Profile cover photo
Profile photo
Kingson Man
We must know. We will know.
We must know. We will know.

Kingson's posts

Post has attachment
a fascinating analogy between fluid dynamics and quantum mechanics, which might account for those spooky effects in a non-spooky deterministic way, and so dispel all those "you don't understand it, you just get used to it" non-explanations. and the video is excellent.

also some love in there for David Bohm

Post has attachment
$2 mil should be enough to cover the shrimp cocktail at the planning meetings

Post has shared content

Post has attachment
Or just circumvent it all by calling it a terrorism-related investigation

Post has shared content
Google I/O: Biologically inspired models of intelligence

Tune into #io14  at 3PM PST today to hear Google Director of Engineering +Ray Kurzweil  ( speak on the exploration of developing artificial intelligence based on biologically inspired models of the neocortex, enhancing functions such as search, answering questions, interacting with the user, and language translation.

Post has shared content
Photographed by Sho Iwai
@Tokyo Street 

Post has shared content
Mining open data from home, 15 year old wins Intel science prize.

From +Intel: "Because of his age, Nathan Han had difficulty finding a Boston area research lab willing to overlook liability issues and allow a fifteen-year-old to conduct research in its facility. But that didn’t stop the high school freshman from forging ahead on his own. Han was determined to study mutations of a gene linked to ovarian cancer, after learning that a family friend had been diagnosed with the disease. The upside for Han? The kind of research he wanted to conduct had far more to do with computer technology and data than traditional lab-ware such as test tubes and petri dishes. So, working from home, Han mined data from publicly available [i.e. open-access] databases and developed a machine learning software tool to study mutations of a gene linked to many kinds of cancer, including ovarian. The BRCA1 tumor suppressor gene has an important function as it is vital in protecting cells from developing cancer. Han detailed the characteristics of this gene and known mutations in order to “teach” his software to differentiate between mutations that cause disease and those that do not. In testing, Han’s device exhibited an 81 percent accuracy rate, demonstrating its potential as a breakthrough tool for accurately identifying cancer threats from BRCA1 gene mutations. Applications for Han’s research are vast, extending to research on other diseases, as well as to advancements in the developing fields of genomics, bioinformatics and big data. For this work, Han beat out nearly 1,800 other global finalists competing at the 2014 Intel International Science and Engineering Fair (Intel ISEF), a program of Society for Science and the Public, when he won the Gordon E. Moore Award, including a $75,000 scholarship."

Comment. I love this. First, it's a beautiful example of initiative and ingenuity. Congratulations to Nathan Han. Second, it's a compelling example of the benefits of open data. This kind of research by a 15 year old, working from home, without an affiliation at a well-endowed institution, would be impossible if the relevant data were locked behind paywalls. It would be equally impossible for a 45 year old. For researchers affiliated with well-endowed institutions, the research would still be impossible if the relevant data were sitting offline on a private hard drive. It reaffirms one of +Rufus Pollock's arguments for open data: the best use of your data will be discovered by someone else. 

#oa #openaccess #opendata #mining

Post has attachment
the "emotional attentional blink" - i love it! reduced sensitivity to fearful faces right after the end of each heartbeat 

"fearful faces were detected more easily and were rated as more intense at systole than at diastole. Correspondingly, amygdala responses were greater to fearful faces presented at systole relative to diastole"

Garfinkel et al. 2014 J Neurosci "Fear from the Heart: Sensitivity to Fear Stimuli Depends on Individual Heartbeats"

Post has attachment
using cross-validation to estimate your classifier's accuracy? better test it against a permutation distribution, and not the binomial distribution 

i've always wondered if reusing data so efficiently (lots of CV folds) might subtly introduce some non-independence. seems so. 

Noirhomme et al. 2014 Neuroimage Clin "Biased binomial assessment of cross-validated estimation of classification accuracies illustrated in diagnosis predictions."

Post has shared content
"infuriating" is about right:

"The important thing to understand here is that we do not need journals to publish anymore. Instead, they are being used by scientists to perform quality control and to filter out bad science. For this service we:

"pay publishers extortionate subscription fees or OA article publishing charges if we choose (or are forced to choose) the gold open access model and open access options;accept that our work will be locked in the drawers of editors and reviewers for months or even years;grant publishers all rights to disseminate and make profit from our work;offer our review services for free and without any academic recognition;allow all the qualitative information provided by reviewers’ thoughtful reports to be condensed into a digital yes or no decision, and hide them from the public;subject ourselves to high rejection rates and spend valuable time re-formatting the same paper over and over again to comply with different publication guidelines;split our research into many different papers to accumulate more publications or to satisfy page count criteria;remove colour from figures to keep publication costs down;compress the methods section, thereby depriving the public of important details needed for reproducibility;adapt and self-censor our research and writing style to accommodate the tastes of journal editors;throw away important negative or seemingly less-significant experimental results;miss out on the chance to have a constructive dialogue and even collaborate with reviewers to advance the work; andfeel obliged to investigate hot and sexy topics rather than exciting phenomena at the fringes of a field, where paradigm change is often found."
Wait while more posts are being loaded