Online Unsupervised Pattern Discovery in Speech Using Parallelization

Segmental dynamic time warping (DTW) has been demonstrated to be a useful technique for finding acoustic similarity scores between segments of two speech utterances. Due to its high computational requirements, it had to be computed in an offline manner, limiting the applications of the technique. In this paper, we present results of parallelization of this task by distributing the workload in either a static or dynamic way on an 8-processor cluster and discuss the trade-offs among different distribution schemes. We show that online unsupervised pattern discovery using segmental DTW is plausible with as low as 8 processors. This brings the task within reach of today's general purpose multi-core servers. We also show results on a 32-processor system, and discuss factors affecting scalability of our methods.

Robustness Study of a Dynamic Inversion Control Law for a High Performance Aircraft

In the design of °ight control system modeling uncertainties in the form of param-eter variations is one of the major problems. It is even more critical for high performance aircrafts,since such aircrafts are purposefully designed unstable to enhance their performance (especially ma-neuverability). Hence the °ight control system needs to be quite e®ective in both assuring accurate tracking of pilot commands, while simultaneously assuring overall stability of the aircraft. In addi-tion, the control system must also be su±ciently robust to cater for possible parameter variations and inaccuracies . The primary aim of this paper is to carry out a robustness study of a dynamic inversion based nonlinear control design for a high performance aircraft, which has been developed recently [1].

Open source drug discovery- A new paradigm of collaborative research in tuberculosis drug development

It is being realized that the traditional closed-door and market driven approaches for drug discovery may not be the best suited model for the diseases of the developing world such as tuberculosis and malaria, because most patients suffering from these diseases have poor paying capacity. To ensure that new drugs are created for patients suffering from these diseases, it is necessary to formulate an alternate paradigm of drug discovery process. The current model constrained by limitations for collaboration and for sharing of resources with confidentiality hampers the opportunities for bringing expertise from diverse fields. These limitations hinder the possibilities of lowering the cost of drug discovery. The Open Source Drug Discovery project initiated by Council of Scientific and Industrial Research, India has adopted an open source model to power wide participation across geographical borders. Open Source Drug Discovery emphasizes integrative science through collaboration, open-sharing, taking up multi-faceted approaches and accruing benefits from advances on different fronts of new drug discovery. Because the open source model is based on community participation, it has the potential to self-sustain continuous development by generating a storehouse of alternatives towards continued pursuit for new drug discovery. Since the inventions are community generated, the new chemical entities developed by Open Source Drug Discovery will be taken up for clinical trial in a non-exclusive manner by participation of multiple companies with majority funding from Open Source Drug Discovery. This will ensure availability of drugs through a lower cost community driven drug discovery process for diseases afflicting people with poor paying capacity. Hopefully what LINUX the World Wide Web have done for the information technology, Open Source Drug Discovery will do for drug discovery. (C) 2011 Elsevier Ltd. All rights reserved.

The standard enthalpy and entropy of formation of Rh2O3 - A third-law optimization

The standard Gibbs energy of formation of Rh203 at high temperature has been determined recently with high precision. The new data are significantly different from those given in thermodynamic compilations.Accurate values for enthalpy and entropy of formation at 298.15 K could not be evaluated from the new data,because reliable values for heat capacity of Rh2O3 were not available. In this article, a new measurement of the high temperature heat capacity of Rh2O3 using differential scanning calorimetry (DSC) is presented.The new values for heat capacity also differ significantly from those given in compilations. The information on heat capacity is coupled with standard Gibbs energy of formation to evaluate values for standard enthalpy and entropy of formation at 289.15 K using a multivariate analysis. The results suggest a major revision in thermodynamic data for Rh2O3. For example, it is recommended that the standard entropy of Rh203 at 298.15 K be changed from 106.27 J mol-' K-'given in the compilations of Barin and Knacke et al. to 75.69 J mol-' K". The recommended revision in the standard enthalpy of formation is from -355.64 kJ mol-'to -405.53 kJ mol".

Generalized distributive law for ML decoding of STBCs

The Generalized Distributive Law (GDL) is a message passing algorithm which can efficiently solve a certain class of computational problems, and includes as special cases the Viterbi's algorithm, the BCJR algorithm, the Fast-Fourier Transform, Turbo and LDPC decoding algorithms. In this paper GDL based maximum-likelihood (ML) decoding of Space-Time Block Codes (STBCs) is introduced and a sufficient condition for an STBC to admit low GDL decoding complexity is given. Fast-decoding and multigroup decoding are the two algorithms used in the literature to ML decode STBCs with low complexity. An algorithm which exploits the advantages of both these two is called Conditional ML (CML) decoding. It is shown in this paper that the GDL decoding complexity of any STBC is upper bounded by its CML decoding complexity, and that there exist codes for which the GDL complexity is strictly less than the CML complexity. Explicit examples of two such families of STBCs is given in this paper. Thus the CML is in general suboptimal in reducing the ML decoding complexity of a code, and one should design codes with low GDL complexity rather than low CML complexity.

A unified view of the Apriori-based algorithms for frequent episode discovery

Frequent episode discovery framework is a popular framework in temporal data mining with many applications. Over the years, many different notions of frequencies of episodes have been proposed along with different algorithms for episode discovery. In this paper, we present a unified view of all the apriori-based discoverymethods for serial episodes under these different notions of frequencies. Specifically, we present a unified view of the various frequency counting algorithms. We propose a generic counting algorithm such that all current algorithms are special cases of it. This unified view allows one to gain insights into different frequencies, and we present quantitative relationships among different frequencies.Our unified view also helps in obtaining correctness proofs for various counting algorithms as we show here. It also aids in understanding and obtaining the anti-monotonicity properties satisfied by the various frequencies, the properties exploited by the candidate generation step of any apriori-based method. We also point out how our unified view of counting helps to consider generalization of the algorithm to count episodes with general partial orders.