We show evidence of electrical and thermal conductivity percolation in polymer based carbon nanotube (CNT) composites, which follow power law variations with respect to the CNT concentrations in the matrix. The experi...
详细信息
Neurons in the brain form complicated networks through synaptic connections. Traditionally, functional connectivity between neurons has been analyzed using simple metrics such as correlation, which do not provide dire...
详细信息
ISBN:
(纸本)9781424441211
Neurons in the brain form complicated networks through synaptic connections. Traditionally, functional connectivity between neurons has been analyzed using simple metrics such as correlation, which do not provide direction of influence. Recently, an information theoretic measure known as directed information has been proposed as a way to capture directionality in the relationship, thereby moving towards a model of effective connectivity. This measure is grounded upon the concept of Granger causality and can be estimated by modeling neural spike trains as point process generalized linear models. However, the added benefit of using directed information to infer connectivity over conventional methods such as correlation is still unclear. Here, we propose a novel estimation procedure for the directed information. Using physiologically realistic simulations, we demonstrate that directed information can outperform correlation in determining connections between neural spike trains while also providing directionality of the relationship, which cannot be assessed using correlation.
Model-based methods for sequential organization in cochannel speech require pretrained speaker models and often prior knowledge of participating speakers. We propose an unsupervised approach to sequential organization...
详细信息
Model-based methods for sequential organization in cochannel speech require pretrained speaker models and often prior knowledge of participating speakers. We propose an unsupervised approach to sequential organization of cochannel speech. Based on cepstral features, we first cluster voiced speech into two speaker groups by maximizing the ratio of between- and within-group distances penalized by within-group concurrent pitches. To group unvoiced speech, we employ an onset/offset based analysis to generate time-frequency segments. Unvoiced segments are then labeled by the complementary portions of segregated voiced speech. Our method does not require any pretrained model and is computationally simple. Evaluations and comparisons show that the proposed method outperforms a model-based method in terms of speech segregation.
Prior models of speech have been used in robust automatic speech recognition to enhance noisy speech. Typically, a single prior model is trained by pooling the entire training data. In this paper we propose to train m...
详细信息
Prior models of speech have been used in robust automatic speech recognition to enhance noisy speech. Typically, a single prior model is trained by pooling the entire training data. In this paper we propose to train multiple prior models of speech instead of a single prior model. The prior models can be trained based on distinct characteristics of speech. In this study, they are trained based on voicing characteristics. The trained prior models are then used to reconstruct noisy speech. Significant improvements are obtained on the Aurora-4 robust speech recognition task when multiple priors are used; in conjunction with an uncertainty transform technique, multiple priors yield a 13.7% absolute improvement in the average word error rate over directly recognizing noisy speech.
This paper presents a complete set of methods for the performance analysis and design of closed loop systems with nonlinear actuators and sensors (so-called Linear Plant/Nonlinear Instrumentation, or LPNI, systems). A...
详细信息
The study demonstrates how the on-line processing of eye movements in First Person Shooter (FPS) games helps to predict player decisions regarding subsequent actions. Based on action-control theory, we identify distin...
详细信息
ISBN:
(纸本)9781450306805
The study demonstrates how the on-line processing of eye movements in First Person Shooter (FPS) games helps to predict player decisions regarding subsequent actions. Based on action-control theory, we identify distinct cognitive orientations in pre- and post-decisional phases. cognitive orientations differ with regard to the width of attention or "receptiveness": In the pre-decisional phase players process as much information as possible and then focus on implementing intended actions in the post-decisional phase. Participants viewed animated sequences of FPS games and decided which game character to rescue and how to implement their action. Oculomotor data shows a clear distinction between the width of attention in pre- and post-decisional phases, supporting the Rubicon model of action phases. Attention rapidly narrows when the goal intention is formed. We identify a lag of 800-900 ms between goal formation ("cognitive Rubicon") and motor response. Game engines may use this lag to anticipatively respond to actions that players have not executed yet. User interfaces with a gaze-dependent, gazecontrolled anticipation module should thus enhance game character behaviours and make them much "smarter". Copyright 2011 ACM.
An FIR filter is implemented in this work. Enhancing the arithmetic operations of the filter is considered. For the addition operation, the signed-digit number system is utilized. For the multiplication operation, Boo...
详细信息
Detecting pitch values for singing voice in the presence of music accompaniment is challenging but useful for many applications. We propose a trend estimation algorithm to detect the pitch ranges of a singing voice in...
详细信息
Detecting pitch values for singing voice in the presence of music accompaniment is challenging but useful for many applications. We propose a trend estimation algorithm to detect the pitch ranges of a singing voice in each time frame. The detected trend substantially reduces the difficulty of singing pitch detection by reducing a large number of wrong pitch candidates either produced by musical instruments or the overtones of the singing voice. The proposed algorithm can be applied to improve the performance of singing pitch detection. Quantitative evaluations show that proposed trend estimation improves an existing algorithm significantly. The results from the MIREX 2010 competition show that our system achieves the best overall raw-pitch accuracy for vocal songs.
Language comprises a central component of what the co-founder of modern evolutionary theory, Alfred Russell Wallace, called “man's intellectual and moral nature” - the human capacities for creative imagination, ...
详细信息
Language comprises a central component of what the co-founder of modern evolutionary theory, Alfred Russell Wallace, called “man's intellectual and moral nature” - the human capacities for creative imagination, language and symbolism generally, a complex that is sometimes simply called “the human capacity.” This complex seems to have crystallized fairly recently among a small group in East Africa of whom we are all descendants, distinguishing contemporary humans sharply from all other animals, with enormous consequences for the whole of the biological world, as well as for the study of computational cognition. How can we explain this evolutionary leap? On the one hand, common descent has been important in the evolution of the brain, such that avian and mammalian brains may be largely homologous, particularly in the case of brain regions involved in auditory perception, vocalization and auditory memory. On the other hand, there has been convergent evolution of the capacity for auditory-vocal learning, and possibly for structuring of external vocalizations, such that apes lack the abilities that are shared between songbirds and humans. Language's recent evolutionary origin suggests that the computational machinery underlying syntax arose via the introduction of a single, simple, combinatorial operation. Further, the relation of a simple combinatorial syntax to the sensory-motor and thought systems reveals language to be asymmetric in design: while it precisely matches the representations required for inner mental thought, acting as the “glue” that binds together other internal cognitive and sensory modalities, at the same time it poses computational difficulties for externalization, that is, parsing and speech or signed production. Despite this mismatch, language syntax leads directly to the rich cognitive array that marks us as a symbolic species, including mathematics, music, and much more.
Visualization of highway traffic environments in virtual reality plays a key role on fidelity of simulation and validity of driving behaviors in driving simulator studies. This paper describes a procedure to replicate...
详细信息
Visualization of highway traffic environments in virtual reality plays a key role on fidelity of simulation and validity of driving behaviors in driving simulator studies. This paper describes a procedure to replicate a signalized intersection with many complex features into a driving simulator's 3-D databases through building a graphical visual database, constructing a road motion database, and creating simulated traffic. The validity of the visualization was assessed via the comparison of the drivers' speeds in the simulator and the speed data at the real intersection. It was found that both speed data follow normal distributions and have equal means for each intersection approach;however, that the speeds measured in the driving simulator have a larger variability than those measured in the field. The users' subjective evaluation results indicated that 92% subjects could recognize the simulated intersection in the driving simulator experiment. Therefore, the strategy of geo-specific environment modeling would be useful for studying driving behaviors in virtual environments.
暂无评论