Your Notebook exported from Why Machines Learn: The Elegant Math Behind Modern AI is
https://www.goodreads.com/notes/195887568-why-machines-learn/114528832-mark-gerstein?ref=rsp
Posts Tagged ‘quote’
iPhone Notebook export for Why Machines Learn: The Elegant Math Behind Modern AI
July 30, 2025iPhone Notebook export for Why Nobody Understands Quantum Physics: The bestselling guide to our universe
July 30, 2025Your Notebook exported from Why Nobody Understands Quantum Physics: The bestselling guide to our universe is
https://www.goodreads.com/notes/218376411-why-nobody-understands-quantum-physics/114528832-mark-gerstein?ref=rsp
My Neighbors Smoke Weed All Day, and It’s Stinking Up the Joint – The New York Times
April 11, 2025Ramos, J. T. (2025, March 29). My neighbors smoke weed all day, and it’s stinking up the joint. The New York Times.
https://www.nytimes.com/2025/03/29/realestate/marijuana-odor-apartment.html
QT:{{”
Q: I am a shareholder in a northern Manhattan co-op. New neighbors have moved in and they smoke marijuana all day, starting early in the morning. The smell is very strong, permeates the hallways, and makes its way into my unit, which is down the hall and around the corner. I know they have a right to smoke in their apartment, but the smell is difficult to tolerate….
A: You can certainly write a letter to your neighbors, asking them to better ventilate their unit. But that might not solve the problem….
You can try to protect your apartment by adding weatherstripping to your door and using an under-door draft stopper, said Mark Foley, president of the Folson Group, a business consultant for condo and co-op buildings. You could also offer to install these things for your neighbors’ door, he said.
“}}
How Your Family Tree Could Catch a Killer | The New Yorker
March 13, 2025The woman featured, CeCe Moore, is quite clever, going far beyond simply examining genealogical matches to track potential suspects.
parabon-nanolabs.com
https://www.newyorker.com/magazine/2021/11/22/how-your-family-tree-could-catch-a-killer
QT:{{”
Genealogists grew interested in genetics at the turn of the
millennium, when it became possible to analyze bits of information from the Y chromosome—known as Y-DNA—on a commercial scale. Because the Y chromosome is passed from father to son with little mutation, and because surnames historically were passed down the same way, it seemed worth exploring whether the confluence could be useful to researchers. In the late nineties, Bryan Sykes, an Oxford geneticist, persuaded forty-eight men who shared his surname to take Y-DNA tests. “Sykes” comes from a Middle English word meaning “spring” or “stream,” and the name was thought to have arisen separately among unrelated families that lived near various sources of water. But the genetics suggested that the men descended from a single ancestral line. “If this pattern is reproduced with other surnames, it may have important forensic and genealogical applications,” Sykes concluded.
Theoretically, researchers could use Y-DNA to establish the pedigree of a man with an unknown identity. Sykes made a similar case for mt-DNA, which is passed down on the maternal line, in a book titled “The Seven Daughters of Eve.”
….
The first step was to establish a DNA profile for the adoptee in a database like GEDmatch, to look for partial genetic matches with other users. The people linked with those matches were not always easy to identify; some users logged on without any personal information or, worse, under aliases. But, when the genealogists succeeded, they could trace back family trees until they identified common ancestors. Then they would reverse the process: starting from the common ancestors, they would build a complete tree of all the descendants, knowing that the adoptee’s parents had to be among them. The amount of DNA that the adoptee shared with matches in the database was a key clue to where he or she belonged in the larger tree; personal details, like birth dates and geography, could also provide clues.
“}}
Media query (Nature Medicine): Quantum computing and health
January 19, 2025Guenot, M. (2025). Can quantum computing crack the biggest challenges in health? Nature Medicine. https://doi.org/10.1038/s41591-024-03369-w
Great story by @Marianne_Guenot, providing good & *not* good news about QC for biomedicine.
QT:{{
The potential power of quantum computers in cracking problems that classical computers cannot is not all good news, says Mark Gerstein, a professor of biomedical informatics at Yale University who recently co-authored a review about quantum computing and health for Nature Methods3.
Experts predict that quantum computers could become fiendishly good at breaking through current encryption algorithms, says Gerstein, which could pose a problem for the privacy of confidential patient data. “There’s a huge push right now to get post-quantum cryptography to work,” he says.
…
The idea, then, is not for quantum computers to replace classical computers, says Gerstein. Instead, they should be considered as adding a node to a computing chain, as each can contribute different strengths to solve a problem. “The art here is figuring out which bit of this big calculation you can quantize,” says Gerstein.
