Artificial Intelligence Definition
Science and engineering of making intelligent computer programs capable of performing tasks that require subtleties of judgement, interpretation and generalisation that we associate with human intelligence
Narrow AI
AI systems built to perform a single task without any skills that can be transferred to other tasks
AGI and definition
Artificial General Intelligence. AI systems that can autonomously solve a variety of complex problems in a variety of different domains and learn and adapt autonomously
Generative AI
Artificial Intelligence technology that automatically generates new/original content in response to prompts typed in by the user. Should be used with caution as not always correct(phenomena called hallucination)
Generative AI outputs
Text
Images
Videos
Music
Software code
Computer vision percentages meaning
Level of confidence returned by model that the image has been correctly classified as the associate label.
Applications of computer vision
- Self driving cars to detect dangers and reduce accidents
- Track endangered species
- Early detection of diseases through medical imaging
- Facial recognition enhances security by tracking individuals
Database definition
Collection of structured data organised in a way such that the data can be easily and efficiently retrieved and maintained
Record definition
(Aka tuple)
Single row of data containing related elements
Primary key
a column in a relational database table that’s distinctive for each record
Foreign Key
a column in one table that references the primary key of another table
Machine learning societal benefits
- Early detection of diseases
- Recommendations for entertainment (movies, tv, music)
- Facial recognition in security systems
- Online shopping recommendations
bias
a systematic error or prejudice within a model that leads to skewed or unfair predictions. If the data on which the algorithm makes its decision contains bias, the algorithm’s output will also be biased.
Raw data
data that hasn’t been processed, coded, formatted, or yet analyzed for useful information.
Methods of cleaning data
- Removing/ replacing “NA” values
- Checking/ removing/ replacing outlier values
- Changing strings to integers/ floating point numbers
