Profile

Cover photo
Andrea Casalotti
Attended Insead
Lives in London
279 followers|446,048 views
AboutPostsCollectionsPhotos

Stream

Andrea Casalotti

Shared publicly  - 
 
Yonatan Zunger originally shared to Today I Learned::
 
One of the biggest challenges in information retrieval (the branch of computer science that includes search and content recommendation) is how to find good content which humans haven't already found. To date, the most reliable signals have been other human judgments: for example, PageRank is a measure of how "good" a site is based on links people have made to that site (with the challenge being how to separate "meaningful" and trustworthy links from the rest), and collaborative filtering is based on what other users have chosen (with the challenge being how to find users with similar enough taste to be relevant).

The challenge is that, when new material shows up on the scene, you don't yet have any human interactions -- and quite often, good material, things people would love, simply goes unnoticed and never builds up the interaction signals which help. To detect quality in these things requires understanding the content itself, and the aspects of it which matter to people.

There are several hard aspects to this. One is simply understanding the content at the right granularity: "the color of the top-left pixel" or "the frequency of the word 'whenever'" are too fine-grained to give us a hint about whether people will like something, so we need to be able to group the content into more meaningful structures. For images, that might be "an image of a face in 3/4-profile," a certain color balance or contrast, a perspective or a cropping, and advances in image recognition in the past few years have (finally) made it possible to reliably identify such features. For text, it's much harder: there isn't yet even a clear idea of what features both could be measured about text and determine people's tastes. (How do you measure "intellectually meaty" or "hinting at scandal?")

This paper has used the recent advances in image processing, together with recent advances in AI in general, to get a sense of which pictures people will like. It started by taking several thousand images, and having them rated by humans for quality; that was used as "ground truth." Then, those thousands of images are analyzed into meaningful features, and a neural network is trained to find patterns of image features which predict human taste.

This is what neural networks, and other kinds of "supervised" machine learning systems, do in general: they take as inputs a bunch of signals, and combine them using a large number of parameters -- the "weights" -- to produce predictions of some values that you want to measure. The weights are set by taking a large number of test examples ("golden data" or "ground truth") with known values of both the signals and the test values; weights are chosen ("trained") to maximize the quality of the system's predictions for this data. To make sure that the training doesn't just teach it to recognize those specific examples, the golden data is randomly split into two groups; one is used for training, and then it's tested against the other group to make sure that the predictions with the trained weights are good. If they are, then you have a model which can predict -- given any set of measured signals -- the truth values.

In this case, the signals are these features of the image, measured by a second machine learning system; the quantity being predicted is whether people will like it. Because these are all "content-based signals" -- that is, they're based on the contents of the image, and not on people's responses to it -- the resulting model can be applied to any image. 

The team then applied this model to a set of 9 million images from Flickr with fewer than five "favorites." They tested the quality of its picks by having human raters compare that result set with the set of popular images on Flickr; the result was excellent, with its "hidden gems" scoring statistically the same as the most popular images on the site.

I would expect a lot more work on related techniques over the next few years, and for this to have a significant impact on the way that content recommendation is done. The main upshot will be that more little-known works get the spotlight they deserve -- something critical, as more and more people are creating things of value that they want the world to see. 

h/t +Wayne Radinsky and +Daniel Estrada
Beautiful images are not always popular ones, which is where the CrowdBeauty algorithm can help, say computer scientists.
114 comments on original post
1
Add a comment...

Andrea Casalotti

commented on a video on YouTube.
Shared publicly  - 
 
The yield signs on the cycle tracks before the pavements are confusing. Many people will think that they have to yield to motor vehicles as well. It would have been better to give priority to people on bikes here.
1
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
The optimist in me believes that soon the argument will flip over from the defeatist nationalist one prevailing at the moment to a more enlightened view that finding permanent residence for millions of refugees is not only a moral imperative but quite advantageous economically and from the point of view of national defence. The UK for example good divert the billions of pounds spent on futile weapons of mass destruction and invest them hosting hundreds of thousands of refugees in the many dilapidated cities.