“}}
Turning on pseudogenes | Interviews | Naked Scientists
January 3, 2025QT:{{”
Kat – Do you think that there are other pseudogenes lurking the genome that have this kind of roles that are actually not drunk and not dead, and could be very active and important?
Howard – Yes, indeed. So, I think over the last several years, the other investigators have found have resembled from the ENCODE project that many pseudogenes are actually being transcribed, that they’re made. They’re evidence of their activation and that work was led by a Professor Mark Gerstein from Yale. More recently, other people have realised that pseudogenes because they’re copies of normal genes, they have many of the same regulatory sequences embedded in them. And so, when the pseudogenes are made, because of these regulatory sequences, compete for different cellular factors.
“}}
https://www.thenakedscientists.com/articles/interviews/turning-pseudogenes
For Her Scientist Husband, an Architect Creates a Boldly Contemporary Retreat
December 26, 2024A private study
QT:{{”
asked her to design a studio for him on their weekend property in upstate New York….studio, intended solely as his private refuge for reading, writing and thinking, she wanted the structure to be as open to its setting as possible. Located near a cliff above a large pond, it is surrounded by woodland and meadows.
“}}
For Her Scientist Husband, an Architect Creates a Boldly Contemporary Retreat
Source: Architectural Digest
A Refuge in the Woods
The state of academic publishing in 3 graphs, 6 trends, and 4 thoughts | Dynamic Ecology
December 22, 2024QT:{{”
Publishing is growing exponentially – While the number of scientists is also growing exponentially, it is at a slower rate than papers. We are producing more papers per scientist every year. This is a profoundly important fact. Every ecologist knows the power and unsustainability of exponential growth. This also makes it abundantly clear that the publishers only deserve half the blame. Scientists have created a Red Queen situation in which we’re aggressively chasing opportunities to publish. Do we really need 1,000,000 (about +40%) more publications than 10 years ago! (Figure 1).
“}}
Variation in human water turnover associated with environmental and lifestyle factors | Science
November 10, 2024from podcast
QT:{{”
….Okay. So when you got measurements from this big group
of people, you end up
with kind of a range of water turnover and it varies by person and location. So how does this range
that you measured in this diverse group of people? How does it stack up with eight, eight ounce
glasses a day, which is like two liters of water a day, that recommendation we discussed?
Most people are not going to need to drink, eight
glasses of water a day, two liters of
water a day. If you measure how much water flows through your body, how much water comes in
and goes out every day, there’s a lot of variation, but it’s something like three to four liters a day
total. And that includes not just the water that you drink, but that includes the water that’s in the
food that you eat.
“}}
The study of ancient DNA is helping to solve modern crimes
November 2, 2024v. related to plight + deconvolution
QT:{{”
Dr Willerslev had been able to obtain SNPs from Meng’s trousers. Then came the task of trying to identify whom they could have come from. In 2023 a man called Philip Patrick Westh was arrested in connection with a kidnapping case in the same area; because of similarities between the cases, the police believed that he had killed Meng too (Mr Westh denies most of the accusations related to the kidnapping case and pleaded not guilty to the charge of killing Meng). To assess the probability that the genetic material from Meng’s trousers had come from Mr Westh, Dr Willerslev made use of a DNA database of ordinary, healthy Danes. If the SNPs found on the trousers were identical to those in Mr Westh’s DNA, went the logic, and enough of them were sufficiently rare variants, the probability that they did indeed come from Mr Westh went up. Dr Willerslev testified that this particular pattern of SNP variants would be at least one million times more likely to turn up if the sample included DNA from Mr Westh, or a close relative, than if it did not.
…
But identifying a suspect from SNPs is another matter. Degraded samples can be extremely complex to analyse, says Dr Simonsen, in particular if material from several people has been mixed together. There are no standardised forensic protocols for separating out those signals, nor for how to confidently calculate the probability that the DNA belongs to a suspect. That matters because the stakes are somewhat higher in a criminal case than in the study of mammoths and dodos. But, says Dr Simonsen, “We expect that nut to be cracked.” He hopes to learn from Dr Willerslev’s team and develop new forensic tools. The work has already started: in September, researchers from Aalborg University and the University of Copenhagen, with whom he
collaborates, published a paper describing an approach for doing identification calculations based on SNPs.
He is not the only one to see potential. An American company called Astrea Forensics has recently spun out from the palaeogenomics group at the University of California in Santa Cruz, to offer aDNA expertise to law enforcement. Their speciality is the nuclear DNA found within hair, which has long been considered too fragmented and scarce to be of any use.
“}}
linkstream2 microblog 