IBM's Watson plays Jeopardy “Unconsciously”

The video above shows Watson, a computer system with artificial intelligence developed by IBM, participating in a televised game show against Jeopardy record holders Brad Rutter and Ken Jennings. The computer uses algorithms (which have been around for a while) to parse out phrases from the jeopardy questions and match them against information in its internal memory. Watson has open access to a memory bank of 200 million pages of text including dictionaries, thesauri, newswire articles, literary works and all of the encyclopedic entries of Wikipedia. This information takes up four terabytes of disk space in its RAM based memory. It is not the first computer to solve these kinds of questions but it is the first to do so in the order of a couple of seconds with very high accuracy. 

Watson parses the clues into different keywords and sentence fragments in order to find statistically related phrases. Through the simultaneous execution of thousands of language analysis algorithms, Watson comes up with a number of potential answers (mostly proper nouns or verbs), that are statistically most likely to match the cues given by Jeopardy's Alex Trebek. Watson chooses the answer that has been found most frequently by different, independently operating, algorithmic searches.  

Watson is not conscious. It has proven that it can best the most skilled human trivia practitioners in the world but it doesn't "understand" the questions that it answers. Watson might be able to narrow its associative search down to a single city but it cannot appreciate why it chose the city, cannot visualize what it would be like to visit this city and has no desire to do so. Although it can come up with a precise list of the most closely semantically related words and concepts, it cannot reflect on these in a creative, insightful or emotional way. In some ways though, Watson's method of processing is very much like our own. Its multiassociative approach is much like the way our unconscious mind can select a single answer when given a set of related cues.

Like Watson we have mental representations of words, perceptions and concepts organized in our head and these are associated to one another in various ways. The associations are determined by our experiences and each concept, is associated with many other concepts. To access the memory of a concept, several other closely related concepts must be activated at the same time. In fact, when a set of concepts are coactivated (like the words twinkle, distant and night sky), sometimes there is only one representation that all of these descriptions sum together to activate. Interestingly, even our brain cells, at a very fundamental level are organized for multiassociativity. Individual neurons must have messages being sent to them by many other neurons, for them to increase their firing rate. In the cognitive, neuroscientific and AI sense, representations made active by a jeopardy question send out outputs to many different dormant representations and the ones that are converged upon the most become active. In other words, multiple cues activate tons of different representations, but only the representations that best match the suite of cues - taken all together - are activated maximally to become conscious.  The important question is, are the associative links in your network of associations wired up in such a way that the representation corresponding to the correct answer will become active?

An answer on Jeopardy a few days ago was: The name of a planet, an element and a Roman god. Within three seconds, the question, "what is mercury," popped into my head. It didn't pop in within a few hundred milliseconds so I must have done some conscious thinking beforehand. But I also know that I did not go sequentially down the list of elements and planets in my head in search for a commonality. The node in my brain (which might correspond to a cluster of neurons or even of cortical minicolumns) that corresponds to the concept of mercury, was converged upon and made active by these other concepts in an automatic and unconscious way. 

Snap judgments are made, unconsciously, in this way. But humans can do more than just make blind, automatic multiassociative guesses. Given time and the motivation to deliberate, humans can contemplate questions from a variety of different perspectives. It seems to me that most Jeopardy questions offer a number of associative cues that will automatically cue up the answer in a contestant’s head if they know it. Some Jeopardy questions seem to offer cues that require additional processing and the use of inferential and deductive thinking. Interestingly, Watson seems to be able to make up for its lack of human logic, on questions like these, with processing power and extensive memory.

To me the important question is, how does one imbue a system like Watson’s with human-like consciousness beyond instantaneous multiassociation? I think that the answer lies in endowing a computing system like Watson with the equivalent of a prefrontal cortex to keep certain representations active through the span of several seconds. Of course, this form of artificial intelligence would have to have a prolonged series of developmental experiences, similar to a childhood, to learn which representations to keep active in which scenarios. By maintaining the activiation of a representation, something that happend moments ago can impact future activity. This may not help much in Jeopardy where each question is meant to stand on its own independent from the other questions. However, in my mind, extended activation is totally responsible for the creativity, insight and emotionality that is necessary for most human activities. 

