Gerard DREYFUS
ESPCI ParisTech, Electronique, Faculty Member
Research Interests:
Research Interests:
Research Interests:
Research Interests: Computer Science, Artificial Intelligence, Mild Cognitive Impairment, Electroencephalography, Speech Recognition, and 14 moreBlind Source Separation, Neural Network, Wavelet, Time-Frequency Analysis, Feature Selection, Artificial Neural Networks, Linear Discriminant Analysis, Wavelet Transform, Time Frequency Analysis, Sparse representation, Early Detection, Sensitivity and Specificity, Time Frequency Representation, and Blind Signal Separation
Research Interests:
Anosognosia, or the lack of awareness of one’s own impairment, is frequent for memory deficits in patients with Alzheimer’s disease (AD). Although often related to frontal dysfunctions, the neural mechanisms of anosognosia remain largely... more
Anosognosia, or the lack of awareness of one’s own impairment, is frequent for memory deficits in patients with Alzheimer’s disease (AD). Although often related to frontal dysfunctions, the neural mechanisms of anosognosia remain largely unknown. We hypothesized that anosognosia in AD may result from a failure in the error-monitoring system, thus preventing patients from being aware of and learning from their own errors. We therefore investigated the event-related potentials evoked by erroneous responses during a memory task in two groups of amyloid positive individuals who had only subjective memory complaints at study entry: 1) those who progressed to AD; and 2) those who remained cognitively normal after five years of follow-up. Our findings revealed direct evidence of a failure in the error-monitoring system at early stages of AD, suggesting that it may be the critical neural substrate of anosognosia in this neurodegenerative disorder.
Research Interests:
Research Interests:
Research Interests:
Research Interests: Computer Science, Algorithms, Artificial Intelligence, Active Learning, Nonlinear dynamics, and 15 moreNeural Networks, Experimental Design, Neural Network, Medicine, Multidisciplinary, X Rays, Computer Simulation, Optimality, Bootstrap, Discrete random variable, Optimal Design, Monte Carlo Method, Orthogonal Polynomial, Bagging, and Neural Network Model
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests: Materials Engineering, Materials Science, Chemistry, Machine Learning, Neural Networks, and 12 moreNeural Network, Numerical Method, Phase Diagrams, Non crystalline solids, Glass Transition, Mixtures, Regression Model, Parameterization, Polynomial Regression, Database, Non linear regression, and melange
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests: Algorithms, Artificial Intelligence, Machine Learning, Statistical Analysis, Nonlinear dynamics, and 15 moreElectroencephalography, Neural Networks, Multidisciplinary, Information Processing, Brain Mapping, Fourier Analysis, Brain, Humans, Animals, Neurons, Local Field Potential, Olfactory Bulb, Automatic Classification, Elementary Functions, and Choice Behavior
Research Interests:
Research Interests:
... Page 8. Gerard Dreyfus ESPCI, Laboratoire d'Electronique 10 rue Vauquelin 75005 Paris, France E-mail: Gerard. Dreyfus@ espci. fr Library of Congress Control Number: 2005929871 Original French edition published by... more
... Page 8. Gerard Dreyfus ESPCI, Laboratoire d'Electronique 10 rue Vauquelin 75005 Paris, France E-mail: Gerard. Dreyfus@ espci. fr Library of Congress Control Number: 2005929871 Original French edition published by Eyrolles, Paris (1st edn. 2002, 2nd edn. ...
