Search

Recent advances in technology bring about novel operating environments where the role of human participants is reduced to passive observation. While opening new frontiers in productivity and lifestyle, such environments also create hazards related to the inability of human individuals to maintain focus and concentration during passive control tasks. A passive brain-computer interface for monitoring mental attention states of human individuals (focused, unfocused, and drowsy) by using electroencephalographic (EEG) brain activity imaging and machine learning data analysis methods is developed in this work. An EEG data processing pipeline and a machine learning mental state detection algorithm using the Support Vector Machine (SVM) method were designed and compared with k-Nearest Neighbor and Adaptive Neuro-Fuzzy System methods. To collect 25 h of EEG data from 5 participants, a classic EEG headset was modified. We found that the changes in EEG activity in frontal and parietal lobes occurring at 1–5 Hz and 10–15 Hz frequency bands were associated with the changes in individuals’ attention state. We demonstrated the ability to use such changes to identify individuals’ attention state with 96.70% (best) and 91.72% (avg.) accuracy in experimental settings using a version of continuous performance task with SVM-based mental state detector. The findings help guide the design of future systems for monitoring the state of human individuals by means of EEG brain activity data.

The study of keyword spotting, a subfield within the broader field of speech recognition that centers around identifying individual keywords in speech audio, has gained particular importance in recent years with the rise of personal voice assistants such as Alexa. As voice assistants aim to rapidly expand to support new languages, keywords, and use cases, stakeholders face the issue of limited training data for these unseen scenarios. This paper details some initial exploration into the application of Text-To-Speech (TTS) audio as a "helper" tool for training keyword spotters in these low-resource scenarios. In the experiments studied in this paper, the careful mixing of TTS audio with human speech audio during training led to a reduction of over 11% in the detection-error-tradeoff (DET) area under the curve (AUC) metric.

Small footprint embedded devices require keyword spotters (KWS) with small model size and detection latency for enabling voice assistants. Such a keyword is often referred to as \textit{wake word} as it is used to wake up voice assistant enabled devices. Together with wake word detection, accurate estimation of wake word endpoints (start and end) is an important task of KWS. In this paper, we propose two new methods for detecting the endpoints of wake words in neural KWS that use single-stage word-level neural networks. Our results show that the new techniques give superior accuracy for detecting wake words' endpoints of up to 50 msec standard error versus human annotations, on par with the conventional Acoustic Model plus HMM forced alignment. To our knowledge, this is the first study of wake word endpoints detection methods for single-stage neural KWS., Comment: Proceedings of INTERSPEECH

Wake word (WW) spotting is challenging in far-field not only because of the interference in signal transmission but also the complexity in acoustic environment. Traditional WW model training requires large amount of in-domain WW-specific data with substantial human annotations. This prevents the model building in the situation of lacking such data. In this paper we present data-efficient solutions to address the challenges in WW modeling, such as domain-mismatch, noisy conditions, limited annotation, etc. Our proposed system is composed of a multi-condition training pipeline with stratified data augmentation, which improves the model robustness to a variety of predefined acoustic conditions, together with a semi-supervised learning pipeline to extract the WW and adversarial examples from untranscribed speech corpus. Starting from only 10 hours of domain-mismatched WW audio, we are able to enlarge and enrich the training dataset by 20-100 times to capture the complexity in acoustic environments. Our experiments on real user data show that the proposed solutions can achieve comparable performance of a production-grade model by saving 97% of the amount of WW-specific data to collect and 86% of the bandwidth for annotation.

In this paper, we investigate novel quantization approaches to reduce memory and computational footprint of deep neural network (DNN) based keyword spotters (KWS). We propose a new method for KWS offline and online quantization, which we call dynamic quantization, where we quantize DNN weight matrices column-wise, using each column's exact individual min-max range, and the DNN layers' inputs and outputs are quantized for every input audio frame individually, using the exact min-max range of each input and output vector. We further apply a new quantization-aware training approach that allows us to incorporate quantization errors into KWS model during training. Together, these approaches allow us to significantly improve the performance of KWS in 4-bit and 8-bit quantized precision, achieving the end-to-end accuracy close to that of full precision models while reducing the models' on-device memory footprint by up to 80%.

