Monday, 11 January 2010

A long thank you note!

Almost an year & a half back, I met an interesting family at the Bangalore airport. They were a middle aged couple with 2 kids. A boy named Akash, who was in Class 7 and a girl named Nidhi who was in Class 6. I think the dad was somewhere in his early 40s. I remember them today, because a simple little conversation I had with them has turned out into a really powerful idea in the education space.

I was talking to them & telling them about how I run a company called Vita Beans. They were naturally surprised that a guy who looks like a high school kid had started a company & they wanted to know more. And since I anyway love talking about my startup, I started telling them about how we started as a research project almost 2 years before & how we built our first simulation engine that simulates human behaviour in a variety of environments. We had used a myriad of research frameworks from neuroscience, cognitive science, artificial intelligence & fuzzy logic. They didn't seem to understand much except that we were tremendously successful in getting high accuracies & quality results from our work.

Somewhere down the conversation, the mother said something very interesting. She said "There is something I always wanted to know. You know when I take Nidhi with me to a grocery store or a shopping market, the shopkeeper tells me the price of all the items I have purchased before billing me.... Item1 - 30Rs, Item2 - 217Rs, Item3 - 125Rs & so on. And within seconds of him completing the list, my daughter looks at me & says - Mom, it's 1231 rupees. Now... I know she's right because she has developed a habit of doing that every time and she is always right. But the shopkeeper does not know that and so he takes out a calculator, punches in the numbers... Voila! 1231 rupees it is. He then gives me a smile & a complicated look that says - Wow, you have a genius for a daughter there!. And I'm all smiles & proud of her" she was already beaming at her daughter when saying this.

She continued "But then we come home & an hour later, she shows me her marks card & it reads - mathematics... 42/100. I don't know why but I suddenly loose my cool & start yelling at her - You're so dumb at maths. You need to work hard, I'm sending you to tuition starting tomorrow. I know she is good at maths, but I don't know why she scores less. I don't know what to do"

She then looks at me & says "You said you do all those stuff about human behaviour simulation. Do you understand why she scores less in exams when she really is intelligent?"

Frankly, I did not know! When we were in school, my younger brother too was something similar. He could remember the batting averages of every damn cricketer from 1983 to the present day. But his super memory didn't work as well when he had to remember his lessons. And I had no clue why! Neither did my parents.

I just shrugged, looked at the mother & said "I don't know why! But I'm very sure it has nothing to do with mathematics as such" But that got me thinking. When I came back, we discussed it in my team, we spoke to some of the professors working in the field of learning & cognitive sciences. We took time, but we tried to understand why Nidhi was scoring less at maths.

After a while, we found that we were not alone. A few other research groups have also been excited about the connection between learning & cognitive functions. We looked more into such results & came up with frameworks of our own to translate many of the research findings that we learnt into something that kids can understand & use. It seemed like we were finally able to answer - why is it that kids fare differently in situations which demand the use of same skills, but in different ways.

But before we converted our frameworks into some sort of a product, there was another challenge that caught our attention. This has been a problem which grown ups (read parents, teachers, adults) would most of the time not even acknowledge as something that is a problem. And that is - we, as kids, are inherently lazy. We are lazy to get up in the morning, we are lazy to iron our clothes & if there is an exam on that day, we are lazy to even have breakfast. We wait for our Mothers to spoon feed our breakfast as we feverously glance through our textbooks. In spite of knowing about our laziness, our entire system of education & learning tries to pull us out of this laziness before it can teach us something that can change our lives. A tough job, ain't it!

On the other hand, we are also super-active & enthusiastic about certain other things. Things which we enjoy doing, things which make us happy. Grown ups often cannot even understand our excitement in many cases. So we looked at these two contrasting tendencies in children & thought - why not build a tool that puts the kids in a fun & addictive gaming environment & weave the learnings into the environment so that they automatically pick it up as a side effect?

And so we started working on the design for a tool, called MindMaps. It is a cognitive tool that lets you profile & train your brain functions in an addictive & enjoyable way. The brain functions can be the ones like Memory, Attention, Problem Solving, Learning & Flexibility or they can even be skills like Speech, Maths & Vocabulary.

We have been working non-stop to convert our designs into a tool that focuses as much on user experience (if not more) as on the utility and usefulness of the product. The learning has been great so far & the first opinions from folks in the industry has also been exceptionally good. And somewhere in between the excitement - I feel like thanking the couple & their kids, Akash & Nidhi, who got us into this space an year and a half ago.

