London Blog Day 91: Riot In London

I took usual overground service from Sydenham to West Croydon. The train was very quiet this morning, considering rush hour. Not many Metro papers scattered on seats, few silent passengers, gloomy weather behind windows. When I got out in Croydon the first thing that stuck me was sweet burning smell. I got out of the station and that’s when the scale of damage really hit me. Ruskin Street sealed of by police, debris scattered everywhere, police helicopter, people gathering around, speaking in rush on mobile phones. And the burning I can smell, even when I am inside the office building now.

The sense of anxiety is real and it gets to you - believe me.

It is obvious that riots in London have striking resemblance to those in France (2005 and 2007) and Greece (2008). In all those cases, police accidentally killed teenager(s). In all those cases the riot started in a large population city - Paris, Athens, London. In all previous cases it was a long (around 3 weeks), persistent and extremely violent event, with hundreds of people arrested, hundreds of vehicles burned, billions lost by businesses. In all cases there was a ‘domino effect’ - riots spread to other major cities in the country.

London is in day three now. It looks bad.

Many things work against quick fix of the situation.

The first and major - population and city size. You hear those numbers in media - “100, 150, 200 thugs demolished supermarket”. If it happens in a few points of the city, it is pretty hard problem for police to tackle. Large groups also lead to ‘lemming crowd effect’ - sense of social ‘rightness’, power and complete lack of personal responsibility.

Second - Twitter effect. Information spreads like virus, infecting others to join the madness.  Social networks are very effective in organising large groups of people. It was great for Middle East political riots, or Iran protests, but it is a big challenge and problem for London crisis.

Third - police powers. Due to strong political correctness and high respect for individual’s well-being, police is using very ‘soft’ methods to deal with rioters. Police doesn’t use some very effective options like water cannons, tear gas or baton rounds, because those were criticised as being too drastic. It ends up with police in a very difficult situation between people who judge them for either not enough or too much force used with rioters. Mind you, police powers and force are very important psychological factors. If police is not sufficiently threatening to hostile rioters, the rioters will gradually grow in power and loose any respect for authority...

Four - media reality show. It gets to you, especially when you live in London. People around you are talking constantly about it. You see it on every channel, every newspaper. Then, when you watch the actual place where it happened (like Croydon) the anxiety suddenly gets to you. Media play a massive role in public awareness and attitude towards situation. Public anxiety feeds the confidence of hostile rioters. Media also make police very self-aware, holding them from using more force, which is sometimes good, but not necessary in case of London riots, as mentioned.

Let's hope it won't get worse...

From Croydon,

Geek On Acid

London Blog Day 78: Surveillance of the Future (and why You are creating it)

In Minority Report (the movie based on Philip K. Dick novel) there is no crime. Thanks to a group of people with precognitive abilities, government is able to detect the crime before it actually occurs, and contain the person who is about to commit the crime. Things obviously go wrong, Tom Cruise has to run away from the entire police force chasing him, but that’s not really the point.

The point is - such system is the holy grail of surveillance. But instead of precognitive humans, we hope to create precognitive artificial intelligence (AI), that will be able to detect crime before it occurs. 

Such precognition would have nothing to do with 'Minority Report's’ precogs' collective unconscious. Surveillance AI would analyse patterns of body movement, voice tonality, facial expression, group dynamics, object detection and other measurable factors to decide whether something bad is about to happen. It would be linked with national identity database, screening biometric information for any personal records that would estimate the chance of particular person committing crime. Collecting all those information, it would display risk probability and give a final decision to human operator. Or just send the police van to the site without human decision. 

As you can imagine, it’s extremely complex to create such system. The reason is mainly due to complexity of any social scene out there in the world - there are too many viewing angles, lighting conditions, resolution differences, distances, and ways people interact and behave. Another problem is how to integrate all this complex information. You would need proper supercomputer and very good software, operating on a very sophisticated level of non-linear statistical integration, chaos mathematics, similar to what is used to predict weather, but on a larger scale. 

Still, we are closer to this goal than 10 years ago, and UK is one of the places where intelligent AI is most likely to happen. UK has biggest CCTV network in the world. The rough estimate is 4 million CCTV cameras, but it’s likely to be more than that, and counting. While there are some older systems in place, majority of the mainstream network (e.g. London or Glasgow) has been upgraded to high-resolution, high-speed digital cameras capable of running face recognition software. London on it’s own has more CCTV cameras than the entire west coast of the US. Your live is pretty much a silent reality show in London. Our computers are also getting better very fast, and we are able to store and process increasingly more data. Most important - we are becoming increasingly more available for screening via social networks - we open ourselves online and give away all necessary information, without any government enforcing us to do that. We are connected 24 hours to a very precise tracking system - the mobile phone. Thats where the future of surveillance really exist. In you 1000 Facebook photos, your Twitts and text messages, your blogs and ‘like’ comments. Now, the only task is to really connect all those points into one integrated system, that will extract the patterns of information automatically and intelligently, and we have a Minority Report. You wrote it.

