John McCarthy is often called the Father of Artificial Intelligence.
In 1950 Alan Turing gave the idea of the Turing Test, asking whether machines can think like humans.
1955–1956: John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon started the Dartmouth AI project; this is known as the beginning of Artificial Intelligence.
2012: Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton made a deep-learning breakthrough with AlexNet, starting the modern AI revolution.
2017: Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan Gomez, Łukasz Kaiser, and Illia Polosukhin introduced the Transformer model, which became the base of modern chatbots.
2022: OpenAI released ChatGPT, which made AI popular worldwide. 2023-GPT-4 and 2024-GPT40
Current AI Developers / Leaders: Sam Altman founded OpenAI in 2015, Demis Hassabis founded Google DeepMind in 2010, Dario Amodei founded Anthropic/Claude in 2021, Elon Musk founded xAI in 2023, and Arthur Mensch founded Mistral AI in 2023.
Must Read: Important AI, Machine Learning Techniques are
- Model size
- Training speed
- Randomness/creativity of outputs
- Hardware heat
اس سوال کو وضاحت کے ساتھ پڑھیں
- They self-train
- They struggle to handle uncertainty and scale
- They need no rules
- They learn automatically
اس سوال کو وضاحت کے ساتھ پڑھیں
- Path planning
- Speech recognition + machine translation
- Image segmentation
- File compression
اس سوال کو وضاحت کے ساتھ پڑھیں
- Total accuracy
- Training time
- Number of layers
- Of predicted positives, how many were correct
اس سوال کو وضاحت کے ساتھ پڑھیں
- Speed of model
- Memory usage
- Number of features
- Of actual positives, how many were detected
اس سوال کو وضاحت کے ساتھ پڑھیں
- Bigger models
- Lower cost
- Right to explanation/redress
- Faster inference
اس سوال کو وضاحت کے ساتھ پڑھیں
- Padding
- SMOTE
- Pooling
- Dropout
اس سوال کو وضاحت کے ساتھ پڑھیں
- Supervised regression
- Reinforcement learning
- Encryption
- Unsupervised clustering
اس سوال کو وضاحت کے ساتھ پڑھیں
- It iteratively minimizes the loss function
- It deletes neurons
- It maximizes loss
- It encrypts weights
اس سوال کو وضاحت کے ساتھ پڑھیں
- SGD uses no data
- SGD uses one/few samples per update
- SGD never converges
- SGD ignores gradients