A comparative study of speaker adaptation methods

For the problem of speaker adaptation in speech recognition, the performance depends on the availability of adaptation data. In this paper, we have compared several existing speaker adaptation methods, viz. maximum likelihood linear regression (MLLR), eigenvoice (EV), eigenspace-based MLLR (EMLLR), segmental eigenvoice (SEV) and hierarchical eigenvoice (HEV) based methods. We also develop a new method by modifying the existing HEV method for achieving further performance improvement in a limited available data scenario. In the sense of availability of adaptation data, the new modified HEV (MHEV) method is shown to perform better than all the existing methods throughout the range of operation except the case of MLLR at the availability of more adaptation data.

Reuse of treated sewage in Delhi city: Microbial evaluation of STPs and reuse options

Microbiological quality of the treated wastewater is an important parameter for its reuse. The data oil the Fecal Coliform (FC) and Fecal Streptococcus (FS) at different stages of treatment in the Sewage Treatment Plants (STPs) in Delhi watershed is not available, therefore in the present study microbial profiling of STPs was carried out to assess the effluent quality for present and future reuse options. This Study further evaluates the water quality profiles at different stages of treatment for 16 STPs in Delhi city. These STPs are based on conventional Activated Sludge Process (ASP), extended aeration, physical, chemical and biological treatment (BIOFORE), Trickling Filter and Oxidation Pond. The primary effluent quality produced from most of the STPs was suitable for Soil Aquifer Treatment (SAT). Extended Hydraulic Retention Time (HRT) as a result Of low inflow to the STPS Was responsible for high turbidity, COD and BODs removal. Conventional ASP based STPs achieved 1.66 log FC and 1.06 log FS removal. STPs with extended aeration treatment process produced better quality effluent with maximum 4 log order reduction in FC and FS levels. ``Kondli'' and ``Nilothi'' STPs employing ASP, produced better quality secondary effluent as compared to other STPs based oil similar treatment process. Oxidation Pond based STPs showed better FC and FS removals, whereas good physiochemical quality was achieved during the first half of the treatment. Based upon physical, chemical and microbiological removal efficiencies, actual integrated efficiency (IEa) of each STP was determined to evaluate its Suitability for reuse for irrigation purposes. Except Mehrauli'' and ``Oxidation Pond'', effluents from all other STPs require tertiary treatment for further reuse. Possible reuse options, depending Upon the geographical location, proximity of facilities of potential users based oil the beneficial uses, and sub-soil types, etc. for the Delhi city have been investigated, which include artificial groundwater recharge, aquaculture, horticulture and industrial uses Such as floor washing, boiler feed, and cooling towers, etc. (C) 2009 Elsevier B.V. All rights reserved.

SPSA algorithms with measurement reuse

Four algorithms, all variants of Simultaneous Perturbation Stochastic Approximation (SPSA), are proposed. The original one-measurement SPSA uses an estimate of the gradient of objective function L containing an additional bias term not seen in two-measurement SPSA. As a result, the asymptotic covariance matrix of the iterate convergence process has a bias term. We propose a one-measurement algorithm that eliminates this bias, and has asymptotic convergence properties making for easier comparison with the two-measurement SPSA. The algorithm, under certain conditions, outperforms both forms of SPSA with the only overhead being the storage of a single measurement. We also propose a similar algorithm that uses perturbations obtained from normalized Hadamard matrices. The convergence w.p. 1 of both algorithms is established. We extend measurement reuse to design two second-order SPSA algorithms and sketch the convergence analysis. Finally, we present simulation results on an illustrative minimization problem.

Climate change and forests in India: adaptation opportunities and challenges

Climate change is projected to lead to shift of forest types leading to irreversible damage to forests by rendering several species extinct and potentially affecting the livelihoods of local communities and the economy. Approximately 47% and 42% of tropical dry deciduous grids are projected to undergo shifts under A2 and B2 SRES scenarios respectively, as opposed to less than 16% grids comprising of tropical wet evergreen forests. Similarly, the tropical thorny scrub forest is projected to undergo shifts in majority of forested grids under A2 (more than 80%) as well as B2 scenarios (50% of grids). Thus the forest managers and policymakers need to adapt to the ecological as well as the socio-economic impacts of climate change. This requires formulation of effective forest management policies and practices, incorporating climate concerns into long-term forest policy and management plans. India has formulated a large number of innovative and progressive forest policies but a mechanism to ensure effective implementation of these policies is needed. Additional policies and practices may be needed to address the impacts of climate change. This paper discusses an approach and steps involved in the development of an adaptation framework as well as policies, strategies and practices needed for mainstreaming adaptation to cope with projected climate change. Further, the existing barriers which may affect proactive adaptation planning given the scale, accuracy and uncertainty associated with assessing climate change impacts are presented.

