This FAQ is very incomplete at present. If you have more questions you’d like to see addressed please email one the mailing lists and the answers may wind up here.
Q: What’s the Secret Sauce?
In a phrase: cognitive synergy.
The human brain appears to be an integration of an assemblage of diverse structures and dynamics, built using common components and arranged according to a sensible cognitive architecture. However, its algorithms and structures have been honed by evolution to work closely together – they are very tightly inter-adapted, in the same way that the different organs of the body are adapted to work together. Due their close interoperation they give rise to the overall systemic behaviors that characterize human-like general intelligence.
We believe that the main missing ingredient in AI so far is cognitive synergy: the fitting-together of different intelligent components into an appropriate cognitive architecture, in such a way that the components richly and dynamically support and assist each other, interrelating very closely in a similar manner to the components of the brain or body and thus giving rise to appropriate emergent structures and dynamics. Which leads us to the central hypothesis motivating the proposed research: that the cognitive synergy ensuing from integrating multiple symbolic and subsymbolic learning and memory components in an appropriate cognitive architecture and environment, can yield robust human-like intelligence.
The reason this sort of intimate integration has not yet been explored much is that it’s difficult on multiple levels, requiring the design of an architecture and its component algorithms with a view toward the structures and dynamics that will arise in the system once it is coupled with an appropriate environment. Typically, the AI algorithms and structures corresponding to different cognitive functions have been developed based on divergent theoretical principles, by disparate communities of researchers, and have been tuned for effective performance on different tasks in different environments. Making such diverse components work together in a truly synergetic and cooperative way is a tall order, yet we believe that this — rather than some particular algorithm, structure or architectural principle — is the “secret sauce” needed to create human-level AGI based on technologies available today.
Cognitive synergy is achieved in the OpenCog design via a number of mechanisms including
- A common neural-symbolic knowledge representation (the AtomSpace) that multiple cognitive processes can all work on
- A common semantics spanning multiple cognitive processes, based on probability theory and artificial economics
- Careful design of each cognitive process involved, ensuring that it can identify when it’s “stuck” and appeal to other cognitive processes appropriately
- An overall architecture founded in cognitive science, emulating the brain’s high-level synergetic dynamics
Q: Why Do You Think this Can Work When Other Attempts to Create Powerful AGI Have Failed?
The short answer is…
- Now is the time for AGI because: computers are far better now; our understanding of cognitive science and neuroscience is a lot better now; and our arsenal of computational learning algorithms is a lot better now
- Due to the short-term focus of the current business community, and an anti-AGI attitude on the part of most current government research funding sources, not all that much R&D work on AGI is currently getting done, in spite of the ripeness of the time for it (and in spite of an increasing amount of attention paid to AI in general)
The time is now, the opportunity is here, but due to historical and practical reasons, very few are making a serious effort to grasp hold of the opportunity. The OpenCog project is doing so.
A more technical answer is that
- intelligence depends on the emergence of certain high-level structures and dynamics across a system’s whole knowledge base
- we have not discovered any one algorithm or approach capable of yielding the emergence of these structures
- achieving the emergence of these structures within a system formed by integrating a number of different AI algorithms and structures requires careful attention to the manner in which these algorithms and structures are integrated; and so far the integration has not been done in the correct way
We believe that doing this integration in the right way is difficult but not insurmountably so, and that the OpenCog design contains one viable approach for doing so.
The role of robotics along the path to advanced artificial general intelligence is somewhat controversial. Some theorists consider roughly human-like embodiment essential to the creation of human-level AGI; others consider robotics basically a distraction from the core issues of cognition and language. OpenCog contributors take a variety of perspectives, but the guiding philosophy behind OpenCog from the start, on the part of the project founders, was that
- human-like embodiment is probably not necessary for human-level AGI but is potentially very useful
- virtual-world embodiment and robotic embodiment are both potentially very useful for AGI development; they contribute overlapping but different things, and may fruitfully be pursued in parallel
The benefits obtained by embodiment in general are numerous, including the broad facts that
- humans will have an easier time relating to AGIs that have vaguely humanlike embodiment
- the human brain is oriented toward controlling a body, so using the human brain as inspiration is much easier in the case of embodied AI
- embodiment provides a well-understood context for combining all the aspect of intelligence, including cognition, language, learning, perception, action, creativity, socialization, self-modeling, etc.
The benefits conferred by robotics as opposed to virtual embodiment include:
- one can be more certain one’s AI is achieving tasks that really are similar to human real-world tasks, rather than tasks that merely visually appear similar
- the physical world provides a great deal more sensorimotor richness than any existing virtual world, and this richness may be important for cognition and language as well as perception and action, via providing grist for metaphors and analogies, and concrete exemplars for more abstract cognitive tasks
Thus, although we are not committed to robotics as the “golden path to AGI,” we feel it has a very important role to play. Exploring the nature of this role in detail is one of the objectives of the OpenCogBot project.