My motivation to study machine perception is my long-term passion for mathematics, programming, and the human brain. Before knowing much about machine perception, my main motivation was to participate and securing the first prize in Bangladesh Mathematical Olympiad for three times. Following this achievement, I graduated in Electrical Engineering taking major classes on theoretical signal processing and accomplishing a bachelor thesis on the modeling of brain hemodynamics. This made me enthusiastic for data-driven science. Subsequently, having had data analysis experience at MathWorks, University of Cambridge and The Johns Hopkins University (JHU), I am pursuing graduate research on sensory perception and machine learning at Baylor College
…show more content…
However, proprioception and touch have traditionally been studied as two separate domains. Therefore, how proprioception and cutaneous sensing over multiple fingers interact is poorly understood. In this study, we investigated how tactile information is processed over fingers and this processing depends on proprioceptive information indexed as the relative positions of the fingers. Using human psychophysics, we show that the vibrotactile frequencies simultaneously presented on fingers are integrated even when one cue is ignored. Importantly, the distance between the attended (target) and ignored (distractor) fingers determines the strength of this integration - Fingers spread further apart interact less than fingers held closely …show more content…
The data were collected from a patient at Methodist Hospital, when the patient was hearing to random sentences of human speech. Spatio-temporal signals from 70 electrodes placed on the auditory cortex were used as the predictors (total 420 features; 70 nodes x 6 frequency bands at each node). The speech signal was converted to 32 response variables using Fourier transformation, which turned the problem into a multivariate regression. We splitted the data into training, validation, query, and testing sets. The models were trained using the training sets, and the parameters were tuned with validation set. The query set was used for model selection and ensemble learning with model
However, proprioception and touch have traditionally been studied as two separate domains. Therefore, how proprioception and cutaneous sensing over multiple fingers interact is poorly understood. In this study, we investigated how tactile information is processed over fingers and this processing depends on proprioceptive information indexed as the relative positions of the fingers. Using human psychophysics, we show that the vibrotactile frequencies simultaneously presented on fingers are integrated even when one cue is ignored. Importantly, the distance between the attended (target) and ignored (distractor) fingers determines the strength of this integration - Fingers spread further apart interact less than fingers held closely …show more content…
The data were collected from a patient at Methodist Hospital, when the patient was hearing to random sentences of human speech. Spatio-temporal signals from 70 electrodes placed on the auditory cortex were used as the predictors (total 420 features; 70 nodes x 6 frequency bands at each node). The speech signal was converted to 32 response variables using Fourier transformation, which turned the problem into a multivariate regression. We splitted the data into training, validation, query, and testing sets. The models were trained using the training sets, and the parameters were tuned with validation set. The query set was used for model selection and ensemble learning with model