Variational Beta Process Hidden Markov Models with Shared Hidden States for Trajectory Recognition

Hidden Markov model (HMM) is a vital model for trajectory recognition. As the number of hidden states in HMM is important and hard to be determined, many nonparametric methods like hierarchical Dirichlet process HMMs and Beta process HMMs (BP-HMMs) have been proposed to determine it automatically. Among these methods, the sampled BP-HMM models the shared information among different classes, which has been proved to be effective in several trajectory recognition scenes. However, the existing BP-HMM maintains a state transition probability matrix for each trajectory, which is inconvenient for classification. Furthermore, the approximate inference of the BP-HMM is based on sampling methods, which usually takes a long time to converge. To develop an efficient nonparametric sequential model that can capture cross-class shared information for trajectory recognition, we propose a novel variational BP-HMM model, in which the hidden states can be shared among different classes and each class chooses its own hidden states and maintains a unified transition probability matrix. In addition, we derive a variational inference method for the proposed model, which is more efficient than sampling-based methods. Experimental results on a synthetic dataset and two real-world datasets show that compared with the sampled BP-HMM and other related models, the variational BP-HMM has better performance in trajectory recognition.

Взаимосвязь между молекулярной динамикой полистирола и его модификаций и параметрами терагерцового поглощения в ИК спектрах

The IR spectra of polystyrene (PS) and its modifications obtained by replacing hydrogen atoms in the benzene ring with methyl groups and / or chlorine and bromine atoms have been measured and analyzed. All these spectra have an anomalously wide asymmetric absorption band with a maximum in the terahertz range from 40 to 80 cm – 1, depending on the nature of the substituent. It can be attributed to absorption associated with libration (rotational vibrations) of phenolic rings of monomer units, i.e., to absorption by the Poly mechanism. The relationship between the spectral parameters of this absorption and the molecular characteristics of the studied polymers makes it possible to directly analyze the role of the molecular structure and intermolecular forces in the dynamics of their macromolecules. It was found that the heights of potential barriers to libration of monomeric units of PS and its modifications, estimated from the analysis of Poly absorption, are close to the activation energies of low-temperature delta-relaxation in these glassy polymers. Comparison of the barriers to libration of monomeric units of macromolecules of PS and its modifications with the activation barriers of local segmental mobility in the same PS confirms that the beta process as an elementary act of segmental dynamics is caused by correlated librational movement of the chain section, which is statistically independent of neighboring regions. In this sense, the universal delta process is a high-frequency precursor of the beta process in polymers like PS.