I don't know what the names of Akash & Nidhi's parents are! I never asked, they never told. But here's me saying Thank You! with all my heart.

Friday, 13 March 2009

Revisiting - Blink !


Much of our understanding of mind-reading comes from two remarkable scientists, a teacher and his pupil: Silvan Tomkins and Paul Ekman. Tomkins was the teacher.

He was born in Philadelphia, at the turn of the last century, the son of a dentist from Russia. He was short, and thick around the middle, with a wild mane of white hair and huge black plastic-rimmed glasses. He taught psychology at Princeton and Rutgers, and was the author of "Affect, Imagery, Consciousness," a four-volume work so dense that its readers were evenly divided between those who understood it and thought it was brilliant and those who did not understand it and thought it was brilliant. He was a legendary talker.

At the end of a cocktail party, a crowd of people would sit, rapt, at Tomkins's feet, and someone would say, "One more question!" and they would all sit there for another hour and a half, as Tomkins held forth on, say, comic books, a television sitcom, the biology of emotion, his problem with Kant, and his enthusiasm for the latest fad diets, all enfolded into one extended riff.

During the Depression, in the midst of his doctoral studies at Harvard, he worked as a handicapper for a horse-racing syndicate, and was so successful that he lived lavishly on Manhattan's Upper East Side. At the track, where he sat in the stands for hours, staring at the horses through binoculars, he was known as the Professor. "He had a system for predicting how a horse would do based on what horse was on either side of him, based on their emotional relationship," Ekman remembers. If a male horse, for instance, had lost to a mare in his first or second year, he would be ruined if he went to the gate with a mare next to him in the lineup. (Or something like that-- no one really knew for certain.) Tomkins believed that faces--even the faces of horses--held valuable clues to our inner emotions and motivations.

He could walk into a post office, it was said, go over to the "Wanted" posters, and, just by looking at mug shots, tell you what crimes the various fugitives had committed. "He would watch the show 'To Tell the Truth,' and without fault he could always pick the person who was lying and who his confederates were," his son, Mark, recalls. He actually wrote the producer at one point to say it was too easy, and the man invited him to come to New York, go backstage, and show his stuff." Virginia Demos, who teaches psychology at Harvard, recalls having long conversations with Tomkins. "We would sit and talk on the phone, and he would turn the sound down while Jesse Jackson was talking to Michael Dukakis, at the Democratic National Convention. And he would read the faces and give his predictions on what would happen. It was profound."

Paul Ekman first encountered Tomkins in the early 1960's. Ekman was then a young psychologist, just out of graduate school, and he was interested in studying faces. Was there a common set of rules, he wondered, that governed the facial expressions that human beings made? Silvan Tomkins said that there were. But most psychologists said that there weren't. The conventional wisdom of the time held that expressions were culturally determined--that we simply used our faces according to a set of learned social conventions.

Ekman didn't know who to believe. So he traveled to Japan, Brazil, and Argentina--and to remote tribes in the jungles of the Far East--carrying photographs of men and women making a variety of distinctive faces. To his amazement, everywhere he went people agreed on what those expressions meant. Tomkins was right.

Not long afterwards, Tomkins came to visit Ekman at his laboratory in San Francisco. Ekman had just tracked down a hundred thousand feet of film that had been shot by the virologist Carleton Gajdusek in the remote jungles of Papua New Guinea. Some of the footage was of a tribe called the South Fore, who were a peaceful and friendly people. The rest was of the Kukukuku, who were hostile and murderous and who had a homosexual ritual where pre-adolescent boys were required to serve as courtesans for the male elders of the tribe. For six months, Ekman and his collaborator, Wallace Friesen, had been sorting through the footage, cutting extraneous scenes, focusing just on close-ups of the faces of the tribesmen, in order to compare the facial expressions of the two groups. Ekman set up the camera. Tomkins sat in the back. He had been told nothing about the tribes involved; all identifying context had been edited out. Tomkins looked on intently, peering through his glasses.

At the end, he went up to the screen and pointed to the faces of the South Fore. "These are a sweet, gentle people, very indulgent, very peaceful," he said. Then he pointed to the faces of the Kukukuku. "This other group is violent, and there is lots of evidence to suggest homosexuality." Even today, a third of a century later, Ekman cannot get over what Tomkins did. "My God! I vividly remember saying, "Silvan, how on earth are you doing that?" Ekman recalls. "And he went up to the screen and, while we played the film backward, in slow motion, he pointed out the particular bulges and wrinkles in the face that he was using to make his judgment. That's when I realized, 'I've got to unpack the face.' It was a gold mine of information that everyone had ignored. This guy could see it, and if he could see it, maybe everyone else could, too."

