Machine Learning Artificial Intelligence Natural Language Processing Semantic Web Ontology Digital Transformation Probabilistic Generative Model Web Technology Deep Learning Online Learning Reinforcement Learning User Interface Knowledge Information Processing Reasoning Technology Navigation of this blog
Autonomous artificial intelligence
Autonomous AI (AI) refers to AI that has the ability to set its own goals and make decisions and take action according to the situation with minimal external direction. Such AI systems have the ability to recognise and understand information from the environment and select and execute appropriate actions.
Characteristics include the following.
- Self-determinism: sets goals, makes plans and executes them without instructions.Updates goals in response to changes in the environment and circumstances.
- Learning capability: learns from experience and self-improves (e.g. reinforcement learning and generative AI). Decision-making based on long-term data accumulation.
- Adaptability: flexibly adapts to unknown environments and changing circumstances. Adaptable to versatile tasks.
- Safety and ethics: autonomous behaviour, requiring ethical decisions and safety assurance. Control mechanisms are required to act in harmony with human society.
Possible fields of application include
- Automated vehicles: recognise road conditions and traffic rules and drive safely and efficiently.
- Robotics: autonomous decision-making and action in warehouse picking operations and disaster relief.
- Financial trading: trade decisions based on real-time market data.
- Space exploration: self-directed exploration activities in inaccessible environments.
- Medical assistance: assistance in diagnosis and treatment planning.
Related technologies include
- Reinforcement Learning: autonomous behavioural learning.
- Deep Learning: learning patterns from data and understanding the environment.
- Multi-agent systems: co-ordination of multiple autonomous AIs.
- Generative AI: creative output and prediction.
Self-expanding machines
Self-enhancing machines refer to artificial intelligence and robotic systems with the ability to improve and evolve their own capabilities and structures. These systems self-improve through learning and interaction with their environment, becoming more efficient and advanced over time.
Characteristics include the following.
- Learning-based evolution: improved performance based on experience, utilising machine learning and deep learning. Enhanced ability to respond to unknown challenges.
- Physical self-improvement: ability to redesign or extend structures (e.g. robots create and use new tools themselves).
- Software updates: increased efficiency through automatic code generation and model updates.
- Goal adaptability: redefining its own goals and adapting to changing circumstances.
Areas of application include
- Robotics: self-improving designs, improving work efficiency and adaptability.
- Medical: automated diagnostic systems improve their own diagnostic accuracy.
- Manufacturing: factory robots learn new work procedures to increase productivity.
- Space exploration: spacecraft adapt to unknown environments and operate for longer periods of time.
Related technologies include.
- Evolutionary Algorithms: optimisation and refinement using simulation.
- Generative AI: automatic generation of new designs and programmes.
- Self-healing technologies: systems detect and repair faults.
- Adaptive Reinforcement Learning: dynamically optimises the learning process.
Technical Topics
Autonomous artificial intelligence technology can be defined as technology that has the ability to enable artificial intelligence to learn and solve problems on its own. In order to achieve this, functions such as self-learning, self-judgment, self-repair and self-replication are considered necessary.
In order to further analyse the issues described in ‘Problem-solving methods, thinking methods and experimental design’, it is necessary to find hypotheses. To find hypotheses, it is necessary to accumulate experience and analytical skills, and in this article, we will consider AI technologies that can support these.
Paul Nurse, the author of this book, saw a butterfly fluttering into his garden one early spring day and felt that, although very different from himself, the butterfly was unmistakably alive, just like himself, able to move, feel, and react, and moving toward its “purpose. What does it mean to be alive? WHAT IS LIFE” is a tribute to the physicist Erwin Schrodinger’s “What is Life?
To “awareness” means to observe or perceive something carefully, and when a person notices a situation or thing, it means that he or she is aware of some information or phenomenon and has a feeling or understanding about it. Becoming aware is an important process of gaining new information and understanding by paying attention to changes and events in the external world. In this article, I will discuss this awareness and the application of artificial intelligence technology to it.
Behavioural economics, described in “The economy is driven by ’emotions’”, is one of the current trends in modern psychology, which focuses on irrational thoughts and behaviours and tries to reveal the common laws of irrationality, in contrast to conventional economics based on rational human activities. Two systems are envisaged in our minds, one being a quick, automatic, unintentional and unconscious system, such as intuitive judgement, and the other being a deliberate and conscious system, such as logical judgement, which is time-consuming but controllable. AI technology, which is also described in this blog, aims to take the rational decision-making of System 2 to the extreme instead of humans, and IA, which is also described in “Overview of Intelligence Augmentation (IA) and its application examples”, is an approach to how to connect the useful parts of System 1 and System 2. The IA described in “Overview of IA (Intelligence Augmentation) and its application examples” can be said to be an approach to how to connect the useful parts of System 1 with System 2.
Priming is also an interesting concept in the field of AI, and there is active research into using the concept of priming to improve human-AI interactions. For example, in AI-based experience (UX) design, when a user performs a specific task, an AI assistant can more accurately understand the user’s intentions and present relevant information and context in advance, so that subsequent operations can proceed smoothly, such as Priming can be considered.
The frame problem in agent systems refers to the difficulty for agents to properly understand the state and changes in the environment and to make decisions when acquiring new information. This is specifically the case in the following cases.
When the soft deterministic idea of free will is considered in terms of the use of artificial intelligence technology, it is possible to derive options that machines can ‘do otherwise as well’ beyond the possible human options, and among these, not simply algorithms that can also be realised by machines, but ‘causal reasoning and considerations towards the realisation of strong AI’. If problems can be solved with algorithms based on deep imagination and models based on that imagination, as described in ‘Considerations for causal reasoning and strong AI’, then humans could play a role that machines cannot play.
- Information Integration Theory and its applications
Information Integration Theory (IIT) is a theory proposed by psychologist Norman H. Anderson that is used to understand the process by which people integrate multiple pieces of information to make decisions and judgements It is a model. This model plays a particularly important role in cognitive and social psychology and represents how people’s judgements and evaluations are formed.
- How does the brain see the world?
The question ‘How does the brain see the world?’ has long been explored in fields such as neuroscience, psychology and philosophy, and provides insight into how the brain works to produce the world we perceive, interpret and are aware of. This section examines whether these perspectives are feasible in AI.
コメント