Decoding algorithm

Neuroprosthetic brain-computer interfaces are systems that decode neural activity into useful control signals for effectors, such as a cursor on a computer screen. It has long been recognized that both the user and decoding system can adapt to increase the accuracy of the end effector. Co-adaptation is the process whereby a user learns to control the system in conjunction with the decoder adapting to learn the user's neural patterns. We provide a mathematical framework for co-adaptation and relate co-adaptation to the joint optimization of the user's control scheme ("encoding model") and the decoding algorithm's parameters. When the assumptions of that framework are respected, co-adaptation cannot yield better performance than that obtainable by an optimal initial choice of fixed decoder, coupled with optimal user learning. For a specific case, we provide numerical methods to obtain such an optimized decoder. We demonstrate our approach in a model brain-computer interface system using an online prosthesis simulator, a simple human-in-the-loop pyschophysics setup which provides a non-invasive simulation of the BCI setting. These experiments support two claims: that users can learn encoders matched to fixed, optimal decoders and that, once learned, our approach yields expected performance advantages.

Neuroprosthetic brain-computer interfaces are systems that decode neural activity into useful control signals for effectors, such as a cursor on a computer screen. It has long been recognized that both the user and decoding system can adapt to increase the accuracy of the end effector. Co-adaptation is the process whereby a user learns to control the system in conjunction with the decoder adapting to learn the user's neural patterns. We provide a mathematical framework for co-adaptation and relate co-adaptation to the joint optimization of the user's control scheme ("encoding model") and the decoding algorithm's parameters. When the assumptions of that framework are respected, co-adaptation cannot yield better performance than that obtainable by an optimal initial choice of fixed decoder, coupled with optimal user learning. For a specific case, we provide numerical methods to obtain such an optimized decoder. We demonstrate our approach in a model brain-compute...

Arabic is considered to have a rich morphology compared to English language. This fact adversely affects the performance of English-Arabic Statistical Machine Translation (SMT). Phrasebased SMT models have a limitation of mapping phrases or blocks from the source to the target languages without any use of linguistic information. Incorporating linguistic tools, such as part-of-speech (POS) taggers can have an impact on translation quality. In this paper, the use of POS tagging is incorporated as a linguistic feature in a factored translation model. The use of factored translation model and its impact on translation quality for English-Arabic machine translation is reported.

Arabic is considered to have a rich morphology compared to English language. This fact adversely affects the performance of English-Arabic Statistical Machine Translation (SMT). Phrasebased SMT models have a limitation of mapping phrases or blocks from the source to the target languages without any use of linguistic information. Incorporating linguistic tools, such as part-of-speech (POS) taggers can have an impact on translation quality. In this paper, the use of POS tagging is incorporated as a linguistic feature in a factored translation model. The use of factored translation model and its impact on translation quality for English-Arabic machine translation is reported.

A robust method to help identify the population of neurons used for decoding motor tasks is developed. We use sensitivity analysis to develop a new metric for quantifying the relative contribution of a neuron towards the decoded output, called “fractional sensitivity.” Previous model-based approaches for neuron ranking have been shown to largely depend on the collection of training data. We suggest the use of anensembleof models that are trained on random subsets of trials to rank neurons. For this work, we tested a decoding algorithm on neuronal data recorded from two male rhesus monkeys while they performed a reach to grasp a bar at three orientations (45∘,90∘, or135∘). An ensemble approach led to a statistically significant increase of 5% in decoding accuracy and 25% increase in identification accuracy of simulated noisy neurons, when compared to a single model. Furthermore, ranking neurons based on the ensemble fractional sensitivities resulted in decoding accuracies 10%–20% gre...

Electroencephalogram (EEG) signals have great importance in the area of brain-computer interface (BCI) which has diverse applications ranging from medicine to entertainment. BCI acquires brain signals, extracts informative features and generates control signals from the knowledge of these features for functioning of external devices. The objective of this work is twofold. Firstly, to extract suitable features related to hand movements and secondly, to discriminate the left and right hand movements signals finding effective classifier. This work is a continuation of our previous study where beta band was found compatible for hand movement analysis. The discrete wavelet transform (DWT) has been used to separate beta band of the EEG signal in order to extract features.  The performance of a probabilistic neural network (PNN) is investigated to find better classifier of left and right hand movements EEG signals and compared with classical back propagation based neural network. The obtai...