Forest policies and programs affecting vulnerability and adaptation to climate change

Due to large scale afforestation programs and forest conservation legislations, India's total forest area seems to have stabilized or even increased. In spite of such efforts, forest fragmentation and degradation continues, with forests being subject to increased pressure due to anthropogenic factors. Such fragmentation and degradation is leading to the forest cover to change from very dense to moderately dense and open forest and 253 km(2) of very dense forest has been converted to moderately dense forest, open forest, scrub and non-forest (during 2005-2007). Similarly, there has been a degradation of 4,120 km(2) of moderately dense forest to open forest, scrub and non-forest resulting in a net loss of 936 km(2) of moderately dense forest. Additionally, 4,335 km(2) of open forest have degraded to scrub and non-forest. Coupled with pressure due to anthropogenic factors, climate change is likely to be an added stress on forests. Forest sector programs and policies are major factors that determine the status of forests and potentially resilience to projected impacts of climate change. An attempt is made to review the forest policies and programs and their implications for the status of forests and for vulnerability of forests to projected climate change. The study concludes that forest conservation and development policies and programs need to be oriented to incorporate climate change impacts, vulnerability and adaptation.

A macroeconomic analysis of adaptation to climate change impacts on forests in India

We examine the potential for adaptation to climate change in Indian forests, and derive the macroeconomic implications of forest impacts and adaptation in India. The study is conducted by integrating results from the dynamic global vegetation model IBIS and the computable general equilibrium model GRACE-IN, which estimates macroeconomic implications for six zones of India. By comparing a reference scenario without climate change with a climate impact scenario based on the IPCC A2-scenario, we find major variations in the pattern of change across zones. Biomass stock increases in all zones but the Central zone. The increase in biomass growth is smaller, and declines in one more zone, South zone, despite higher stock. In the four zones with increases in biomass growth, harvest increases by only approximately 1/3 of the change in biomass growth. This is due to two market effects of increased biomass growth. One is that an increase in biomass growth encourages more harvest given other things being equal. The other is that more harvest leads to higher supply of timber, which lowers market prices. As a result, also the rent on forested land decreases. The lower prices and rent discourage more harvest even though they may induce higher demand, which increases the pressure on harvest. In a less perfect world than the model describes these two effects may contribute to an increase in the risk of deforestation because of higher biomass growth. Furthermore, higher harvest demands more labor and capital input in the forestry sector. Given total supply of labor and capital, this increases the cost of production in all the other sectors, although very little indeed. Forestry dependent communities with declining biomass growth may, however, experience local unemployment as a result.

Symmetry adaptation of correlated states in the valence bond basis

Symmetry?adapted linear combinations of valence?bond (VB) diagrams are constructed for arbitrary point groups and total spin S using diagrammatic VB methods. VB diagrams are related uniquely to invariant subspaces whose size reflects the number of group elements; their nonorthogonality leads to sparser matrices and is fully incorporated into a binary integer representation. Symmetry?adapated linear combinations of VB diagrams are constructed for the 1764 singlets of a half?filled cube of eight sites, the 2.8 million ??electron singlets of anthracene, and for illustrative S?0 systems.

Low-power pipelined LMS adaptive filter architectures with minimal adaptation delay

The use of delayed coefficient adaptation in the least mean square (LMS) algorithm has enabled the design of pipelined architectures for real-time transversal adaptive filtering. However, the convergence speed of this delayed LMS (DLMS) algorithm, when compared with that of the standard LMS algorithm, is degraded and worsens with increase in the adaptation delay. Existing pipelined DLMS architectures have large adaptation delay and hence degraded convergence speed. We in this paper, first present a pipelined DLMS architecture with minimal adaptation delay for any given sampling rate. The architecture is synthesized by using a number of function preserving transformations on the signal flow graph representation of the DLMS algorithm. With the use of carry-save arithmetic, the pipelined architecture can support high sampling rates, limited only by the delay of a full adder and a 2-to-1 multiplexer. In the second part of this paper, we extend the synthesis methodology described in the first part, to synthesize pipelined DLMS architectures whose power dissipation meets a specified budget. This low-power architecture exploits the parallelism in the DLMS algorithm to meet the required computational throughput. The architecture exhibits a novel tradeoff between algorithmic performance (convergence speed) and power dissipation. (C) 1999 Elsevier Science B.V. All rights resented.

