Identifying Mixture Models
Leonard J. Schulman (Caltech)
https://simons.berkeley.edu/talks/identifying-mixture-models
Algorithmic Aspects of Causal Inference
We give an algorithm for source identification of a mixture of k product distributions on n bits. This problem (which we call k-MixProd) is a fundamental problem in machine learning with many applications. Our algorithm identifies the source parameters of an identifiable mixture, given, as input, approximate values of multilinear moments; the complexity is ~ exp(k log k) arithmetic operations and similar sample size. Our analysis gives a quantitative version of a qualitative characterization of identifiable sources, that is due to Tahmasebi, Motahari, and Maddah-Ali (2018). Key to our work are a method of ``synthetic bits'' and a condition number lower bound for Hadamard Extensions. Based on joint works with Chaitanya Swamy (Waterloo), Yuval Rabani (Hebrew U), Jian Li (Tsinghua), Spencer Gordon (Caltech) and Bijan Mazaheri (Caltech)
Видео Identifying Mixture Models канала Simons Institute
https://simons.berkeley.edu/talks/identifying-mixture-models
Algorithmic Aspects of Causal Inference
We give an algorithm for source identification of a mixture of k product distributions on n bits. This problem (which we call k-MixProd) is a fundamental problem in machine learning with many applications. Our algorithm identifies the source parameters of an identifiable mixture, given, as input, approximate values of multilinear moments; the complexity is ~ exp(k log k) arithmetic operations and similar sample size. Our analysis gives a quantitative version of a qualitative characterization of identifiable sources, that is due to Tahmasebi, Motahari, and Maddah-Ali (2018). Key to our work are a method of ``synthetic bits'' and a condition number lower bound for Hadamard Extensions. Based on joint works with Chaitanya Swamy (Waterloo), Yuval Rabani (Hebrew U), Jian Li (Tsinghua), Spencer Gordon (Caltech) and Bijan Mazaheri (Caltech)
Видео Identifying Mixture Models канала Simons Institute
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Adversarial Examples in Deep Learning
Quantum Advantage Without Structure
Remarks on the Discrete Cube
Fast Reinforcement Learning With Generalized Policy Updates
"The Problem with Qubits"
Stability and Learning in Repeated Games
Planning and Markov Decision Processes Part 1 (reupload)
Panel | Quantum Colloquium
Tractable Probabilistic Circuits
Algorithmic Fairness From The Lens Of Causality And Information Theory
Quantum Computing and Simulation with Atoms
Project CETI Next Steps: Industrial-Scale Whale Bioacoustic Data Collection and Analysis
Causal Matrix Completion
Learning and Incentives (Part I)
Tutorial: Implicit Bias I
Using Theories of Decision-Making Under Uncertainty to Improve Data Visualization
Equivariant RL
Berkeley in the 80s, Episode 1: Shafi Goldwasser
The Role of Conventions in Adaptive Human-AI Interaction
Fully Linear PCPs and their Cryptographic Applications
Mathematical Imaging: From Geometric PDEs and Variational Modeling to Deep Learning for Images