Demystifying -

Ancient brain circuits light up so we can judge people on first impressions

'Its almost instantaneous and you can't stop doing it': neuroscientists match scans to human decision making

Ancient neural circuits - the amygdala and posterior cingulate cortex - are active when people form first impressions of new acquaintances.

Photograph: Daniela Schiller/New York University/Nature Neuroscience

Scientists have recorded the gentle flicker of activity that lights up the brain when we form our first impressions of people. The study shows how age-old brain circuitry that evolved to make snap decisions on the importance of objects in the environment is now used in social situations.

Brain scans taken while volunteers formed opinions of new acquaintances found activity surged in an ancient neural circuit that helps us make a rapid assessment of a person's character.

"Humans have always been engaged in making decisions on what's important and what's not, and social decision making is taking advantage of these primary systems in the brain," said Daniela Schiller, who led the study at New York University.

"Whenever you need to assign value to something, you use the same mechanism, whether it's an inanimate object or a person. It's like there's one common currency in the brain."

Previous work by neuroscientists has shown we form our first impressions well within 30 seconds of meeting people. Often, our opinion changes very little after knowing them for longer.

"When you meet a person, they might say something, or look a certain way, or behave a certain way, but you have very little information on which to form an opinion, but it is almost instantaneous and you can't withhold from doing it," said Schiller.

In the study, published in Nature Neuroscience, Schiller and scientists from Harvard University took brain scans of 19 volunteers who were asked to form a first impression of a series of fictional characters.

The volunteers were shown faces of men on a computer screen, followed by six sentences that described a mix of good and bad aspects of their character. For example, the person might have picked up his room mate's post on the way home from university, or told a fellow student they were stupid. After reading all of the sentences, the participants were asked to rate how much they liked the person on a scale from one to eight, with eight being the most likable.

Schiller's team then looked through the images from the scanner to see what brain regions had been most active while people formed their first impressions.

The scans showed that two brain regions were involved in opinion-forming, the almond-shaped amygdala, which is linked to regulating emotions, and the posterior cingulate cortex, which is active in making financial decisions and putting values on the outcomes of situations.

"Even when we only briefly encounter others, brain regions that are important in forming evaluations are engaged, resulting in a quick first impression," said Elizabeth Phelps, a co-author on the study at New York University.

Understanding the biological circuits involved with opinion forming might help scientists learn what happens when they are disrupted or fail to activate properly. "It might affect the impressions you have of others, and that could feed into the basis of your relationship with them from then on," said Schiller.

Though, Gladwell captured it well without the mumbo-jumbo in his book, Blink.

Rapid cognition is the sort of snap decision-making performed without thinking about how one is thinking, faster and often more correctly than the logical part of the brain can manage.

Gladwell's discussion of 'thin slicing' is arresting: The secret is knowing which information to discard and which to keep. Our brains are able to perform that work unconsciously; when rapid cognition breaks down, the brain has seized upon a more obvious but less correct predictor.

There are things that can be done to redirect our mind along lines more conducive to accurate thin slicing: we can alter our unconscious biases; we can change products' packaging to something that tests better with consumers; we can analyze numerical evidence and make decision trees; we can analyze all possible facial expressions and their shared meanings, then watch for them on videotape; and we can evade our biases by blind screening, hiding the evidence that will lead us to incorrect conclusions.

Friday, 6 March 2009

They Read Minds


... or thought reading (as they say)

For the basics about multivariate fMRI "mind-reading" techniques, see the video below. Some of it is based on this 2007 Haynes et al paper from Current Biology, described in more detail following the video.

What Haynes et al have done is to ask 8 subjects to freely decide either to add or subtract two numbers, and to select among 4 options an answer corresponding to the task they chose. After repeating this process many times, the authors ran a pattern classifier on the metabolic activity recorded in the brain.

This pattern classifier was run on the unsmoothed fMRI data - smoothing is normally applied because fMRI is thought to be a relatively noisy recording technique. Critically, the use of a pattern classifier allows the use of unsmoothed data (and in fact requires it) because buried within the noise is a distributed signal reflecting the distributed neural patterns encoding the subject's intention. Such data is presumably lost in averaging/smoothing operations.