Recent advancements in brain computer interfaces (BCI) have demonstrated control of robotic systems by mental processes alone. Together with invasive BCI, electroencephalographic (EEG) BCI represent an important direction in the development of BCI systems. In the context of EEG BCI, the processing of EEG data is the key challenge. Unfortunately, advances in that direction have been complicated by a lack of large and uniform datasets that could be used to design and evaluate different data processing approaches. In this work, we release a large set of EEG BCI data collected during the development of a slow cortical potentials-based EEG BCI. The dataset contains 60 h of EEG recordings, 13 participants, 75 recording sessions, 201 individual EEG BCI interaction session-segments, and over 60 000 examples of motor imageries in 4 interaction paradigms. The current dataset presents one of the largest EEG BCI datasets publically available to date. Machine-accessible metadata file describing the reported data (ISA-Tab format)

Operators who use a vehicle have less control load with fast improvements of robotic and autonom systems so that situation causes losing of attention an operator while important control processes. In this paper, a passive brain computer interface for monitoring mental attention state of human individuals by using electroencephalographic (EEG) brain activity imaging is developed using a machine learning data analysis method Support Vector Machine. Also a mental state detection system using EEG data is evolved as well. It has been determined that changes in EEG activity in the frontal and parietal lobes occurring in the 1–5 Hz and 1015 Hz frequency bands are associated with changes in attention state. Such changes were detected with 90% to 95% accuracy in experimental settings. The results of the work done will guide the design of future systems to monitor the status of the operators via EEG brain activity data.

The high dimensionality of typical feature spaces, small size of training data, and time-varying nature of EEG signals are the current difficulties of brain computer interfaces that significantly limit their applications in clinical practice. Brain-computer interface data processing approach using dynamic models can be expected to be useful for solving these problems. In this study we aim to develop a dynamic Hidden Markov Model of EEG signals for the use in EEG-based brain computer interfaces. We present a dynamic Hidden Markov Model of EEG signals and develop the algorithm for calculating its parameters from real data.

We investigate the properties of recently proposed "shotgun" sampling approach for the common inputs problem in the functional estimation of neuronal connectivity. We study the asymptotic correctness, the speed of convergence, and the data size requirements of such an approach. We show that the shotgun approach can be expected to allow the inference of complete connectivity matrix in large neuronal populations under some rather general conditions. However, we find that the posterior error of the shotgun connectivity estimator grows quickly with the size of unobserved neuronal populations, the square of average connectivity strength, and the square of observation sparseness. This implies that the shotgun connectivity estimation will require significantly larger amounts of neuronal activity data whenever the number of neurons in observed neuronal populations remains small. We present a numerical approach for solving the shotgun estimation problem in general settings and use it to demonstrate the shotgun connectivity inference in the examples of simulated synfire and weakly coupled cortical neuronal networks.

Physical organization of the nervous system is a topic of perpetual interest in neuroscience. Despite significant achievements here in the past, many details of the nervous system organization and its role in animals' behavior remain obscure, while the problem of complete connectivity reconstructions has recently re-emerged as one of the major directions in neuroscience research (i.e. connectomics). We describe a novel paradigm for connectomics reconstructions that can yield connectivity maps with high resolution, high speed of imaging and data analysis, and significant robustness to errors. In essence, we propose that physical connectivity in a neural circuit can be sampled using anatomical fluorescent synaptic markers localized to different parts of the neural circuit with a technique for randomized genetic targeting, and that high-resolution connectivity maps can be extracted from such datasets. We describe how such an approach can be implemented and how neural connectivity matrix can be reconstructed statistically using the methods of Compressive Sensing. Use of Compressive Sensing is the key to allow accurate neural connectivity reconstructions with orders-of-magnitude smaller volumes of experimental data. We test described approach on simulations of neural connectivity reconstruction experiments in C. elegans, where real neural wiring diagram is available from past electron microscopy studies. We show that such wiring diagram can be in principle re-obtained using described approach in 1-7 days of imaging and data analysis. Alternative approaches would require currently at least 1-2 years to produce a single comparable reconstruction. We discuss possible applications of described approach in larger organisms such as Drosophila.

One of surprising recent discoveries in biology is that the gene and protein expression profiles in cells with identical genetic and environmental makeup can exhibit large variability. The nature and the significance of this variability had been posed as one of the current fundamental questions in biology. In this letter, we argue that the observed variability in cellular gene and protein expression can be understood as an outcome of homeostatic regulation mechanisms controlling the gene and protein expression profiles.