London Blog Day 71: Obsessive-compulsive rules.

Bathroom 1: Desist...

Bathroom 2: Shit poem...

Bathroom 3: Swine flu reminder.

Kitchen 1: Good grammatical construction.

Kitchen 2: Just a tiny notice...

Kitchen 3: First microwave notice from graphic designer

Kitchen 4: Second microwave notice, just in case...

And it's all from one place ;-)

[photos by Geek On Acid] 

London Blog Day 53: 'Lemming Crowd Syndrome'

LeBon (1985) saw the crowd as a social structure where individual transform into the primitive, irrational state of a mass of people, characterised by loss of self control, anonymity of the individual, primitive behaviour, suggestibility and a collective mind. A dominant notion here, by analogy with a disease, was that of mental contagion or the spread of emotion. LeBon, like a number of crowd theorists who followed, was concerned with social unrest and the apparent mob-like state of agitated crowds.

I like to think about crowd more in context of ‘Lemming crowd syndrome’ ('Lemming syndrome' as a term is reserved for a cultural context, so I redefine it here slightly). It’s not that you become absolutely mindless and irrational like LeBon suggested, it’s more that you options become limited – specifically your behavioural options. For example, in a busy London street you can try to resist the urge to cross the pedestrian walk on red light when the crowd around you suddenly starts moving, but it’s likely that you will follow them anyway. The rapid movement of crowd makes the option of waiting on red much less attractive. This doesn’t work in Poland of course, where you get fined for crossing the street on red…

‘Lemming Crowd syndrome’ doesn’t only apply to simple directional aspects. You might feel uncomfortable if you act against crowd mental behaviour too. Recently I joined extremely busy train from Bath Spa to London Paddington on Sunday afternoon. The train was obviously overbooked, a mass of people crowded in the passages, lots of backpacks and baggage’s (it was Bath Festival weekend). I didn’t book a ticket in advance. You know how this advance booking works – sits which are taken have a small display that tells you from which station they are taken. There are also free sits for those who bought the ticket on the train station last minute. I had a big backpack, it was very hot, and I had a nasty cold. Normally, I would conform but this time something cracked. I started walking/crawling towards the end of the train. People were not happy, but nobody said anything. I am small enough to pass even through most ridiculous bag barrier, so I slowly progressed through 3 coaches. Not bad. But so far all sits were taken everywhere. I reached the final coach. Two guys, rugby-size, blocking the entrance. I am trying to pass, one is asking ‘do you have a sit booked there?’, I said ‘yes’ with surprised certainty, but feeling extremely uncomfortable lying and they let me pass. There was one sit in this couch booked from Bath, but not taken. I dropped my bag on the overhead, and joined the table, smiling with a certain guilt displayed on my face. Everyone at the table looked at me with suspicion, or if not – I definitely felt like they were. I was stressed for another 10 minutes until we stopped in some smaller station (Chippenham), nobody joined there, and that was it – I got a sit all the way to Paddington (almost 2 hours trip mind you).

What was interesting about this is how this massive crowd set the behavioural rules for the train, blocking passages, and even enquiring me, whether I follow some non-existent crowd ethics – that is exactly what I mean by Lemming syndrome. It was a combination of attribution bias, cognitive dissonance and blind obedience. The limited space, limited options and large crowd worked as perfect decision blocker. And how uncomfortable I felt going against it was just a final prove.

I am sure you got you own stories of ‘Lemming Crowd syndrome’. Share some if you do.

London Blog Day 14th: Your common-sense view of lying and deception is probably wrong.

What makes us judge people as suspicious? There are thousands of papers published by folks from vision and computing science, who try to find a ‘holy grail of security’ - automatic system for detecting hostility. But there is surprisingly small number of papers looking at psychological aspects of detecting suspicious behaviours. Sure, there is a legion of self-help books, NLP-breed mutant-books on earning £10K per week in 5 simple steps and ‘reading people like books’ or common-sense guidelines. These are often ultra-simplistic approaches, missing the point entirely. Science frequently shows us, that counter-intuitive approach turns out to be true. For example, lets take a topic of lying and detecting deception - one of the very important questions I am looking at in my current project. 

