Uncategorized
Counterfactuals: Reasoning About What Did Not Happen
To ask what would have happened is to reason about a world that never was.
Stories by Mahendra
Uncategorized
The Frame Problem: Knowing What Stays the Same
When something changes, most things do not.
Uncategorized
Credit Assignment: Which Action Earned the Reward
A reward arrives, but which of many past choices deserves the thanks? When cause and consequence are separated by time and tangled with chance, assigning credit is the central puzzle of learning to act..
Uncategorized
Catastrophic Forgetting: The Price of Learning Anew
To learn a new thing can be to lose an old one.
Uncategorized
Disentanglement: Factoring the World Into Its Causes
The world's appearances arise from a few underlying factors, mixed together.
Uncategorized
Representation: Learning What to Represent
Before a problem can be solved it must be described, and the description decides the difficulty.
Uncategorized
The Ergodic Hypothesis: Time Averages and Ensemble Averages
To predict a system's behavior one can watch it for a long time or imagine many copies at once.
Uncategorized
Learning From Finite Data: The Problem of Induction
No finite set of examples logically compels a general conclusion.
Uncategorized
The Constructive Role of Noise
Noise is usually the enemy of signal, yet in learning and in nature it can help.
Uncategorized
Optimization Landscapes: Hills, Valleys, and Saddles
To train a model is to descend a landscape of error, seeking its lowest point.