The video of cognitive robot XCR-1, featured five characteristics of the sentient beings, and demonstrated a machine too can learn simple associations. The robot XCR-1 is a three-wheeled robot, without an envelope that gives a shape and form, with gripping hands, with multiple sensory capabilities and with a capacity to produce words. The robot has limited motor capacity (gripping, moving forward/ backward, talking and touching), but has excellent sensory systems for visual, auditory, touch, pain and tap stimulus. Obviously it has been created for studies and experiments in the field of machine learning. The experiment pioneered by Dr. Haikonen illustrates five behavioral phenomena in machine i.e. searching for objects, feeling pain, associating pain with an object, verbal learning and sound direction detection. It seems that the robot does not operate on microprocessors or logical programs of any kind. The behaviors are not based on the programming logic as in a computer application but on the model for cognitive computation, namely an associative neural processing characteristic that essentially and integrally combines sub-symbolic and symbolic process.
In the experiments with the XCR-1, the stepping stones to higher order learning by a machine are laid and the cognitive and affective processes as in humans and other sentient beings are attributed to the machine. For reproducing the processes of perception, imagery, speech, pain, pleasure, emotions, etc. Haikonen (2003) proposes a special cognitive architecture. The robot works on the basis of HCA (Haikonen Cognitive Architecture) and associative neural processing. HCA is a set of organized associative neurons and neuron groups (electronic). The robot can respond to objects and can record its experience. For instance, the robot could recognize a green object and withdrew from it because it had a painful pat when it touched the object. The robot seems to be “aware” of the object and the consequence of touching it. The question is, does the robot have consciousness or merely carrying out set of operations? The experiments demonstrate the machine can respond to sounds and acquire associative learning as in the Pavlov’s classical conditioning.
In the XCR-1 experiment, the robot was put through the experience of associative learning or respondent conditioning. The robot formed association between the unconditioned stimulus (the painful pat on the rear side) and the conditioned stimulus (the green object). Before the pat, the green object was a neutral stimulus, after the painful pat the green object became the conditioned stimulus. The pat gave the experience of pain and the unconditioned response to pat was withdrawal. Later green object became conditioned response and elicited withdrawal behavior (conditioned response). In the second experiment, the word blue was neutral stimulus. The experimenter called the word ‘blue’ and then gave a pat to the robot and then placed a blue object in front of the robot. The robot avoided the blue object by moving backward, away from the blue object. Pat was the unconditioned stimulus and the avoidance was the unconditioned response. The neutral word blue became the conditioned stimulus, and avoidance was the conditioned response to it.
Other two experiments illustrate perception with associative memory and response based on perception. In the experiment, the robot declares what it perceives and acts based on what it says. It can be inferred that the robot can associate meaning with the objects existing in its learned environment. Through the neural networking architecture, the use of representational symbols and symbol systems; “words” and “language” can be taught to a machine.
However, a common question bordering psychology and neural networks are ‘how can immaterial phenomenon arise from the material system such as brain or electronic circuits?’ The theories of consciousness attempt to compare the processes of the mind (immaterial) to brain and to patterns of electrical neural activity (material). No plausible answer has emerged to the question why the mind is immaterial, and has a subjective sense. The experiment with the robotics is successful in creating experiential learning to the robot, but the questions on consciousness of the machine and the humans are yet not answered fully.
A very ingenious suggestion by Haikonen, 2011, takes the problem of consciousness to a different direction. He suggests we are not able to explain consciousness because we are approaching the problem from a wrong direction. If we ask, why we perceive mind as immaterial, the answer to this question will make vanish the problems associated with the understanding of consciousness and make then redundant.
Franklin, S, B J Baars, U Ramamurthy, and Matthew Ventura.. The role of consciousness in memory. 2005, Brains, Minds and Media 1: 1–38,
Haikonen, Pentti O. A. "XCR-1: An Experimental Cognitive Robot Based on an Associative Neural Architecture.." Cognitive Computation 3, no. 2, 2011: 360-366.
Haikonen, Pentti, The Cognitive Approach to Conscious Machines, Exeter, UK: 2003, Imprint Academic, I
Searle, J. R, "Minds, brains, and programs", 1980, Behavioral and brain sciences, 3(3), 417-457
Schlagel, R. H, Why not artificial consciousness or thought?", 1999, Minds and Machines, 9(1), 3-2