the easy problem.
- Generation of consciousness is not simply about a particular region, or about the number of neurons. Consciousness is related to how brain regions ‘talk’ to each other, information propagation. Conscious states are less compressible (in an information theory sense). It is possible to assign a number to this capacity for compresibility. This should correlate with awareness. Highly important to be able to numerically and reliably measure a phenomenon one wants to study
- Remark re 22.53 and cognitive impenetrability. Focusing hard enough on patches A and B does make it possible (for me) to see them as the same colour. I expect that you will also be able to do it.
- Expectations modify conscious experience
- DOG VISION! (Around 39.10)
- Christiano, P. Decisions and desiderate for AI control
- Yin, W. et al. Comparative study of CNN and RNN for Natural Language Processing
- In the street, you are meant to heuristically use CNN for images and RNN for NLP
- Here’s an actual comparison
- Disagreements between Paul Christiano and MIRI researchers on alignability of ‘messy AI’
- Generative Adversarial Networks in 50 lines of code
- Creating Human level-AI, how and when
- Not that much interesting if you have been following the area
- Deep Learning Paper reading roadmap
- Includes links to full Goodfellow-Bengio Deep Learning texbook!
- Entrevista con Jesús Fernández Villaverde
- Hafer, R.W. New estimates on the relationship between IQ, economic growth, and welfare
- What you expect
- The Decentralized yet Durable Empire (Holy Roman Empire)
- Really cool look into the inner workings and history of the HRE