We analyze the scaling and cost-performance characteristics of current and projected connectomics approaches, with reference to the potential implications of recent advances in diverse contributing fields. Three generalized strategies for dense connectivity mapping at the scale of whole mammalian brains are considered: electron microscopic axon tracing, optical imaging of combinatorial molecular markers at synapses, and bulk DNA sequencing of trans-synaptically exchanged nucleic acid barcode pairs. Due to advances in parallel-beam instrumentation, whole mouse brain electron microscopic image acquisition could cost less than $100 million, with total costs presently limited by image analysis to trace axons through large image stacks. Optical microscopy at 50 to 100 nm isotropic resolution could potentially read combinatorially multiplexed molecular information from individual synapses, which could indicate the identifies of the pre-synaptic and post-synaptic cells without relying on axon tracing. An optical approach to whole mouse brain connectomics may be achievable for less than $10 million and could be enabled by emerging technologies to sequence nucleic acids in-situ in fixed tissue via fluorescent microscopy. Novel strategies relying on bulk DNA sequencing, which would extract the connectome without direct imaging of the tissue, could produce a whole mouse brain connectome for $100k to $1 million or a mouse cortical connectome for $10k to $100k. Anticipated further reductions in the cost of DNA sequencing could lead to a $1000 mouse cortical connectome.

In recent years, the problem of reconstructing the connectivity in large neural circuits ("connectomics") has re-emerged as one of the main objectives of neuroscience. Classically, reconstructions of neural connectivity have been approached anatomically, using electron or light microscopy and histological tracing methods. This paper describes a statistical approach for connectivity reconstruction that relies on relatively easy-to-obtain measurements using fluorescent probes such as synaptic markers, cytoplasmic dyes, transsynaptic tracers, or activity-dependent dyes. We describe the possible design of these experiments and develop a Bayesian framework for extracting synaptic neural connectivity from such data. We show that the statistical reconstruction problem can be formulated naturally as a tractable L 1-regularized quadratic optimization. As a concrete example, we consider a realistic hypothetical connectivity reconstruction experiment in C. elegans, a popular neuroscience model where a complete wiring diagram has been previously obtained based on long-term electron microscopy work. We show that the new statistical approach could lead to an orders of magnitude reduction in experimental effort in reconstructing the connectivity in this circuit. We further demonstrate that the spatial heterogeneity and biological variability in the connectivity matrix--not just the "average" connectivity--can also be estimated using the same method.

Monte Carlo approaches have recently been proposed to quantify connectivity in neuronal networks. The key problem is to sample from the conditional distribution of a single neuronal spike train, given the activity of the other neurons in the network. Dependencies between neurons are usually relatively weak; however, temporal dependencies within the spike train of a single neuron are typically strong. In this paper we develop several specialized Metropolis--Hastings samplers which take advantage of this dependency structure. These samplers are based on two ideas: (1) an adaptation of fast forward--backward algorithms from the theory of hidden Markov models to take advantage of the local dependencies inherent in spike trains, and (2) a first-order expansion of the conditional likelihood which allows for efficient exact sampling in the limit of weak coupling between neurons. We also demonstrate that these samplers can effectively incorporate side information, in particular, noisy fluorescence observations in the context of calcium-sensitive imaging experiments. We quantify the efficiency of these samplers in a variety of simulated experiments in which the network parameters are closely matched to data measured in real cortical networks, and also demonstrate the sampler applied to real calcium imaging data., Comment: Published in at http://dx.doi.org/10.1214/11-AOAS467 the Annals of Applied Statistics (http://www.imstat.org/aoas/) by the Institute of Mathematical Statistics (http://www.imstat.org)

The concept of quasi-bosons or composite bosons (like mesons, excitons etc.) has a wide range of potential physical applications. Even composed of two pure fermions, the quasi-boson creation and annihilation operators satisfy non-standard commutation relations. It is natural to try to realize the quasi-boson operators by the operators of deformed (nonlinear) oscillator, the latter constituting widely studied field of modern quantum physics. In this paper, it is proved that the deformed oscillators which realize quasi-boson operators in a consistent way really exist. The conditions for such realization are derived, and the uniqueness of the family of deformations under consideration is shown., 7 pages; minor changes made, the published version

We present a framework for reconstructing structure of complete neural circuits in the brain using collections of independent measurements of connectivity performed with existing anatomical or functional fluorescent probes, and designed to provide complementary information about neural circuit’s structure by targeting slightly different its parts either in deterministic or stochastic succession. We discuss specific implementation of this procedure using synaptic fluorescent marker GRASP and Cre/Lox system Brainbow to collect ensemble of observations of the sets of synapses between stochastically labeled samples of neurons. By representing such measurements mathematically as weak constraints on circuit’s connectivity matrix and by solving a constrained optimization problem, we are able to exactly deduce the wiring diagram in C. Elegans in an in-silico experiment from only ~10,000 measurements. This offers possibility for routinely reconstructing complete connectivity in smaller organisms, such as C. Elegans, using exclusively light microscopy instruments over the span of single weeks.