Haynes et al trained their pattern classifier (a linear support vector machine) using a "multivariate searchlight" (described here) approach. This means that for every recorded voxel, they fed the classifier information about both that voxel and those surrounding it. The classifier was trained on 87.5% of the data (using 8-fold validation), and maps were produced of the classifier's accuracy at each voxel in the brain. These "accuracy maps" were averaged across subjects to produce the following figure:

Crop from Figure 2 from Haynes et al

As you can see above, the results showed that intention is decodable both prior to and during the intended response in numerous regions in the prefrontal cortex. In particular, the anterior & posterior medial prefrontal cortices as well as lateral frontopolar cortex, right middle frontal gyrus, and left operculum contained information that allowed the decoding of intentions at a level significantly above chance. Intentions prior to responses were also decoded based on activity in the temporo-parietal junction, although it is not illustrated in the above figure (see the supporting online material here). Much debate focuses on the precise roles of these regions, but their involvement here would be predicted by the majority of cognitive neuroscientists.

Conspicuously absent from these maps is the intraparietal sulcus (IPS), which has been argued to reflect numerical processing. An interesting possibility is that the numerical processing accomplished by this region cannot be distinguished based on the numerical operation (addition vs. subtraction), which would support a process-independent representation of quantity. Note that this conflicts with some theories of numerical processing in the IPS.

What's fairly amazing about this work is that they used a pretty standard scanner (only 3 tesla) with a reasonable sampling time (a TR of just over 2.7s). Peter Bandettini has suggested that this unsmoothed multivariate approach would benefit from higher resolution MRI, but Haynes et al have demonstrated surprising success with much more widely-available technology.

Related Posts:

However, all we ask you to do is create your Vita profile sitting on your bean bag in your apartment by just clicking your mouse/pressing space bar. And, I can assure you that although it wouldn't involve any mumbo-jumbo and wouldn't put a hole in your pocket; it would for sure help you understand and explore things about yourself that you'd never expect a machine to be able to figure out! !

Thursday, 5 March 2009

Split Brain - Literally


A marvel to observe and know which side of your brain does what !
To reduce the severity of his seizures, Joe had the bridge between his left and right cerebral hemisphers (the corpus callosum) severed. As a result, his left and right brains no longer communicate through that pathway.
Here's what happens as a result:

Psychology quote of the day

"Common sense and a sense of humor are the same thing, moving at different speeds. A sense of humor is just common sense, dancing."

- William James, American psychologist and philosopher (1842 - 1910)

Letting my common sense do its bit ... new study shows this ! ( why do people even bother and take all that pain, I fail to grasp ?)

Friday, 27 February 2009

Special ain't so special?

In 1920, a few villagers caught 2 girls in remote countryside west of Calcutta. They had been spotted previously with adult wolves and were found in a wolf den with two wolf cubs. The den was dug up, the mother wolf killed and the girls taken away. J A L Singh, an Anglican missionary, who ran an orphanage, took them in and gave them their names - Kamala & Amala.

Kamala was thought to be five or six years of age and Amala around two years old. They were dishevelled, ate raw meat in the manner of dogs, and howled but could not talk. The were indifferent to temperature - a characteristic of people leading rugged lives - had sharp hearing, good vision in the dark and a strange gleaming look in their eyes.

Kamala and Amala stood and walked on all fours. Kamala was so adept as a quadruped that she could outstrip on four legs anyone on two legs and climb and jump easily. But like many other children of her feral background she never seriously mastered walking upright and resorted to hands and knees when needing to. Amala died the year after she was found. Kamala survived into her teens and managed to learn only some three dozen words.

It is interesting to note that Kamala & many other feral children successfully managed to blend very well into the lifestyles of animals they were raised by. If Kamala continued to live with the wolves, she would never realize that she had the ability to speak & master a language, or the ability to walk upright. To Kamala, these ordinary human abilities might have appeared as superpowers!

It makes me wonder if there are many such superpowers hidden inside all of us. Take the example of Kim Peek (The Rain Man). Kim remembers everything, literally everything. Ofcourse it comes with a price - social & developmental disabilities (especially motor skills). While the origin of Kim's abilities have been traced to congenital abnormalities in the brain, it becomes exciting once we remember that the human brain is extremely malleable in our early years. In other words, it is highly possible that a child can grow these abnormalities during it's early development purely because of the way it is raised. Kamala & other feral children have more than proved this to be true.

