Semisupervised Clustering by Queries and Locally Encodable Source Coding

Source coding is the canonical problem of data compression in information theory. In a locally encodable source coding, each compressed bit depends on only few bits input. this paper, we show that recently popular model semisupervised clustering equivalent to locally encodable coding. model, task perform multiclass labeling unlabeled elements. At beginning, can ask parallel set simple queries an oracle who provides (possibly erroneous) binary answers queries. The cannot involve more than two (or fixed constant number of) Now all elements clustering) must be performed based noisy query answers. goal recover correct labelings while minimizing such equivalence codes leads us find lower bounds required variety scenarios. We provide querying schemes pairwise ‘same cluster’ - and AND queries, provable performance guarantees for schemes.