Research Interests:
The electrophysiological data recorded in the glomerular stage of the insect olfactory pathway show both a coherent global oscillating behavior of the neurons of this stage- carrier waveform?-, and a reproducible complex activity pattern-... more
The electrophysiological data recorded in the glomerular stage of the insect olfactory pathway show both a coherent global oscillating behavior of the neurons of this stage- carrier waveform?-, and a reproducible complex activity pattern- code?- of some of these neurons, in phase with the global oscillation. We propose a possible interpretation of this type of biological activity patterns, using a simple model of the glomerular stage of the insect olfactory pathway that has been previously designed. This model is analytically tractable, even when synaptic noise, random synaptic weights, inputs or delays are taken into account. This model exhibits the property of coding its inputs through spatio-temporal patterns which are the attractors of its dynamics. These attractors can be long cycles, robust against synaptic noise and also to input fluctuations, provided that the latter occur within well-defined limits. We give an example of a set of adapted synaptic weights and inputs leading ...
Research Interests:
Research Interests:
Research Interests:
The article describes a video-only speech recognition system for a “silent speech interface ” application, using ultrasound and optical images of the voice organ. A one-hour audio-visual speech corpus was phonetically labeled using an... more
The article describes a video-only speech recognition system for a “silent speech interface ” application, using ultrasound and optical images of the voice organ. A one-hour audio-visual speech corpus was phonetically labeled using an automatic speech alignment procedure and robust visual feature extraction techniques. HMM-based stochastic models were estimated separately on the visual and acoustic corpus. The performance of the visual speech recognition system is compared to a traditional acoustic-based recognizer. Index Terms: speech recognition, audio-visual speech description, silent speech interface, machine learning
Research Interests:
We present an original initialization procedure for the parameters of feedforward wavelet networks, prior to training by gradient-based techniques. It takes advantage of wavelet frames stemming from the discrete wavelet transform, and... more
We present an original initialization procedure for the parameters of feedforward wavelet networks, prior to training by gradient-based techniques. It takes advantage of wavelet frames stemming from the discrete wavelet transform, and uses a selection method to determine a set of best wavelets whose centers and dilation parameters are used as initial values for subsequent training. Results obtained for the modeling of two simulated processes are compared to those obtained with a heuristic initialization procedure, and demonstrate the effectiveness of the proposed method.
Research Interests:
One of the most widespread misconceptions about neural networks is the fact that they are "black boxes " which (i) do not make any use of prior knowledge of the process to be modelled, and (ii) cannot be "understood "... more
One of the most widespread misconceptions about neural networks is the fact that they are "black boxes " which (i) do not make any use of prior knowledge of the process to be modelled, and (ii) cannot be "understood " by the expert of the process. We show that, on the contrary, neural networks can be used as "grey box models", and that the designer can take full advantage of the mathematical knowledge which may exist on the process. Using a
Research Interests:
The fundamental property of feedforward neural networks - parsimonious approximation - makes them excellent candidates for modeling static nonlinear processes from measured data. Similarly, feedback (or recurrent) neural networks have... more
The fundamental property of feedforward neural networks - parsimonious approximation - makes them excellent candidates for modeling static nonlinear processes from measured data. Similarly, feedback (or recurrent) neural networks have very attractive properties for the dynamic nonlinear modeling of artificial or natural processes; however, the design of such networks is more complex than that of feedforward neural nets, because the designer has additional degrees of freedom. In the present paper, we show that this complexity may be greatly reduced by (i) incorporating into the very structure of the network all the available mathematical knowledge about the process to be modeled, and by (ii) transforming the resulting network into a "universal" form, termed canonical form, which further reduces the complexity of analyzing and training dynamic neural models.
Latest results on continuous speech phone recognition from video observations of the tongue and lips are described in the context of an ultrasound-based silent speech interface. The study is based on a new 61-minute audiovisual database... more
Latest results on continuous speech phone recognition from video observations of the tongue and lips are described in the context of an ultrasound-based silent speech interface. The study is based on a new 61-minute audiovisual database containing ultrasound sequences of the tongue as well as both frontal and lateral view of the speaker’s lips. Phonetically balanced and exhibiting good diphone coverage, this database is designed both for recognition and corpus-based synthesis purposes. Acoustic waveforms are phonetically labeled, and visual sequences coded using PCA-based robust feature extraction techniques. Visual and acoustic observations of each phonetic class are modeled by continuous HMMs, allowing the performance of the visual phone recognizer to be compared to a traditional acoustic-based phone recognition experiment. The phone recognition confusion matrix is also discussed in detail.
