Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 5 of 5 for “"Provable Algorithms"”.
-
Provable Algorithms for Learning and Variational Inference in Undirected Graphical Models
… are the first computationally efficient algorithms for provably (1) learning a (possibly ill-conditioned) walk-summable Gaussian Graphical Model from samples, (2) learning a Restricted Boltzmann Machine (or other latent variable Ising model) from data, and (3) performing naive mean-field …
-
Rethinking Algorithm Design for Modern Challenges in Data Science
… why they work so well. Can we design learning algorithms with rigorous guarantees to either match, outperform, or augment these heuristics? In the first part of this thesis, we present new provable algorithms for learning rich function classes like neural networks in natural learning settings …
-
Learning from expert advice framework: Algorithms and applications
… scenarios, to limited resources. We propose provable algorithms for such systems, along with theoretical and experimental results. In the first part of the thesis, we focus our attention to a generalized model of learning from expert advice in which experts could abstain from participating at …
-
Polynomial methods in statistical inference: Theory and practice
… and to design sample optimal and time efficient algorithms with provable guarantees. The first part investigates the problem of property estimation. Consider the problem of estimating the Shannon entropy of a distribution over $k$ elements from $n$ independent samples. We obtain the minimax …
-
Efficient New Computational Protein Design Algorithms, with Applications to Drug Resistance Prediction and HIV Antibody Design
… plasticity. To this end, I have developed algorithms that accurately and more efficiently search large sequence and conformational spaces to compute proteins that satisfy binding affinity, specificity, and stability requirements. Crucially, my algorithms maintain the state-of-the-art in …