Arabic is considered to have a rich morphology compared to English language. This fact adversely affects the performance of English-Arabic Statistical Machine Translation (SMT). Phrasebased SMT models have a limitation of mapping phrases or blocks from the source to the target languages without any use of linguistic information. Incorporating linguistic tools, such as part-of-speech (POS) taggers can have an impact on translation quality. In this paper, the use of POS tagging is incorporated as a linguistic feature in a factored translation model. The use of factored translation model and its impact on translation quality for English-Arabic machine translation is reported.

Arabic is considered to have a rich morphology compared to English language. This fact adversely affects the performance of English-Arabic Statistical Machine Translation (SMT). Phrasebased SMT models have a limitation of mapping phrases or blocks from the source to the target languages without any use of linguistic information. Incorporating linguistic tools, such as part-of-speech (POS) taggers can have an impact on translation quality. In this paper, the use of POS tagging is incorporated as a linguistic feature in a factored translation model. The use of factored translation model and its impact on translation quality for English-Arabic machine translation is reported.

Electroencephalogram (EEG) signals have great importance in the area of brain-computer interface (BCI) which has diverse applications ranging from medicine to entertainment. BCI acquires brain signals, extracts informative features and generates control signals from the knowledge of these features for functioning of external devices. The objective of this work is twofold. Firstly, to extract suitable features related to hand movements and secondly, to discriminate the left and right hand movements signals finding effective classifier. This work is a continuation of our previous study where beta band was found compatible for hand movement analysis. The discrete wavelet transform (DWT) has been used to separate beta band of the EEG signal in order to extract features.  The performance of a probabilistic neural network (PNN) is investigated to find better classifier of left and right hand movements EEG signals and compared with classical back propagation based neural network. The obtai...

Arabic is considered to have a rich morphology compared to English language. This fact adversely affects the performance of English-Arabic Statistical Machine Translation (SMT). Phrasebased SMT models have a limitation of mapping phrases or blocks from the source to the target languages without any use of linguistic information. Incorporating linguistic tools, such as part-of-speech (POS) taggers can have an impact on translation quality. In this paper, the use of POS tagging is incorporated as a linguistic feature in a factored translation model. The use of factored translation model and its impact on translation quality for English-Arabic machine translation is reported.

Wireless Sensor Networks (WSN) are increasingly being envisioned for the collection of data, such as physical or environmental properties. Unlike detection of an event, tracking requires ensuring continuous monitoring. Resource constrained nature of wireless sensor networks makes energy efficient tracking a challenging task. Prediction based approaches try to save energy by reducing an avoidable communication. A Kalman-based approach has been widely used for target tracking but is inaccurate in the case of maneuvering target due to its inability to incorporate nonlinearity. In this paper, dynamic neural network based approaches called Time Delay Neural Network (TDNN) and Nonlinear Autoregressive network with Exogenous inputs (NARX) are proposed for non-cooperative target tracking application. The performance of NARX is compared with Kalman based approach and TDNN in terms of tracking accuracy. Simulation results show that NARX outperforms both Kalman approach and TDNN for target tra...

Brain-machines capture brain signals in order to restore communication and movement to disabled people who suffer from brain palsy or motor disorders. In brain regions, the ensemble firing of populations of neurons represents spatio-temporal patterns that are transformed into outgoing spatio-temporal patterns which encode complex cognitive task. This transformation is dynamic, non-stationary (time-varying) and highly nonlinear. Hence, modeling such complex biological patterns requires specific model structures to uncover the underlying physiological mechanisms and their influences on system behavior. In this study, a recent multi-electrode technology allows the record of the simultaneous neuron activities in behaving animals. Intra-cortical data are processed according to these steps: spike detection and sorting, than desired action extraction from the rate of the obtained signal. We focus on the following important questions about (i) the possibility of linking the brain signal time events with some time-delayed mapping tools; (ii) the use of some suitable inputs than others for the decoder; (iii) a consideration of separated data or a special representation founded on multi-dimensional statistics. This paper concentrates mostly on the analysis of parallel spike train when certain critical hypotheses are ignored by the data for the working method. We have made efforts to define explicitly whether the underlying hypotheses are actually achieved. In this paper, we propose an algorithm to define the embedded memory order of NARX recurrent neural networks to the hand trajectory tracking process. We also demonstrate that this algorithm can improve performance on inference tasks.

We present an Online Server for Machine Translation Evaluation that offers improvements over the standard usage of the typical scoring scripts. Users are able to interactively define their own test sets, experiments and pre-processing steps. Several scores are automatically calculated for submitted translations and the hypotheses and scores are organized and archived for later review. The server offers a nice web based user interface.