We present methods for training high quality object detectors
very quickly. The core contribution is a pair of fast
training algorithms for piece-wise linear classifiers, which
...
Abstract. This paper proposes a general approach named ExpectationMiniMax (EMM) for clustering analysis without knowing the cluster number. It describes the contrast function of Ex...
Max and min operations have interesting properties that facilitate the exchange of information between the symbolic and real-valued domains. As such, neural networks that employ m...
A dynamic geometric data stream consists of a sequence of m insert/delete operations of points from the discrete space {1, . . . , ∆}d [26]. We develop streaming (1 + )-approxim...
We propose an adaptive technique to design the spectrum of an orthogonal frequency division multiplexing (OFDM) waveform to improve the radar's wideband ambiguity function (WA...