Picture Credit: Public Works Knowledge Network
History of knowledge silos
Beginning in the later half of the 17th century, science became the activity of collecting and filing all available knowledge as well as the delineating and systematic arranging of topics. Disciplines became a new organizational way of creating and filing knowledge. Extending scientific analysis to every subject und opening up a potentially unlimited succession of research activities is producing a steadily growing number of knowledge silos with disciplines being subdivided into new disciplines. The science publication platform ‘Frontiers’ for example subdivides research reports about neuroscience into more than twenty sub-categories and this trend is growing. The fact that information and knowledge is documented digitally and distributed globally via internet provides additional support to this trend. It is estimated that over 80% of all digital information accessible today has been created within the last two years. As a consequence the complexity to manage knowledge has risen to the point where humans have difficulty to organize themselves in this vast pool of knowledge and to maintain a top-down view amid the information overflow generated by hundreds of scientific papers and reports published every day. Google, the dominant search engine to access digital knowledge receives about one million inquiries per second. The listing of answers typically contains hundreds of titles that match the inquiry, another indication of the inefficiency of knowledge silos.
Current efforts to unlock knowledge silos
Efforts to overcome the negative consequences of knowledge and information silos are present both in corporations as well as in research and educational institutions:
- In corporations so-called matrix organizations have been introduced to unlock business information silos and increase cooperation and communication to unlock resources and talent that have been inaccessible to the rest of the organization. A matrix organization helps develop individuals with broader perspectives and skills who can deliver value across the business and manage in a more complex and interconnected environment. Operating in matrix organizations can be quite demanding for managers as they have limited control over their subordinates setting their priorities. Matrix managers need to build trust in distributed and diverse teams and to empower people, even though they may rarely get to meet them face-to-face.
- In research and educational institutions so-called ‘Critical-Thinking’ initiatives have been launched. The world renowned Swiss Institute of Technology (ETH Zurich) has implemented a program that advises students to take courses outside their ‘silo’ discipline with the aim to train them in thinking critically and independently. During the course of their education, they should not only acquire methodological skills and disciplinary knowledge, but are also given opportunities to work on complex, interdisciplinary and system-oriented issues.
As all these efforts support the idea of interdisciplinary thinking and collaboration, in the long run it might not be enough to deal with our present-day information overflow as we continue to store our knowledge in hierarchical silos.
The impact of neuroscience research
The huge research efforts to understand the functionality of the human brain with about 100’000 scientists working on this question worldwide also includes the study of the human capacity for learning and memorizing. Only in the last decade neuroscience researchers have been able to go inside the brain and observe how learning actually occurs at the molecular level. New technologies like functional magnetic resonance imaging (fMRI) have opened up the brain’s inner workings and allowed scientists to “see” what is going on inside the brain when people are engaged in learning. Through a network of neurons connected via synapses, sensory information is transmitted along neural pathways called axons and stored temporarily in a short-term memory that acts like a receiving center for the flood of sensory information we encounter in our daily lives. Once processed in short-term memory, our brain’s neural pathways carries this information to our long-term memory, a vast repository of everything we have ever experienced and learned in our lives.
This process is not always perfect. In fact, as information races across billions of neurons’ connections which transmit signals to the next neuron via synapse, some degradation is common. This can result in a blurred acquisition of information and possibly irrationalize decision making. If the new information proves useful to us, it becomes part of our long-term memory. If this new information does not seem useful or if we do not trust its source, we are likely to forget it or even reject it altogether, preferring to stick with the information we already possess. There is an academic debate as to how much information our brain can store. In February 2016 researchers at the Salk Institute for Biological Studies reported that human brain’s memory-storage capacity is an order of magnitude greater than previously thought. Their study has found that the storage capacity of the human brain may be around a staggering quadrillion bytes and this at the brain’s power consumption of just 20 Watts! According to Paul Reber, professor of psychology at Northwestern University the limit to human memory in a lifetime is not hard drive disk space, but download speed. “It’s not that our brains are full,” Reber says. “The information we’re experiencing comes in faster than our memory system can write it all down.” The resulting experience of information overflow supports our behavior to document and store more and more information externally in digital silos.
From Neuroscience to Machine Learning
The human brain is unique for its enormous complexity and processing capacity: its hundred billion neurons and several hundred trillion synaptic connections can process and exchange prodigious amounts of information over a distributed neural network in a matter of milliseconds. Since the artificial intelligence community has begun to mimic the human brain by solving problems such as image recognition or language analysis based on computational networks modeled after our brain’s biological networks a paradigm shift is under way to create machine knowledge which eventually will match or exceed the cognitive capacity of the human brain. The continuous progress in neuroscience supported by huge government grants such as the EEC’s funded Human Brain Project will accelerate artificial intelligence research and the continuous creation of accessible machine knowledge. This knowledge is mapped in networks equivalent to the brain hence hierarchical knowledge silos are no longer needed. We have to learn to open our mind asking the right questions as opposed to cramming silo-knowledge into our overloaded brains. The brain’s capacity has limits and it makes far more sense to employ human brains with knowledge tied to emotions and leave rational and logic based memory up to machines. To quote Albert Einstein:
Education is not the learning of facts but the training of the mind to think
Conclusion
What does this neuroscience research suggest about learning? We need to ensure that learning engages all the senses and taps the emotional side of the brain, through methods like humor, storytelling, group activities, experimenting and games, etc. As cognitive knowledge and intelligence becomes a commodity accessible to everyone, learning should foster intelligence tied to emotional issues. If we experience an emotional reaction to something – fear, anger, laughter or love – that emotion becomes part of the memory and strengthens it dramatically.
Within the corporate context breaking-up knowledge silos opens the way towards collective intelligence. The sharing of human intelligence that emerges from collaboration, collective efforts, and competition of many individuals supported with networked machine knowledge will lead to consensus and improved decision making reducing the risk of corporate failures in times of increasing economic complexity.
The neural capacity of a mouse brain is at best the level of artificial intelligence we have reached so far in applying machine learning technology. We still have a long way to go but exponential growth both in computer hardware performance and neuroscience based software drives this paradigm shift at a staggering rate. It is probably no exaggeration to state that machine intelligence could match human cognitive intelligence by the year 2030.