This indeed is the way to rescue the European vision. Demilitarize and divert the money spent by the military economic complex to technological and social programmes that will guarantee genuine peace.
Refugees can revitalize the city, which has a large Arab community.
1
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
When the problem is that the NSA and their accomplice agencies don’t care a bit what the law says, the solution cannot be to change what the law says.
 
The NSA have wiretapped in bulk since 1976. They're not going to care what happens to the Patriot Act of 2001. New column on Privacy News.

Right now, there is a debate about a small section of the Patriot Act in the U.S., and which option best removes the authorization from the U.S. NSA to wiretap the world. Both answers in the debate are wrong. No change in law will stop the NSA’s behavior: they have been wiretapping like this since at least 1976, and will not care about changes to a law from 2001. It just happens to be the most convenient justification of the day. If that justification is removed, there will be countless others.

Let me tell you about an event in 2008, when I happened to be on the same expert panel as the local supervisor of NSA activities.

[...]

You will note from this episode that while there is an outward official legal justification, the top brass know full well that what they’re doing is completely illegal on every level. More importantly, you will also note that they don’t care a bit that it’s illegal, for the simple reason they don’t have to care.

It comes down to this: When the problem is that the NSA and their accomplice agencies don’t care a bit what the law says, the solution cannot be to change what the law says.

https://www.privateinternetaccess.com/blog/2015/05/the-nsa-have-wiretapped-in-bulk-since-1976-theyre-not-going-to-care-what-happens-to-the-patriot-act-of-2001/
View original post
1
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
My goal is to develop a physics without concepts such as charge and force, which, in my view, are not properties of particles, rather observed consequences of their behaviour.

Two points from this article:
1. "Empty dots not only can’t be next to each other, they tend to not even come near each other, repelling one another like particles of the same charge."
2. Avalanches occur when density is higher than 2.125 and one keeps adding grains. But the critical level depends on the rule. So a physical constant can tell us something about the rules of behaviour of matter/energy

2
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
If a person has Alzheimer’s disease, it’s usually the result of a build-up of two types of lesions - amyloid plaques, and neurofibrillary tangles. Amyloid plaques sit between the neurons and end up as dense clusters of beta-amyloid molecules, a sticky type of protein that clumps together and forms plaques. Neurofibrillary tangles are found inside the neurons of the brain, and they’re caused by defective tau proteins that clump up into a thick, insoluble mass. This causes tiny filaments called microtubules to get all twisted, which disrupts the transportation of essential materials such as nutrients and organelles along them, just like when you twist up the vacuum cleaner tube. The new technique uses a particular type of ultrasound called a focused therapeutic ultrasound, which non-invasively beams sound waves into the brain tissue. By oscillating super-fast, these sound waves are able to gently open up the blood-brain barrier, which is a layer that protects the brain against bacteria, and stimulate the brain’s microglial cells to activate. Microglila cells are basically waste-removal cells, so they’re able to clear out the toxic beta-amyloid clumps that are responsible for the worst symptoms of Alzheimer’s. The team reports fully restoring the memory function of 75 percent of the mice they tested it on, with zero damage to the surrounding brain tissue. They found that the treated mice displayed improved performance in three memory tasks - a maze, a test to get them to recognise new objects, and one to get them to remember the places they should avoid.
Australian researchers have come up with a non-invasive ultrasound technology that clears the brain of neurotoxic amyloid plaques - structures that are responsible for memory loss and a decline in cognitive function in Alzheimer’s patients. If a...
1
Add a comment...
Have him in circles
279 people
AGA BARBERINI's profile photo
crystal Isabel Fischetti's profile photo
Lisa Saiki's profile photo
Tom Kearney's profile photo
sebastian casalotti's profile photo
Marjorie Powell's profile photo
Tom Harrison's profile photo
Larry Thomas's profile photo
Aris Rakas's profile photo