Research Interests:
The article describes a video-only speech recognition system for a “silent speech interface” application, using ultrasound and optical images of the voice organ. A one-hour audiovisual speech corpus was phonetically labeled using an... more
The article describes a video-only speech recognition system for a “silent speech interface” application, using ultrasound and optical images of the voice organ. A one-hour audiovisual speech corpus was phonetically labeled using an automatic speech alignment procedure and robust visual feature extraction techniques. HMM-based stochastic models were estimated separately on the visual and acoustic corpus. The performance of the visual speech recognition system is compared to a traditional acoustic-based recognizer.
Research Interests:
The paper proposes a general framework that encompasses the train- ing of neural networks and the adaptation of filters. We show that neural networks can be considered as general nonlinear filters that can be trained adaptively, that is,... more
The paper proposes a general framework that encompasses the train- ing of neural networks and the adaptation of filters. We show that neural networks can be considered as general nonlinear filters that can be trained adaptively, that is, that can undergo continual training with a ...
Research Interests:
Research Interests:
The article presents the results of tests of a portable post-laryngectomy voice replacement system that allows a silently articulating speaker to select and play back short phrases contained in a 60-phrase phrasebook. Such a system could... more
The article presents the results of tests of a portable post-laryngectomy voice replacement system that allows a silently articulating speaker to select and play back short phrases contained in a 60-phrase phrasebook. Such a system could be a useful communication tool for post-laryngectomy patients unable to use tracheo-oesophageal speech. Experiments on two non-pathological speakers and one person having undergone a total laryngectomy in 1998 are presented. Results are promising and provide proof of principle for a more sophisticated system currently being developed.
Research Interests:
In their paper [1], Tsoi and Tan present what they call a "canonical form", which they claim to be identical to that proposed in Nerrand et al [2]. They also claim that the algorithm which they present can be applied to any... more
In their paper [1], Tsoi and Tan present what they call a "canonical form", which they claim to be identical to that proposed in Nerrand et al [2]. They also claim that the algorithm which they present can be applied to any recurrent neural network. In the present comment, we disprove both claims. Back in 1993, Nerrand et al. [2] proposed a general approach to the training of recurrent networks, either adaptively (on-line) or non-adaptively (off-line). One of the main points of that paper was the introduction of the minimal state-space form, or canonical form, defined in relations (4) and (4a) of their paper as: z(n+1) = φ [ z(n), u(n)] (state equation) y(n+1) = ψ[ z(n+1), u(n+1)] (output equation) where z(n) is a state vector, i.e. a minimal set of variables necessary at time n for computing the future output vector y(n+1), the external input vector u(n+1) (control inputs, measured disturbances, ...) being known. A graphic representation of the canonical form is shown in ...
Research Interests:
Paper [1] aimed at providing a unified presentation of neural network architectures. We show in the present comment (i) that the canonical form of recurrent neural networks presented by Nerrand et al. [2] many years ago provides the... more
Paper [1] aimed at providing a unified presentation of neural network architectures. We show in the present comment (i) that the canonical form of recurrent neural networks presented by Nerrand et al. [2] many years ago provides the desired unification, (ii) that what Tsoi and Back call Nerrand's canonical form is not the canonical form presented by Nerrand et al. in [2], and that (iii) contrarily to the claim of Tsoi and Back, all neural network architectures presented in their paper can be tranformed into Nerrand's canonical form. We show that the contents of Tsoi and Back's paper obscures the issues involved in the choice of a recurrent neural network instead of clarifying them: this choice is definitely much simpler than it might seem from Tsoi and Back's paper. In [1], Tsoi and Back present a number of different discrete-time recurrent neural network architectures and intend to clarify the links between them. The authors must be commended for trying to perform s...