Surface chemical characterisation of Paenibacillus polymyxa before and after adaptation to sulfide minerals

A heterotroph Paenibacillus polymyxa bacteria is adapted to pyrite, chalcopyrite, galena and sphalerite minerals by repeated subculturing the bacteria in the presence of the mineral until their growth characteristics became similar to the growth in the absence of mineral. The unadapted and adapted bacterial surface have been chemically characterised by zeta-potential, contact angle, adherence to hydrocarbons and FT-IR spectroscopic studies. The surface free energies of bacteria have been calculated by following the equation of state and surface tension component approaches. The aim of the present paper is to understand the changes in surface chemical properties of bacteria during adaptation to sulfide minerals and the projected consequences in bioflotation and bioflocculation processes. The mineral-adapted cells became more hydrophilic as compared to unadapted cells. There are no significant changes in the surface charge of bacteria before and after adaptation, and all the bacteria exhibit an iso-electric point below pH 2.5. The contact angles are observed to be more reliable for hydrophobicity assessment than the adherence to hydrocarbons. The Lifschitz–van der Waals/acid–base approach to calculate surface free energy is found to be relevant for mineral–bacteria interactions. The diffuse reflectance FT-IR absorbance bands for all the bacteria are the same illustrating similar surface chemical composition. However, the intensity of the bands for unadapted and adapted cells is significantly varied and this is due to different amounts of bacterial secretions underlying different growth conditions.

On the Limits of Spatial Reuse and Cooperative Communication for Dense Wireless Networks

We consider a dense ad hoc wireless network comprising n nodes confined to a given two dimensional region of fixed area. For the Gupta-Kumar random traffic model and a realistic interference and path loss model (i.e., the channel power gains are bounded above, and are bounded below by a strictly positive number), we study the scaling of the aggregate end-to-end throughput with respect to the network average power constraint, P macr, and the number of nodes, n. The network power constraint P macr is related to the per node power constraint, P macr, as P macr = np. For large P, we show that the throughput saturates as Theta(log(P macr)), irrespective of the number of nodes in the network. For moderate P, which can accommodate spatial reuse to improve end-to-end throughput, we observe that the amount of spatial reuse feasible in the network is limited by the diameter of the network. In fact, we observe that the end-to-end path loss in the network and the amount of spatial reuse feasible in the network are inversely proportional. This puts a restriction on the gains achievable using the cooperative communication techniques studied in and, as these rely on direct long distance communication over the network.

Power and Discrete Rate Adaptation for Energy Harvesting Wireless Nodes

Energy Harvesting (EH) nodes, which harvest energy from the environment in order to communicate over a wireless link, promise perpetual operation of a wireless network with battery-powered nodes. In this paper, we address the throughput optimization problem for a rate-adaptive EH node that chooses its rate from a set of discrete rates and adjusts its power depending on its channel gain and battery state. First, we show that the optimal throughput of an EH node is upper bounded by the throughput achievable by a node that is subject only to an average power constraint. We then propose a simple transmission scheme for an EH node that achieves an average throughput close to the upper bound. The scheme's parameters can be made to account for energy overheads such as battery non-idealities and the energy required for sensing and processing. The effect of these overheads on the average throughput is also analytically characterized.

An embedded grid adaptation strategy for unstructured data based finite volume computations

A grid adaptation strategy for unstructured data based codes, employing a combination of hexahedral and prismatic elements, generalizable to tetrahedral and pyramidal elements has been developed.

Enhancing Speedup in Network Processing Applications by Exploiting Instruction Reuse with Flow Aggregation

Instruction reuse is a microarchitectural technique that improves the execution time of a program by removing redundant computations at run-time. Although this is the job of an optimizing compiler, they do not succeed many a time due to limited knowledge of run-time data. In this paper we examine instruction reuse of integer ALU and load instructions in network processing applications. Specifically, this paper attempts to answer the following questions: (1) How much of instruction reuse is inherent in network processing applications?, (2) Can reuse be improved by reducing interference in the reuse buffer?, (3) What characteristics of network applications can be exploited to improve reuse?, and (4) What is the effect of reuse on resource contention and memory accesses? We propose an aggregation scheme that combines the high-level concept of network traffic i.e. "flows" with a low level microarchitectural feature of programs i.e. repetition of instructions and data along with an architecture that exploits temporal locality in incoming packet data to improve reuse. We find that for the benchmarks considered, 1% to 50% of instructions are reused while the speedup achieved varies between 1% and 24%. As a side effect, instruction reuse reduces memory traffic and can therefore be considered as a scheme for low power.