作者:
Weinberger, NirFeder, MeirMIT
IDSS 77 Massachusetts Ave Cambridge MA 02139 USA MIT
LIDS 77 Massachusetts Ave Cambridge MA 02139 USA Tel Aviv Univ
Sch Elect Engn IL-69978 Tel Aviv Israel
The k-vectors algorithm for learning regression functions proposed here is akin to the well-known k-means algorithm. Both algorithms partition the feature space, but unlike the k-means algorithm, the k-vectors algorit...
详细信息
ISBN:
(纸本)9781728131511
The k-vectors algorithm for learning regression functions proposed here is akin to the well-known k-means algorithm. Both algorithms partition the feature space, but unlike the k-means algorithm, the k-vectors algorithm aims to reconstruct the response rather than the feature. The partitioning rule of the algorithm is based on maximizing the correlation (inner product) of the feature vector with a set of k vectors, and generates polyhedral cells, similar to the ones generated by the nearest-neighbor rule of the k-means algorithm. Similarly to k-means, the learning algorithm alternates between two types of steps. In the first type of steps, k labels are determined via a centroid-type rule (in the response space), which uses a surrogate hinge-type loss function to the mean squared error loss function. In the second type of steps, the k vectors which determine the partition are updated according to a multiclass classification rule, in the spirit of support vector machines. It is proved that both steps of the algorithm only require solving convex optimization problems, and that the algorithm is empirically consistent - as the length of the training sequence increases to infinity, fixedpoints of the empirical version of the algorithm tend to fixed points of the population version of the algorithm. Learnability of the predictor class posit by the algorithm is also established.
A distributed system which comprises independent sensors and a fusion center is considered. Optimum quantization of sensor observations by Chernoff Information (CI) and deflection criterion (DC) for detection purpose ...
详细信息
ISBN:
(纸本)9781479930838
A distributed system which comprises independent sensors and a fusion center is considered. Optimum quantization of sensor observations by Chernoff Information (CI) and deflection criterion (DC) for detection purpose is studied. Chernoff information of overall system is derived, and locally optimum quantization of sensor observations is proved to achieve global optimality. A relationship between DC and CI of quantized observations is established. A novel iterative algorithm which works similar to lloyd-max algorithm is proposed to find optimum quantization intervals. Results indicate that optimum 2-bit quantization gives detection performance that is close to what is achievable by collecting soft reports from sensors.
暂无评论