We describe an approach for automation of the process of reconstruction of neural tissue from serial section transmission electron micrographs. Such reconstructions require 3D segmentation of individual neuronal processes (axons and dendrites) performed in densely packed neuropil. We first detect neuronal cell profiles in each image in a stack of serial micrographs with multi-scale ridge detector. Short breaks in detected boundaries are interpolated using anisotropic contour completion formulated in fuzzy-logic framework. Detected profiles from adjacent sections are linked together based on cues such as shape similarity and image texture. Thus obtained 3D segmentation is validated by human operators in computer-guided proofreading process. Our approach makes possible reconstructions of neural tissue at final rate of about 5 microm3/manh, as determined primarily by the speed of proofreading. To date we have applied this approach to reconstruct few blocks of neural tissue from different regions of rat brain totaling over 1000microm3, and used these to evaluate reconstruction speed, quality, error rates, and presence of ambiguous locations in neuropil ssTEM imaging data.

We discuss the higher Fock state (q \bar q g) contributions to the nonzero value of the pion GPD at the crossover point x = zeta between the DGLAP and ERBL regions. Using the phenomenological light-front constituent quark model, we confirm that the higher Fock state contributions indeed give a nonzero value of the GPD at the crossover point. Iterating the light-front quark model wave function of the lowest q \bar q Fock state with the Bethe-Salpeter kernel corresponding to the one-gluon-exchange, we include all possible time-ordered q \bar q g Fock state contributions and obtain the pion GPD satisfying necessary sum rules and continuity conditions., References added

We suggest an exact approach to help remedy the fermion sign problem in diffusion quantum Monte Carlo simulations. The approach is based on an explicit suppression of symmetric modes in the Schrödinger equation by means of a modified stochastic diffusion process (antisymmetric diffusion process). We introduce this algorithm and illustrate it on potential models in one dimension (1D) and show that there it solves the fermion sign problem exactly and converges to the lowest antisymmetric state of the system. Then, we discuss extensions of this approach to many-dimensional systems on examples of quantum oscillator in 2D-20D and a toy model of three and four fermions on harmonic strings in 2D and 3D. We show that in all these cases our method shows a performance comparable to that of a fixed-node approximation with an exact node.

We discuss a treacherous point in light-front dynamics (LFD) which should be taken into account to restore complete equivalence with the manifestly covariant formalism. We present examples that require an inclusion of the arc contribution in the light-front energy contour integration in order to achieve the equivalence between the LFD result and the manifestly covariant result. © 2005 The American Physical Society.

MISHCHENKO, YURIY. Applications of Canonical Transformations and Nontrivial Vacuum Solutions to flavor mixing and critical phenomena in Quantum Field Theory. (Under the direction of Chueng-Ryong Ji.) In this dissertation we consider two recent applications of Bogoliubov Transformation to the phenomenology of quantum mixing and the theory of critical phenomena. In recent years quantum mixing got in the focus of the searches for New Physics due to its unparalleled sensitivity to SM parameters and indications of neutrino mixing. It was recently suggested that Bogoliubov Transformation may be important in proper definition of the flavor states that otherwise results in problems in perturbative treatment. As first part of this dissertation we investigate this conjecture and develop a complete formulation of such a mixing field theory involving introduction of general formalism, analysis of space-time conversion and phenomenological implications. As second part of this dissertati

We discuss a novel form of the variational approach in Quantum Field Theory in which the trial quantum configuration is represented directly in terms of relevant expectation values rather than, e.g., increasingly complicated structure from Fock space. The quantum algebra imposes constraints on such expectation values so that the variational problem is formulated here as an optimization under constraints. As an example of application of such approach we consider the study of ground state and critical properties in a variant of nonlinear sigma model., talk presented at DPF2004 meeting in Riverside, CA; to appear in a supplement in International Journal of Modern Physics A

