Personal Science Week - 8 Sept 2022
AI for science, back pain, and more
Back Pain
Steven Byrnes at Lesswrong writes “The Mind-Body Vicious Cycle” an excellent overview of an intriguing theory about the relationship between chronic back pain and the mind. Anyone who suffers from back pain in the modern West will be familiar with the “orthodox model” that assumes the pain has a physical cause: an out-of-place bone or muscle, often caused by an injury or poor posture. But a small number of physicians have long argued that the pain is directly related to the mind, not necessarily in the psychosomatic sense (“you’re just imagining it”) but perhaps in a deeper way that relates to how the nervous system is wired.
Byrnes summarizes the various theories, particularly those of John Sarno, a rehabilitation professor at New York University. Sarno, who died at age 93 in 2017, believed that all back pain originated in the mind and that it could often be cured by simply understanding that fact.
If you or someone you love suffers from chronic back pain, Bryne’s summary and the related references is a good way to understand an alternative path to a solution.
Biohacking references
Hacking Age, a review article in the Wiley publication Sociology Compass, is an up-to-date (2022) and concise summary of the biohacking world, including the history and key figures like Dave Asprey and others. Mentions uBiome and other “bio-metric” companies. If you’re looking for a good overview of the industry/movement, this is your source.
Applying AI to Science
How to build a GPT-3 for Science wonders what would happen if the same technology underlying GPT-3 could be harnessed for scientific research. GPT-3 and its sister version for images, DALL-E, work by scraping billions of sample documents and images to build a sophisticated model that generates results that bring auto-complete to a whole new level. Like the generator in a search engine or your favorite messaging app, these technologies can extrapolate from a few initial hints to guess what would happen if you typed out the rest of the phrase – or document. Many of the resulting texts and drawings have an uncanny and eerie resemblance to something you’d expect from a human.
In principle, the same idea could work for science. Imagine typing “Write a scientific paper based on my data” – and the system generates what you expect.
Unfortunately, despite decades of heavy data-centric and computer use by scientists, most of the data required to build a “GPT-3 for science” is locked inside publisher paywalls. Even the papers that are open often don’t provide the raw data required to reproduce them, and data that is available is often in an inscrutable format understood only by the lone researcher who generated it.
This is a solvable problem, and like most solvable problems with important applications, it will inevitably be tackled.
For the personal scientist, I’d like to add one more feature if I may. What if, instead of the static plots based on the data collected for that paper, what if you could enter your own data into each paper? A study on, say, the relationship between a particular diet type and weight loss for a specific genotype, could show how you fit into the data. You may find that you’re one of the outliers, to whom this study has no relevance. Or alternatively, you might find the results to be extremely relevant – and lead to even more studies.
About Personal Science
Personal Science is the process of using the scientific method to solve problems and get better results on an individual, personal level. Following the motto of the Royal Society, established in 1660, nullius in verba, we take nobody’s word for it.
This newsletter is a weekly summary of a few observations we think will be interesting to anyone who wants to be a personal scientist. Press the button below to share a (free) subscription with others.