Smartipedia

Markov Chain Monte Carlo

Markov Chain Monte Carlo is a powerful class of algorithms that enables sampling from complex probability distributions by constructing Markov chains whose stationary distributions match the target distributions, revolutionizing computational statistics and probabilistic modeling across numerous scientific fields.

Automorphy

Automorphy is the mathematical concept describing structure-preserving mappings from mathematical objects to themselves, forming the foundation for understanding symmetries across algebra, geometry, and number theory.

fermat's last theorem

Fermat's Last Theorem, which states that *a^n + b^n = c^n* has no positive integer solutions for *n > 2*, was finally proven by Andrew Wiles in 1995 after remaining unsolved for over 350 years, using revolutionary techniques connecting elliptic curves to modular forms.

PAC Learning

PAC Learning is a theoretical framework that provides mathematical foundations for understanding when machine learning problems are solvable, establishing criteria for algorithms to learn concepts that are probably approximately correct with polynomial sample and computational complexity.

training time compute

Training time compute measures the total computational resources required to train machine learning models, which has grown exponentially with modern deep learning systems, creating significant economic and environmental challenges while driving innovation in optimization techniques and specialized hardware.

Leslie Lamport

Leslie Lamport is a pioneering computer scientist whose work on distributed systems, logical clocks, Byzantine fault tolerance, and formal verification has fundamentally shaped modern computing, earning him the Turing Award in 2013.

Test time compute

Test time compute refers to the computational resources required during AI model inference and deployment, encompassing processing power, memory requirements, and optimization strategies that directly impact the practical scalability, cost, and accessibility of artificial intelligence systems.

Sunk cost fallacy

The sunk cost fallacy is a cognitive bias where people continue investing in failing endeavors based on past investments rather than future potential, leading to irrational decision-making and resource misallocation.

Creative destruction

Creative destruction is the economic process by which innovation and competition continuously replace old industries and business models with new, more efficient ones, driving long-term economic growth despite causing short-term disruption.

Externalities

Externalities are economic side effects that occur when market transactions impose uncompensated costs or benefits on third parties, leading to market failures that can be addressed through various policy interventions including taxes, subsidies, regulations, and market-based mechanisms.

Coase theorem

The Coase theorem demonstrates that under ideal conditions with zero transaction costs and well-defined property rights, private parties can negotiate efficient solutions to externality problems without government intervention, regardless of the initial allocation of rights.

Endowment effect

The endowment effect is a cognitive bias where people value objects more highly simply because they own them, leading to systematic differences between willingness to pay and willingness to accept that challenges traditional economic assumptions about rational decision-making.