The Brain and Artificial Intelligence

In parallel developments, research in artificial intelligence and robotics has produced significant results in planning, problem-solving, rule-based reasoning, image analysis, and speech understanding. All of the fields below are active, and there exists an enormous and rapidly growing literature in each of these areas:

• Research in learning automata, neural nets, fuzzy systems, and brain modeling is providing insights into adaptation and learning and knowledge of the similarities and differences between neuronal and electronic computing processes.

• Game theory and operations research have developed methods for decision-making in the face of uncertainty.

• Genetic algorithms and evolutionary programming have developed methods for getting computers to generate successful behavior without being explicitly programmed to do so.

• Autonomous vehicle research has produced advances in realtime sensory processing, world modeling, navigation, path planning, and obstacle avoidance.

• Intelligent vehicles and weapons systems are beginning to perform complex military tasks with precision and reliability.

• Research in industrial automation and process control has produced hierarchical control systems, distributed databases, and models for representing processes and products.

• Computer-integrated manufacturing research has achieved major advances in the representation of knowledge about object geometry, process planning, network communications, and intelligent control for a wide variety of manufacturing operations.

• Modern control theory has developed precise understanding of stability, adaptability, and controllability under various conditions of uncertainty and noise.

• Research in sonar, radar, and optical signal processing has developed methods for fusing sensory input from multiple sources, and assessing the believability of noisy data.

In the field of software engineering, progress is also rapid, after many years of disappointing results. Much has been learned about how to write code for software agents and build complex systems that process signals, understand images, model the world, reason and plan, and control complex behavior. Despite many false starts and overly optimistic predictions, artificial intelligence, intelligent control, intelligent manufacturing systems, and smart weapons systems have begun to deliver solid accomplishments:

• We are learning how to build systems that learn from experience, as well as from teachers and programmers.

• We understand how to use computers to measure attributes of objects and events in space and time.

• We know how to extract information, recognize patterns, detect events, represent knowledge, and classify and evaluate objects, events, and situations.

• We know how to build internal representations of objects, events, and situations, and how to produce computer-generated maps, images, movies, and virtual reality environments.

• We have algorithms that can evaluate cost and benefit, make plans, and control machines.

• We have engineering methods for extracting signals from noise.

• We have solid mathematical procedures for making decisions amid uncertainty.

• We are developing new manufacturing techniques to make sensors tiny, reliable, and cheap.

• Special-purpose integrated circuits can now be designed to implement neural networks or perform parallel operations such as are required for low-level image processing.

• We know how to build human-machine interfaces that enable close coupling between humans and machines.

• We are developing vehicles that can drive without human operators on roads and off.

• We are discovering how to build controllers that generate autonomous tactical behaviors under battlefield conditions.

As the fields of brain research and intelligent systems engineering converge, the probability grows that we may be able to construct what Edelman (1999) calls a "conscious artifact." Such a development would provide answers to many long-standing scientific questions regarding the relationship between the mind and the body. At the very least, building artificial models of the mind would provide new insights into mental illness, depression, pain, and the physical bases of perception, cognition, and behavior. It would open up new lines of research into questions that hitherto have not been amenable to scientific investigation:

• We may be able to understand and describe intentions, beliefs, desires, feelings, and motives in terms of computational processes with the same degree of precision that we now can apply to the exchange of energy and mass in radioactive decay or to the sequencing of amino acid pairs in DNA.

• We may discover whether humans are unique among the animals in their ability to have feelings, and start to answer the questions,

- To what extent do humans alone have the ability to experience pain, pleasure, love, hate, jealousy, pride, and greed?

- Is it possible for artificial minds to appreciate beauty and harmony or comprehend abstract concepts such as truth, justice, meaning, and fairness?

- Can silicon-based intelligence exhibit kindness or show empathy?

- Can machines pay attention, be surprised, or have a sense of humor?

- Can machines feel reverence, worship God, be agnostic?

0 0

Post a comment