But why then is it so rare to find guys like Kim with extraordinary abilities? My bet is that the answer lies in the human tendency to outcast everything that is 'not like them'. Kim himself is still considered a disabled person - someone who needs to cured. No wonder that most people who display special abilities today are either raised by non-humans or picked up these abilities even before they were born amidst us. The rules & guidelines based on which our society is built, only promotes those behaviors & abilities that make it easy for new kids to blend with the adults - be like everyone else.

Maybe it is possible to develop abilities like that of Kim & many more like him, without having to compromise social skills or any other essential abilities of the human mind. If only we were more open to explore lifestyles that appear alien, in comparison with the majority that form our social order.

Sunday, 22 February 2009

Unconnected Dots

Leslie Brothers attached electrodes to the brain of a monkey which was watching videotapes of the face of another monkey. She found neurons selectively responsive to the other monkey’s facial expression of emotions. An identical behavior is found in human infants where the selectively repond to expressions on adult faces.

A child invariably stares longer at an object that you drop out of your hand, but does not fall down. Somehow, it knows that all objects are supposed to fall to the ground - gravity. We all are prejudiced even before birth, that the light (sunlight) always comes from the top, which is why the same shaded-circle appears bulged when seen at an angle & depressed when seen upside-down.

All these support Darwin's view that emotional & cognitive behaviors are remnants of actions that were functional in evolutionary history. Since the feeling of self-awareness & consciousness is invariably linked with emotions, this suggests that the notion of Self itself might be an evolutionary functionality.

On the other hand, William James held that emotions are internal perceptions of physiological processes in our own bodies — tense facial muscles, sweaty palms, and especially the effects of the autonomic nervous system, such as a pounding heart, faster breathing, and higher blood pressure. Recent works on Somatic Theory by Antonio Damascio also strongly uphold this view. This seems to suggest a bodily (somatic) source for Self.

In a very different plane, Donald Griffin has studied the mental abilities of insects and animals. He associates consciousness with complex and novel behavior in changing or unfamiliar circumstances. For instance, Bees can communicate the direction and distance of food sources and can distinguish between water, nectar, and a possible hive site; they do their waggle dance only when other bees are around, but they have limited ability to modify their behavior in new circumstances. African Grey Parrots can talk excellently but they fail to comprehend the meaning of Self. They often refer to themseves in third person saying - "Polly needs water" instead of saying "I need water". Some apes can recognize parts of their body in a mirror & even communicate to some extent symbolically & even with the use of sounds. But their communicative abilities are greatly dwarfed in front of that found in humans to reveal anything more than evolutionary impressions of Self.

But works of Griffin & others bring out a curious observation - the notion of Self began to emerge only after organisms started to indulge in a social life. Greater the complexity of social interactions & social needs, the more expressed the recognition of self & our feelings towards it.

Though there are many such alternate views to view the source of emotions, self & consciousness, they don't seem to be in contradiction with each other. They just seem to be talking about the same source, seen from different angles. Will all the views converge? Can there be a unified theory for defining everything that is human?

Friday, 20 February 2009

I was - I am

If you want to truly understand something, try to change it.

- Kurt Lewin (American Psychologist)
I'm often tempted to believe - I means everything that is unique to my existence. Just to see if such a definition has any relevance in the chain of evolution, let's take a blue-green algae with a boring lifestyle. Broadly speaking, it does 3 things over the course of it's life -
  • Interact with the environment.
  • Reproduce.
  • Die. (Woah! Isn't it remarkably similar to our lives)
So all that can be unique to an algae's existence is - the way it interacts with the environment. Within the lifetime of an algae, this way is completely determined by its genes. Crude as it may seem, millions of years of evolution has added just 1 more category of activity in the lives of humans -
  • Store our interactions & learn from them.
While an algae can learn only from it's genes, humans can learn in innumerable ways. But this small change is sufficient to completely alter the meaning of I. While the way an algae interacts with it's environment evolves over thousands of generations, the way we interact with our environment evolves with every interaction. So apart from the way we interact, the interactions themselves become a part of us, part of I. No prizes for guessing what takes care of this - the human brain.

Looking through the glass, this essentially means -
  • I = genes (in algae) = genes + brain (in humans)
  • Understanding an algae = understanding it's genes + it's environment ; Understanding humans = understanding our genes + brain + environment
  • While I evolves over generations in an algae, it evolves continually in humans
While most of the things I've mentioned above might appear common sensical, they give a direction to delve deeper. On a different note, returning to common sense is also essential to discard some of the long held prejudices & start afresh.

Amidst all this, am I suggesting that - it is I or the notion of Self which has been driving evolution all along?