Mental Disorders at the Office

I have suspected for a long time that almost every one of the characters in NBC’s TV series “The Office” has a psychological disorder - even if it is a slightly subclinical one. Each character has a flaw that causes them to consistently make the same types of misconstruals, misapprehensions or misjudgements.  Often these mistakes lead the characters to commit serious social or occupational blunders making my favourite show on television a madhouse confined within the four walls of second story, Scranton, Pennsylvania office space.  In my Mom's opinion, the writers either intentionally modelled their characters after common psychiatric or psychological disorders or else at least a few of them have an uncanny sense of how and where everyday people can go wrong. What follows is a list of the characters and the psychological disorders that I think they are supposed to represent.

Few of the characters deserve a full blown diagnosis, rather most have tendencies or habits that are characteristic of a specific disorder. Further, some of the characters' patterns of behavior don’t map onto a true disorder but to a personality trait. A person usually only receives a diagnosis if their disorder consistently and adversely affects their well-being, their social life, their economic status or the people around them. On these criteria, you be the judge. Characters are listed in an order that reflects the severity of their disorder, with the first being the most severe. Most of the descriptors below are symptom criteria coming straight out of the DSM (Diagnostic and Statistical Manual of Mental Disorders).

Michael Scott:
ADD: Inattentive, impulsive, inattention to detail, careless mistakes, does not follow through, difficulty organizing tasks and activities, forgetful, easily distracted by extraneous stimuli

Dwight Schrute:
Aspergers Syndrome: Impaired social and emotional interaction, impaired empathy and sympathy, forced to compensate for deficit of empathy by memorizing social patterns, idiosyncratic communication, make-believe inappropriate to developmental level, intense asocial preoccupations, obsessed with details that have no social consequences, other symptoms of autism

Creed Bratton:
Schizophrenia or Schizotypy: Disturbance of consciousness, poor clarity of awareness of the environment, reduced ability to sustain or shift attention, disorientation, memory impairment, frequent derailment of thought, incoherent, odd, eccentric or peculiar,

Angela Martin:
Obsessive Compulsive Disorder: recurrent and persistent thoughts about maintaining structure and order despite anxiety and stress that they may cause, preoccupied with lists, details and organization, feels driven to behaviors that are subject to rigid rules, overconscientious, scrupulous and inflexible about matters of morality, ethics or values. Hyperfrontality: habits pertaining to hygiene or social conventions are clearly excessive

Ryan Howard:
Narcissism: grandiose sense of self and inflated ego, arrogant and egoistic, is interpersonally exploitative and opportunistic, takes advantage of others, exaggerates achievements and talents, preoccupied with fantasies of success and power, has a sense of entitlement, lacks empathy

Erin Hannon:
Avoidant: avoids interpersonal activities that involve social contact in an effort to avoid criticism, disapproval or rejection, unwilling to become involved unless certain of being liked, shows restraint within intimate relationships for fear of shame, ridicule, inadequacy or embarrassment

Kevin Malone:
Mental Retardation: Does not meet the standards of his age group for pragmatics, communication, or self direction.

Kelly Kapoor:
Dependent Personality Disorder: pervasive and excessive fear of separation leading to clinging behavior, necessitates advice and reassurance from others, false veil of openness, extraversion and sociability.

Mr. Mifflin:
Alzheimer’s Disease: Significant memory impairment and advanced cognitive deficit, forgetfulness, disorientation

Toby Flenderson:
Depression: depressed mood throughout the day, feels sad or empty, low self esteem, feelings of hopelessness or melancholy, diminished interest or pleasure in most activities

Meredith Palmer:
Alcohol Dependence, Substance Abuse, Addictive Personality, Hyperactive Sexual Desire

Jan Levinson:
Neurotic: Impelled by distressed, anxious or angry energy

Stanley Hudson:
Antisocial / Psychoticism: failure to be apprehensive of aggressive acts, irritability, lack of remorse, indifference

