Reading group on unsupervised and interactive learning
Fridays at 1
Crowdsourcing
Zhang, Chen, Zhou, Jordan. Spectral methods meet EM: a provably optimal algorithm for crowdsourcing.
Document models
Arora, Li, Liang, Ma, Risteski. Rand-walk: a latent variable model approach to word embeddings.
Arora, Ge, Koehler, Ma, Moitra. Provable algorithms for inference in topic models.
Stratos, Collins, Hsu. Model-based word embeddings from decompositions of count matrices.
Blum, Haghtalab. Generalized topic modeling.
Tensor/spectral methods
Anandkumar, Ge, Hsu, Kakade, Telgarsky. Tensor decompositions for learning latent variable models.
Clustering
Kannan, Kumar. Clustering with spectral norm and the k-means algorithm.
Awasthi, Sheffet. Improved spectral-norm bounds for clustering.
Tang, Monteleoni. On Lloyd's algorithm: new theoretical insights for clustering in practice.
Other unsupervised learning topics
Belkin, Rademacher, Voss. Basis learning as an algorithmic primitive.
Eldridge, Belkin, Wang. Graphons, mergeons, and so on.
Other
Lovasz. Large networks and graph limits.
|