- Activation Function
- Confusion Matrix
- Convolutional Neural Networks
- Forward Propagation
- Generative Adversarial Network
- Gradient Descent
- Linear Regression
- Logistic Regression
- Machine Learning Algorithms
- Multilayer Perceptron
- Naive Bayes
- Neural Networking and Deep Learning
- RuleFit
- Stack Ensemble
- Word2Vec
- XGBoost
- Attention Mechanism
- BERT
- Binary Classification
- Classify Token ([CLS])
- Conversational Response Generation
- GLUE (General Language Understanding Evaluation)
- GPT (Generative Pre-Trained Transformers)
- Language Modeling
- Layer Normalization
- Mask Token ([MASK])
- Probability Distribution
- Probing Classifiers
- SQuAD (Stanford Question Answering Dataset)
- Self-attention
- Separate token ([SEP])
- Sequence-to-sequence Language Generation
- Sequential Text Spans
- Text Classification
- Text Generation
- Transformer Architecture
- WordPiece
- AUC-ROC
- Analytical Review
- Autoencoders
- Bias-Variance Tradeoff
- Decision Optimization
- Explanatory Variables
- Exponential Smoothing
- Level of Granularity
- Long Short-Term Memory
- Loss Function
- Model Management
- Precision and Recall
- Predictive Learning
- ROC Curve
- Recommendation system
- Stochastic Gradient Descent
- Target Leakage
- Target Variable
- Underwriting
A
C
D
G
L
M
N
P
R
S
T
X
- Activation Function
- Confusion Matrix
- Convolutional Neural Networks
- Forward Propagation
- Generative Adversarial Network
- Gradient Descent
- Linear Regression
- Logistic Regression
- Machine Learning Algorithms
- Multilayer Perceptron
- Naive Bayes
- Neural Networking and Deep Learning
- RuleFit
- Stack Ensemble
- Word2Vec
- XGBoost
- Attention Mechanism
- BERT
- Binary Classification
- Classify Token ([CLS])
- Conversational Response Generation
- GLUE (General Language Understanding Evaluation)
- GPT (Generative Pre-Trained Transformers)
- Language Modeling
- Layer Normalization
- Mask Token ([MASK])
- Probability Distribution
- Probing Classifiers
- SQuAD (Stanford Question Answering Dataset)
- Self-attention
- Separate token ([SEP])
- Sequence-to-sequence Language Generation
- Sequential Text Spans
- Text Classification
- Text Generation
- Transformer Architecture
- WordPiece
- AUC-ROC
- Analytical Review
- Autoencoders
- Bias-Variance Tradeoff
- Decision Optimization
- Explanatory Variables
- Exponential Smoothing
- Level of Granularity
- Long Short-Term Memory
- Loss Function
- Model Management
- Precision and Recall
- Predictive Learning
- ROC Curve
- Recommendation system
- Stochastic Gradient Descent
- Target Leakage
- Target Variable
- Underwriting
What is artificial general intelligence?
Artificial general intelligence (AGI) is a system of generalized, humanlike cognitive capabilities designed to learn and complete a vast range of tasks.Technology of this magnitude has not yet launched due to the complexity of mimicking the biological network of the human brain.
Capabilities of AGI
AGI was designed to perform any task that a human brain is capable of, using flexible thinking and reasoning skills. It was also designed to offer more computational advantages than a human brain, in that it has near-instant recall and split-second number computing skills. While artificial intelligence (AI) offers similar advantages to AGI, capabilities that would set AGI ahead are:
Cause and effect
AGI is designed to understand reactions to specific actions, which enables the model to understand cause and effect.
Sensory perception
There are specific types of sensory perception that AGI can understand, such as color recognition, depth perception, and 3D static images.
Fine motor skills
AGI is projected to complete tasks that would normally require fine motor skills such as grabbing keys from a pocket or solving a Rubik’s cube.
Natural language understanding
Currently, AI can be programmed to present relevant information upon request. This does not show AI comprehension or an understanding of context, which, alternatively, is something that AGI would understand.
Problem solving
AGI will be able to understand when a problem has occured, and what is needed to solve it. While AI offers similar capabilities, it requires training and parameters to be met in order to solve the problem.
Creativity
AGI creativity is defined as the model being able to rewrite its own code. This will result in improvements on understanding and reacting to situations.
Social and emotional engagement
Similar to sensory perception, AGI will be designed to recognize emotions in facial expressions, vocal tones, and body language.
AI vs. AGI
AI excels at completing tasks and problem solving, once specific training has been completed. The training offered can range from supervised to unsupervised, but must be used to train the AI on one specific task.
Theoretically, AGI will be able to perform any human task while exhibiting an extensive intellectual range. Additionally, AGI will be able to complete the same tasks as AI, but with less specific training.
Narrow AI vs AGI
Artificial Narrow Intelligence (ANI), also known as “Weak AI”, refers to AI models or tasks that can outperform a human. While beneficial, these tasks are limited to narrow requirements and must be fully structured through set parameters and rules. ANI also requires supervised and unsupervised training.
As previously stated, AGI would be designed to complete a variety of tasks without the need for strict parameters and complex training.
When will AGI be Invented?
There is no clear path to AGI. AI is capable of performing a single task at a time, but when asked to perform that task outside of the programmed function, it is useless. AGI will be capable of performing multiple functions without the limitations that AI faces. It is unknown when or if this technology will be implemented.
AGI Resources H2O.ai
The Key To Unlocking Artificial General Intelligence: Generative Deep Learning