Research Interests:
The recent developments of statistical learning focused mainly on vector machines, i.e. on machines that learn from examples described by a vector of features. There are many fields where structured data must be handled; therefore, it... more
The recent developments of statistical learning focused mainly on vector machines, i.e. on machines that learn from examples described by a vector of features. There are many fields where structured data must be handled; therefore, it would be desirable to learn from examples described by graphs. The presentation describes graph machines, which learn real numbers from graphs. Applications in the field of Quantitative Structure-Activity Relations (QSAR), which aim at predicting properties of molecules from their (graph) structures, are described.
Research Interests:
The development of a continuous visual speech recognizer for a silent speech interface has been investigated using a visual speech corpus of ultrasound and video images of the tongue and lips. By using high-speed visual data and... more
The development of a continuous visual speech recognizer for a silent speech interface has been investigated using a visual speech corpus of ultrasound and video images of the tongue and lips. By using high-speed visual data and tied-state cross-word triphone HMMs, and including syntactic information via domain-specific language models, word-level recognition accuracy as high as 72% was achieved on visual speech. Using the Julius system, it was also found that the recognition should be possible in nearly real-time.
Recent improvements are presented for phonetic decoding of continuous-speech from ultrasound and optical observations of the tongue and lips in a silent speech interface application. In a new approach to this critical step, the visual... more
Recent improvements are presented for phonetic decoding of continuous-speech from ultrasound and optical observations of the tongue and lips in a silent speech interface application. In a new approach to this critical step, the visual streams are modeled by context-dependent multi-stream Hidden Markov Models (CD-MSHMM). Results are compared to a baseline system using context-independent modeling and a visual feature fusion strategy, with both systems evaluated on a onehour, phonetically balanced English speech database. Tongue and lip images are coded using PCA-based feature extraction techniques. The uttered speech signal, also recorded, is used to initialize the training of the visual HMMs. Visual phonetic decoding performance is evaluated successively with and without the help of linguistic constraints introduced via a 2.5kword decoding dictionary.
Research Interests:
This article presents a framework for a phonetic vocoder driven by ultrasound and optical images of the tongue and lips for a “silent speech interface” application. The system is built around an HMM-based visual phone recognition step... more
This article presents a framework for a phonetic vocoder driven by ultrasound and optical images of the tongue and lips for a “silent speech interface” application. The system is built around an HMM-based visual phone recognition step which provides target phonetic sequences from a continuous visual observation stream. The phonetic target constrains the search for the optimal sequence of diphones that maximizes similarity to the input test data in visual space subject to a unit concatenation cost in the acoustic domain. The final speech waveform is generated using “Harmonic plus Noise Model” synthesis techniques. Experimental results are based on a onehour continuous speech audiovisual database comprising ultrasound images of the tongue and both frontal and lateral view of the speaker’s lips.
Research Interests:
The early detection Alzheimer's disease is an important challenge. Using blind source separation, wavelet time-frequency transforms and "bump modeling" of electro-encephalographic (EEG) recordings, a set of features... more
The early detection Alzheimer's disease is an important challenge. Using blind source separation, wavelet time-frequency transforms and "bump modeling" of electro-encephalographic (EEG) recordings, a set of features describing the recordings of mildly impaired patients and of controls subject is built. Feature selection by the random probe method leads to the selection of a few reliable features, which are fed to a neural network classifier. This leads to a sizeable performance improvement over detection results previously published on the same data.
Research Interests:
We describe a new algorithm for the estimation of Cycle Lengths (CL) in the atria. In the spirit of wavelet transforms, the algorithm correlates the electrogram (EGM) signal to a set of functions that are specifically designed to extract... more
We describe a new algorithm for the estimation of Cycle Lengths (CL) in the atria. In the spirit of wavelet transforms, the algorithm correlates the electrogram (EGM) signal to a set of functions that are specifically designed to extract the cycle length present in the ...