At its most basic level, computation involves processing inputs to generate outputs according to specific rules or operations. Your brain does it all the time: it takes in sensory information (inputs), processes it through neural pathways, and produces thoughts, decisions, or actions (outputs). This perspective could lead us to think that if we understand the brain, we understand intelligence, and that’s the end of the story. But this misses an important point: the fact that all species on earth have evolved as embodied agents that physically interact with the world.
But our brain is not touching the world, you might say? In 1960, the anthropologist Sherwood Washburn wrote in Scientific American that “the modern human brain came after the hominid hand.” This is a bold statement, but it does make a point: understanding the brain alone will not lead to an understanding of intelligence. What other computations contribute to the generation of our intelligence?
The principle of multiple computational paradigms states that the robustness and versatility of intelligent behavior results—at least in part—from the synergistic application of multiple computational paradigms.
For example, humans perform neural computation in their nervous systems, but also mechanical computation with their bodies. Below we will present evidence from dexterous manipulation, but we will use this example for illustration here, too. When grasping an object, humans perform neural computations to identify the object, and this triggers movements that bring the hand close to it. However, the specific locations of each finger on the to-be-grasped objects are not explicitly computed. Humans employ the mechanical computation performed by a compliant, soft-material hand: the fingers land as they may. Neural computation has the task of bringing the hand into position such that the mechanical computation of the hand can take over the determination of the final grasp pose. Each computational paradigm solves the sub-problem at which it is good. The result is highly robust and adaptable behavior.
A more in-depth look
We take a computational view on intelligence (Brock, 2024). Computation can be realized in different paradigms, e.g., neural, digital, mechanical, or quantum. Each of these paradigms has advantages over others, depending on the type of problem to be solved. By combining multiple computational paradigms, sub-problems can be addressed using the most suitable paradigm. The principle of multiple computational paradigms states that the robustness and versatility of intelligent behavior results—at least in part—from the synergistic application of different paradigms.
This principle is a specific instantiation of active interconnections that bridge two or more computational paradigms. This interconnection must be adaptive because for each problem there is an optimal balance between the different paradigms. The ability to balance the paradigms leads to generality, robustness, and intelligence.
Evidence
The following video demonstrates an object being rotated by a soft robotic hand. This rotation is achieved using an extremely simple control command. The specific forces needed to execute this motion aren’t explicitly programmed but instead emerge from the mechanical computation happening within the robot hand as it interacts with the object.
The mechanical computation of the soft hand enables generalization. In the following video, a variety of objects are robustly manipulated with the exact same control signal.