Andrea Casalotti

Shared publicly  - 
 
Uncertainty and the Precautionary Principle

We have only one planet. This fact radically constrains the kinds of risks that are appropriate to take at a large scale. Even a risk with a very low probability becomes unacceptable when it affects all of us – there is no reversing mistakes of that magnitude.
It is the degree of opacity and uncertainty in a system, as well as asymmetry in effect, rather than specific model predictions, that should drive the precautionary measures. Push a complex system too far and it will not come back. The popular belief that uncertainty undermines the case for taking seriously the ’climate crisis’ that scientists tell us we face is the opposite of the truth. Properly understood, as driving the case for precaution, uncertainty radically underscores that case, and may even constitute it.
1
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
Definitely.
By +Hugh MacLeod 
1
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
“We know there are numerous brain networks that control distinct cognitive functions such as attention, language and control, with each node of a network densely interconnected with other nodes of the same network, but not with other networks,” Marois said. “Consciousness appears to break down the modularity of these networks, as we observed a broad increase in functional connectivity between these networks with awareness.” The research suggests that consciousness is likely a product of this widespread communication, and that we can only report things that we have seen once they are being represented in the brain in this manner. Thus, no one part of the brain is truly the “seat of the soul,” as René Descartes once wrote in a hypothesis about the pineal gland, but rather, consciousness appears to be an emergent property of how information that needs to be acted upon gets propagated throughout the brain.
1
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
A valiant attempt to understand on what the UK Department of the Energy and Climate Change spends our money. The answer is: very little on programmes that have measurable impact on carbon reduction. The biggest share is to clean up nuclear plants.

The estimated costs of cleaning up the Sellafield nuclear site rose an estimated £5 billion to £53 billion in February this year, according to a March statement from Public Accounts Committee chair Margaret Hodge.
Update 14/5 - A few clarifications are worth making. First, the £8 billion DECC budget in 2013/14 includes large accounting adjustments ("provisions") which do not reflect actual expenditure. This is better reflected by the £3.4 billion "departmental expenditure limits" budget shown further down ...
2
Add a comment...

Andrea Casalotti

Shared publicly  - 
 
One of the myths: more efficient irrigation techniques lower water usage. In reality more efficient irrigation means more efficient growing, i.e. more water retention by food and therefore more water usage.


Water from  less efficient irrigation is generally used downstream by other farms (a small percentage is "lost" through evaporation, but presumably that will fall back down again).
Yes, farmers use a whole lot of Golden State water. But there's no easy solution to the state's drought woes in turning off their taps.
1
Add a comment...
Andrea's Collections
People
Have him in circles
279 people
AGA BARBERINI's profile photo
crystal Isabel Fischetti's profile photo
Lisa Saiki's profile photo
Tom Kearney's profile photo
sebastian casalotti's profile photo
Marjorie Powell's profile photo
Tom Harrison's profile photo
Larry Thomas's profile photo
Aris Rakas's profile photo
Work
Occupation
UK Agent, Christiania Bikes
Employment
  • Futures trader, present
  • K4RGO - Cargobike evangelist
    2013 - present
    One of the world's foremost expert in cycle logistics, cargobikes and their use.
  • Velorution
    Director, 1997 - 2012
  • UBS
    1993 - 1995
  • PaineWebber
    1986 - 1988
  • Sumitomo
    1983 - 1986
Places
Map of the places this user has livedMap of the places this user has livedMap of the places this user has lived
Currently
London
Previously
Milano - Copenhagen - Norwich - Fointainebleau
Story
Tagline
Velorutionary
Introduction
Main activities:
I am currently working on these campaigns:

I have three wonderful children.


Bragging rights
Owned the most beautiful bicycle shop in ...
Education
  • Insead
    MBA, 1983 - 1983
  • UCL
    Economics, 1980 - 1982
  • Copenhagen International School
    1974 - 1976
  • Rygaards
    1973 - 1974
  • Liceo Scientifico Enrico Mattei
    1976 - 1979
Basic Information
Gender
Male
Birthday
January 1, 1961