Andy Bernard:
Intermittent Explosive Disorder: several episodes of failure to resist aggressive impulses that result in harm or destruction of property. Degree of aggressiveness is out of proportion to the precipitating psychosocial stressor, age inappropriate displays of anger

Jo Bennett:
Histrionic: uncomfortable in situations in which she is not the center of attention, uses physical appearance to draw attention, inflated ego, has a style of speech that is excessively impressionistic

Phyllis Lapin-Vance:
Eating Disorder

Oscar Martinez:
Conscientious

Pam Halpert:
Low Self-esteem, but she is much better now

Jim Halpert:
Procrastination, Sarcasm, Competitiveness

Darryl Philbin:
Surgency

Holly Flax:
I guess she is just a bit nerdy

Could Hiccups Represent An Evolutionary Reflexive Response to Choking?

I have long wondered about the evolutionary or adaptive value of hiccups. As a child I used to hiccup a lot and they seemed strange and inexplicable. Some recent experiences helped me to see that hiccups may play a role in swallowing. More precisely they may serve to help mammals recover from choking by clearing the esophagus. On several occasions I have had a mouthful of food block my esophageal airway. When this happens with a particularly large mouthful, it can partially occlude the tracheal airway making it hard to breathe. Several times now I have been saved by a hiccup. The hiccup literally squeezed the food down my esophagus and into my stomach.

A hiccup is like a cough in reverse. During a cough the diaphragm expels air from the lungs as the glottis opens. During a hiccup the diaphragm pulls air into the lungs while the glottis is closed and I firmly believe that this can pull food down the esophagus. Most people experience several hiccups in a row usually not coinciding with a meal or with choking. Perhaps these normal hiccups are the body practicing this important reflex. I would be very interested to know how many other people have experienced a hiccup while choking and had their airway cleared because of it.

It happened to me again today. I was eating way too much food, way too fast and I had an uncomfortable feeling in my throat. It is a strong sensation that usually induces a bit of panic. This often happens when I eat something dry without drinking anything. Have you ever eaten a hard-boiled egg first thing in the morning? When this happens I always hiccup, but only a couple of times. The hiccuping tends to push the food down my esophagus, relieving me of the choking feeling. Could most cases of the hiccups represent misfirings of this important, hair-trigger reflex meant to keep our airway clear?

Important reflexes come out sometimes even when they are not needed. At times we salivate despite the fact that we are not about to eat. We often stress out, releasing adrenaline and cortisol despite the fact that we are not about to engage in strenuous physical activity. For a long time now I have concluded that random bouts of hiccuping are triggered accidentally but represent a  means to constrict the middle respiratory tract forcing large boluses of food down so that they do not obstruct the trachea. I can think of many ways to test this hypothesis experimentally, anyone want to help?




Here are some of my favorite books that explain how common human traits are actually adaptive biological responses:

Psychopathy is a Selfish Behavioral Strategy that may be Programmed by Ecological Adversity

Psychopathy, sometimes known as antisocial personality disorder, is characterized by a lack of empathy and severely amoral conduct. Because people with psychopathy have so much trouble empathizing they are much less afraid to cross social boundaries, insult or hurt others. Because about 1% of the population are psychopaths evolutionary psychologists have come to believe that psychopathy may represent a viable ecological strategy that helps an individual to selfishly gain resources by taking advantage of others. Our hunting and gathering ancestors have relied on cooperation for millions of years and this is why most people are not psychopathic. Perhaps though, in small numbers, genes that influenced individuals to be less helpful and more selfish benefitted those that bore them explaining why psychopathy is heritable and why genes for psychopathy exist in the human genome. It is too bad that these genes can’t tell if the environment is a healthy, friendly one that is conducive to team playing or if it is a bad one which would reward those that defect from the team. Maybe the genes can tell...

