2:46 pm - Wed, Nov 30, 2011

4 notes

Neuroscience and predictive coding in e-discovery

30 November 2011 — U.S. Federal Court Judge Jed Rakoff is the federal judge who raised questions about why he should approve the government’s $285 million civil settlement with Citigroup, and upended the deal.  He has a record of showing hostility toward Securities and Exchange Commission settlements. For more click here.  To an outsider looking in at the US legal system, it might seem like only one judge – Jed Rakoff – decides all the big cases on Wall Street.  He has taken the role of public arbiter in SEC enforcement cases, repeatedly questioning why individuals are not being held accountable for their misdeeds during the financial crisis and why companies that settle with the agency neither admit nor deny guilt.

But what I find very cool about him is his other pursuit.  He became a senior judge last year to allow him to teach a class at Columbia University about neuroscience and the courts, his real passion.  He serves on the Governing Board of the MacArthur Foundation’s Law & Neuroscience Project.  There is a lot of information on the web about the Project.  For a YouTube clip (with links to other related YouTube clips) click here.  I had the opportunity to meet Judge Rakoff last year at a financial forensics e-discovery conference where he was the keynote speaker. 

It also allowed me to get involved with the Project.  The Law and Neuroscience Project, which is now in its fourth year of operation, brings together lawyers, neuroscientists, and scholars from many other disciplines to explore the legal implications of new findings in cognitive neuroscience.  I became involved when one of my companies was tasked with finding candidates for Research Associate positions for the Project.  They were looking for lawyers familiar with  academic research, not necessarily with science backgrounds but some familiarity with one or more of the fields of biology, psychology, or neuroscience.

The Project focus is on set of problems at the intersection of neuroscience and criminal justice:  the mental states of defendants and witnesses, assessing a defendant’s capacity for self-regulating his behavior, etc.  

Obviously there are many other areas of the law being examined in the light of neuroscience.  One of those areas involves electronic discovery (also called e-discovery or ediscovery) which refers to any process in which electronic data is sought, located, secured, and searched with the intent of using it as evidence in a civil or criminal legal case. E-discovery can be carried out offline on a particular computer or it can be done in a network.  

I have two companies involved in e-discovery and this past year I have spent a fair amount of time on the science-side of the industry at several technology and science conferences.  Especially the cognitive science events which demonstrate how cognitive science is such an interdisciplinary effort that unites biology, neuroscience, AI research, psychology, linguistic and philosophy.  As was explained to me its main aim is “to explain how information processing systems, such as the human brain, can produce veridical representations”.   Very simply put verdicality involves knowledge structure and the degree to which an experience, perception, or interpretation accurately represents reality.

Much has been written in the past year about the damage that the gadget-filled, pharmaceutically-enhanced 21st century is doing to our brains.  Today’s technology is already producing a marked shift in the way we think and behave, particularly among the young. It is not all bad.  Earlier this year at a SoC (system-on-a-chip) conference I attended I learned about an electronic chip in development that could allow a paralysed patient to move a robotic limb just by thinking about it. And I also learned about chips embedded with algorithims for predictive review in e-discovery.

The brain is not the unchanging organ that we might imagine. It not only goes on developing, it is also substantially shaped by what we do to it and by the experience of daily life.  At an Oxford University program earlier this year I spoke with a neuroscientist who said that on a microcellular level, the brain is an infinitely complex network of nerve cells that make up the constituent parts of the brain and actually change in response to certain experiences and stimuli.  Technology has run rampant on us.  Yes, there have been a flood of articles (and books) on what technology is doing to our brains.  He referred me to one he liked in Scientific American (click here) plus a special series on the subject in the New York Times (click here).

Which is why I asked him about algorithims and how they change thinking.  And if he might even be able to discuss e-discovery and algorithims.  He was well aware of IBM’s Watson experiment  and was also familiar with the software architecture behind it (click here).  And he was not … suprisingly … unfamiliar with predictive coding and e-discovery.

He approached it all from the algorithms side and the heuristic side.   Sometimes we change our mind by the sheer force of logic which challenges our current thinking (algorithms).  More often we change our mind because a story is told or a metaphor is shared which “connects the dots” in our thinking (heuristics).  These “ah-ha” moments are the ones which lead us to see the world in a new light (you know, that proverbial light bulb flash).

Neuroscience has mapped these “ah-ha” moments to occur in the right hemisphere of your brain: the realm of story and metaphor.  Light bulbs don’t go off in the left hemisphere, it’s too logical, too linear and too blind to new perspectives.  But he assumes the “sell point” for e-discovery vendors is to change that left hemisphere, to use a formulaic, analytical and logical solution via technology and to use the power of algorithms (“a set of rules for solving a problem in a finite number of steps, as for finding the greatest common divisor”) and thereby change the way a lawyer thinks, modifying his/her brain.  Because the left hemisphere knows what it knows.  It cannot know what it doesn’t know … unless it has a REALLY helpful algorithm.

I see a future for him in e-discovery sales.

And what about “predictive coding”, the big buzz phrase in e-discovery today?  After a lengthy discussion of what it means in e-discovery one can see that the world of cognitive and computational neuroscience has lent its meaning to e-discovery.  He said “predictive coding constitutes the fundamental functional and cognitive principle for the brain. Predictive coding is the idea that the representation of the environment requires that brain actively predicts what its sensory input will be, rather than just passively registering it.
It generalizes simple cybernetic models of prediction and control into a biological framework.  It requires a synthesis across a number of input fields.  And so the brain learns.  And then predicts”.

Too much for this simple blog post, but certainly a catalyst for a longer one.  And he has tentatively agreed to a video interview so watch this space.

This year I have learned more … more than the straight e-discovery events I attend … about “the big picture” in e-discovery, information management, Big Data, etc. from these types of events.  High-tech events such as The 451 Group conferences, the SoC conferences, enterprise information management and cognitive science programs.  And not surprisingly, events that are drawing more and more e-discovery vendors.

My uptake on all this:  we ain’t seen nuthin’ yet.  My view is that the future of e-discovery and data management will involve such a development in information technology and which will lead to a corresponding uptake across the entire legal profession that such things as automated production of documents and intelligent e-discovery systems and analysis … the mainstays of several of my companies … will become staggeringly faster and more efficient.  And less costly than even the lowest-paid lawyers.  Yes, we now live in a world where information is potentially unlimited and information is cheap.  But meaning and actionable data is still the key and humans shall prevail.  Well, at least until Asimo gets really, really good.