In this short note we discuss recent observation of linear correlation on log-log scale between distribution of dark and visible mass in gravitationally bound systems. The coefficient of such correlation appears to be essentially the same for various systems of dramatically different scales such as spiral galaxies of different luminosities and galaxy clusters. We briefly touch possible interpretations of this observation and implications for the mass of dark matter particle., presented at the DPF2004 meeting in Riverside, CA; to be published in a supplement in International Journal of Modern Physics A

We study the distribution of dark matter versus visible matter using a set of data obtained from strong gravitational lensing in the galaxy cluster CL0024+1654 and another set of data inferred from the universal rotation curves in spiral galaxies. The important feature of these two dramatically different observations is that the mass density profile of both visible and dark components can be estimated. From these measurements we deduce the mass of the dark matter particle and our estimate of the mass for the dark matter particle is $\mu_d \approx (200-800)$MeV. We contrast our estimates from CL0024+1654 data and the universal rotation curves of the spiral galaxies and discuss their consistency., Comment: 14 pages, 13 figures

We present a general theory of mixing for an arbitrary number of fields with integer or half-integer spin. The time dynamics of the interacting fields is solved and the Fock space for interacting fields is explicitly constructed. The unitary inequivalence of the Fock space of base ~unmixed! eigenstates and the physical mixed eigenstates is shown by a straightforward algebraic method for any number of flavors in boson or fermion statistics. The oscillation formulas based on the nonperturbative vacuum are derived in a unified general formulation and then applied to both two- and three-flavor cases. Especially, the mixing of spin-1 ~vector! mesons and the Cabibbo-Kobayashi-Maskawa mixing phenomena in the standard model are discussed, emphasizing the nonperturbative vacuum effect in quantum field theory.

Replacing the perturbative vacuum by the nonperturbative vacuum, we extend a recent development of a quantum field theoretic framework for scalar and pseudoscalar meson mixing. The unitary inequivalence of the Fock space of base (unmixed) eigenstates and the physical mixed eigenstates is investigated and the flavor vacuum state structure is explicitly found. This is exploited to develop formulas for two flavor boson oscillations in systems of arbitrary boson occupation number. We apply these formulas to analyze the mixing of eta with eta' and comment on the other meson-mixing systems. In addition, we consider the mixing of boson coherent states, which may have future applications in the construction of meson lasers., Comment: 26 pages, 5 figures

Wiring economy has successfully explained the individual placement of neurons in simple nervous systems like that of Caenorhabditis elegans [ [1] , [2] and [3] ] and the locations of coarser structures like cortical areas in complex vertebrate brains [4]. However, it remains unclear whether wiring economy can explain the placement of individual neurons in brains larger than that of C. elegans. Indeed, given the greater number of neuronal interconnections in larger brains, simply minimizing the length of connections results in unrealistic configurations, with multiple neurons occupying the same position in space. Avoiding such configurations, or volume exclusion, repels neurons from each other, thus counteracting wiring economy. Here we test whether wiring economy together with volume exclusion can explain the placement of neurons in a module of the Drosophila melanogaster brain known as lamina cartridge [ [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] and [13] ]. We used newly developed techniques for semiautomated reconstruction from serial electron microscopy (EM) [14] to obtain the shapes of neurons, the location of synapses, and the resultant synaptic connectivity. We show that wiring length minimization and volume exclusion together can explain the structure of the lamina microcircuit. Therefore, even in brains larger than that of C. elegans, at least for some circuits, optimization can play an important role in individual neuron placement., We are grateful to Richard Fetter for assistance in EM imaging and Janelia FlyEM Project Team and Aljoshcha Nern for helpful comparisons with images from Gal4-driven expression of green fluorescent protein in lamina cells. M.R.-A. acknowledges a Formación del Personal de Investigación (FPI) fellowship from Ministerio de Ciencia e Innovación (MICINN) and support from the Janelia Farm Research Campus (JFRC) (Howard Hughes Medical Institute [HHMI]) visitor program; G.G.d.P. acknowledges funding by MICINN (Spain) as Plan Nacional, including a FPI fellowship, and as partners of the ERASysBio+ initiative supported under the EU ERA-NET Plus scheme in FP7, and support from the JFRC (HHMI) visitor program. I.A.M. also acknowledges support from the JFRC (HHMI) visitor program.

We analyze phenomenological aspects of the quantum field theoretical formulation of meson mixing and obtain the exact oscillation formula in the presence of the decay. This formula is different from quantum mechanical formula by additional high-frequency oscillation terms. In the infinite volume limit, the space of the flavor quantum states is unitarily inequivalent to the space of energy eigenstates.