You probably assume that people who are experts in interrogation (e.g. military intelligence) would be better in spotting lies than a random person without any expertise in such field. Right? Well, you are wrong. Turns out that novices are frequently more accurate than experts. Why might this be? First, expertise often entails other factors such as age, high motivation, and access to other sources of information that accompany expert’s occupation. Also, experts usually gain little self-correcting feedback when they commit false-negatives (i.e. believing a dishonest person is honest) and also when they commit false-positives (i.e. believing an honest person is dishonest). So they may become overly confident of their detection abilities. Together, these factors can impair judgement much more than in case of novice, who looks at the situation without any preconceptions. If you run a simple experiment comparing experts with non-expert, you would most likely confirm those assumptions (as others did - see Burgoon et al., 1994 - a classic on interpersonal deception).

Another example: you would assume that when you are suspicious about others, you are better in detecting dishonesty or lie (because presumably you are more sensitive to behavioural cues). Wrong again. First - it will be very obvious to the skilled observers that you are suspicious about them, so they can react and adjust their behaviour, and trick you better. There is actually a lot of ‘non-verbal leaking’ coming from the person, who is suspicious about others, and this is not good when you interact with a smart deceiver (who has time to change his communication strategy). Secondarily - if you are suspicious about others, you are also lie-biased, you perceive people as less trustworthily. So you commit much more false-positive judgements than observer who is not suspicious about others. Again, this cause you to focus overly on unimportant information, overload your cognitive system, and possibly omit the real suspect (e.g. in the airports screening situation, where there is massive crowd to monitor). So overall - observers who are suspicious about others are less accurate in detecting deception and truth telling than non-suspicious observers. This ESPECIALLY applies to EXPERTS who seem to make a lot of stereotyped and biased judgements based on their biased experience, which blends badly with their suspicious-receivers approach (see the first point I made).

This doesn’t mean that we need to get rid of all the airport security every year to keep the security fresh. Because there are advantages of being an experts as well - with experience you get better grab of questioning strategies, that can help you a lot in reading other people intentions. You know, those simple and apparently meaningless questions you get asked on airports randomly? Those are just dummy questions, nonverbal triggers, to facilitate any hidden, bad intention that you might ‘nonverbally leak’. With experience, you also get better in reading nonverbal signals, although this one is more unclear, and seem to be related with many other factors.

There is another level - multisensory integration of verbal and non-verbal cues when you are trying to decode the suspect, but this is a completely different story...  

London Blog Day 4. Paper of the week - drugs, violence and holidaymakers visiting mediterranean destinations.

I spend most of the week browsing through security literature. Because it’s a holiday time, the current paper of the week is ‘Substance Use, Violence, and Unintentional Injury in Young Holidaymakers Visiting Mediterranean Destinations’ from 2001 issue of Journal of Travel Medicine titled. 

Authors used survey (for stats geeks: cross-sectional comparative survey) on airports looking at 6,502 British and German holidaymakers aged 16 to 35 years coming back from Cyprus, Greece, Italy, Portugal and Spain. Result showed that over 95% used alcohol, 66% have reported been drunk, over 10% used illicit drugs, 5.8% suffered unintentional injury and 3.8% have been in physical fight. From those who used drugs, 86.5% smoked cannabis, 31.9% took ecstasy, 18.3% cocaine, 5.8% ketamine, 5.7% amphetamine and 3.8% GBH. British tourists visiting Crete took more drugs and were involved in more violent episodes than tourists visiting other destinations. Also, German tourists visiting Portugal took more drugs than tourists visiting other destination, but not different from British on Crete. Odds for violence were overall highest for Majorca. No other significant differences between Germans and British tourists were reported. Violence was most common for those staying between 8 and 14 days. Also, big surprise, drinking alcohol on holidays was associated both with violence and unintentional injury, while taking drugs - more associated with unintentional injury, as relationship with violence wasn’t clear. Those results don’t seem to apply for tourists smoking cannabis. Discuss.

The Big Smoke

The Big Smoke:

7.754 million city area population,

12.58 million metro area population,

412,000 students,

8000 buses,

700 bus routes,

(over) 300 languages spoken,

270 underground stations, 

250 miles of tracks,

58.7% are christians,

43 universities,

40 theatres,

31% of population is foreign-born,

(generates) 30% of UK's Gross Domestic Product (GDP),

10.6 mph is an average speed of car in rush hour,

8 airports,

3 summer olympics hosted historically (as for 2012).

I couldn't be more excited than being here for 3 months to explore it's vast potential of this multicultural, historical, postmodern and ultra-technological city. And on the other side, to accelerate my understanding of the patterns of suspicious behaviors in ports and airports.

London Blog starts today!