Reproduce “Learning Hidden Markov Models Using Conditional Samples” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Sham M. Kakade, Akshay Krishnamurthy, Gaurav Mahajan, Cyril Zhang
The authors give a polynomial-time algorithm for learning rank-r sequence distributions from conditional samples. It represents predictions using representative histories and one-step update operators, without recovering hidden states. When fidelity is sufficient, it achieves total-variation error at most ε with probability at least 1−δ. This guarantee applies to finite alphabets; learning all low-rank distributions remains open.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.