Introduction to AI §
AI in Popular Culture §
- Early Computers & Turing (1940s) and the British Bombe were significant milestones.
- Clockwork & Automata (1700s-1800s) showcased early mechanical creations like Maillardet’s Automaton and The Writer.
- Early AI research involved problem-solving, logic and games.
- The Perceptron model of a neuron was introduced.
Maillardet’s Automaton §
- Created by Swiss clockmaker Henri Maillardet around the late 18th century.
- Designed to imitate a draftsman or artists, it could create drawings and write messages in beautiful calligraphy using pen and ink.
- It had a set of intricate clockwork mechanisms that allowed it to move its hand with precision.
The Writer (1775) §
- Created by Pierre Jaquet-Droz, a Swiss watchmaker and inventor.
- Was designed to resemble a young boy holding a quill pen and paper, it could write custom messages and draw images.
- Could write any text up to 40 characters in length, even create intricate drawings.
Turing Test (1951) §
- It involves a conversation between a human and a machine, with the goal being to determine if the machine’s response are indistinguishable from those of a human.
- Not a well defined-test, posing significant challenges in AI.
Challenges and Realisations §
- Machine translation struggled with context.
- Handwriting recognition worked well only for easy examples.
- Chess was still dominated by humans.
Expert Systems (1970s-Present) §
- Deep Blue won in chess in 1997.
- Research explored brain studies, algorithms and improved models.
- Computers have become more powerful since the 1970s.
Connectionism §
- Connectionism, focusing on distributed processing, gained importance.
- “Lesion Studies” showed the brain’s ability to recover from damage.
- AI demonstrated the capability to adapt, generalise and assimilate new information.
AI Today §
AI Applications §
- AI is used in various fields, including pattern recognition, reasoning, sensing, responding, learning, logic, and game playing.
- Applications span medicine, law, science, art, and many other domains.
Chatbots and Language Understanding §
- Chatbots like Lamda and Chat GPT have been developed.
- They participate in conversations and pass the Turing Test to some extent.
- However, their fluency relies on massive data, computing power, and probabilistic models.
- They may struggle with understanding context and meaning.
AI Challenges §
- AI systems, like Chat GPT, can provide answers but may not comprehend context.
- Accountability and the potential for misuse remain concerns.
- AI is not about machines taking over the world but about addressing challenges related to groupthink, insularity, and arrogance in the AI community.
Glossary §