A good deal of very interesting research points to the fact that environmental cues can switch on genes that would normally stay dormant, resulting in changes to the body and brain. For instance, water bugs whose mothers are stressed out by the scent of their natural predators make babies that have totally different protective armor. A huge number of similar “predictive adaptive responses” to environmental adversity have been documented in animals and even humans – many arising from prenatal stress. Studies of the Dutch Hunger Winter of 1944-1945 have shown that individuals born just after this famine are more likely to have lower metabolisms. This is thought to represent a predictive response to low food availability. Interestingly, these studies also show that these famine-born individuals are more likely to be psychopathic.

It is difficult to say exactly what this means, but it suggests to me that humans may be programmed by early environmental stress (and perhaps postnatal adversity too) to be selfish, manipulative, guiltless and antisocial. It would make sense for natural selection to program us to be more deceitful and cunning in a hostile environment that is marked by scarcity and thus high levels of competition. Such a change would force the person or animal to look out for number 1 – to look out for its genes. Peruse the following clinical rating scale from the PCL-R and ask yourself: “what kind of environment would these traits go along well with?”

Psychopathic Traits:

Factor 1: Aggressive narcissism

1. Glibness/superficial charm
2. Grandiose sense of self-worth
3. Pathological lying
4. Cunning/manipulative
5. Lack of remorse or guilt
6. Emotionally shallow
7. Callous/lack of empathy
8. Failure to accept responsibility for own actions

Factor 2: Socially deviant lifestyle

1. Need for stimulation/proneness to boredom
2. Parasitic lifestyle
3. Poor behavioral control
4. Promiscuous sexual behavior
5. Lack of realistic, long-term goals
6. Impulsiveness
7. Irresponsibility
8. Juvenile delinquency
9. Early behavioral problems
10. Revocation of conditional release

Here are some of my favorite books that explain how common human traits are actually adaptive biological responses:

Mapping the Spatial Relationships Between Cortical Inputs

The cerebral cortex is a big sheet of neurons about 2 millimeters thick with a surface area of about four 8.5 x 11 pages of paper. Under this thin sheet of gray matter (neuron bodies) lies all of the white matter that creates the connections between neurons. The physical characteristics of neurons differ a little throughout the sheet but the main differences - that determine what the neurons process – are found in the inputs. Inputs from the sense organs and other brain areas feed into the cortex in a distributed way. I believe, that if you want to know, in general, what a small patch of cortex does, the best way to do this is to look at the inputs that feed into it. Particular areas of cortex are programmed by not only the nearest input, but other nearby inputs. These inputs probably work together, triangulate or overlap like circles in a Venn diagram, to determine the relevant functional properties of nearby cortical tissue.

 Input from the retina feeds right into cortex in the back of the brain and this is why occipital cortical areas there process visual stimuli. Up near the top of the head, in the somatosensory strip, are inputs from sensors on the skin. This strip of cortex is thus responsible for processing touch information from all over the body. But what about half-way in between these two areas? What is processed in the space between visual and tactile? The parietal cortex sits in between these two areas and its main function seems to be almost a split between visual and somatosensory processing. The parietal cortex is responsible for our mental representations of personal space, the areas right around our body. Could it be that the function of parietal cortex is determined by a combination of its nearest inputs? I think so. As another example, consider the temporal lobe which receives inputs directly from the cochlea of the ear. As you might guess much of the temporal lobe is involved in identifying the source of heard sounds. Guess what the area intermediate between visual and auditory cortex is involved in? It is involved in the recognition of conceptual objects, normally things that both have visual forms and make audible sounds.

 In more complex areas you have several inputs, very close to each other, coming from sensory organs, other cortical areas and also neuromodulatory systems that release transmitters like dopamine, serotonin and adrenaline. This gets very complicated, very quickly, but I have a strong hunch that the unique connectional trigonometry of patches of cortex determines what kinds of things their neurons are exposed to and how they "perceive" the world. Many diagrams that I have seen in neuroscience display the inputs and outputs associated with different areas but contain no information about the spatial orientation of these inputs relative to each other. For many reasons, I think that the spatial configuration of cortical inputs is very important and will be instrumental in determining function in different cortical areas - especially those with multiple, multimodal, overlapping inputs such as the prefrontal cortex.