The UCT algorithm learns a value function online using sample-based search. The TD() algorithm can learn a value function offline for the on-policy distribution. We consider three...
Bottom-up segmentation based only on low-level cues is a notoriously difficult problem. This difficulty has lead to recent top-down segmentation algorithms that are based on class-...
We present a constraint-based algorithm for the synthesis of invariants expressed in the combined theory of linear arithmetic and uninterpreted function symbols. Given a set of pro...
Dirk Beyer, Thomas A. Henzinger, Rupak Majumdar, A...
"Pattern recognition techniques are concerned with the theory and algorithms of putting abstract objects, e.g., measurements made on physical objects, into categories. Typical...
A multiple description scalar quantization (MDSQ) based coding system can be regarded as a source coder (quantizer) followed by a channel coder, i.e. the combination of index and ...