Papers by Md. Saiful Islam
arXiv (Cornell University), Mar 15, 2022
In this paper, we investigated whether we can 1) detect participants with ataxia-specific gait ch... more In this paper, we investigated whether we can 1) detect participants with ataxia-specific gait characteristics (risk-prediction), and 2) assess severity of ataxia from gait (severityassessment). We collected 155 videos from 89 participants, 24 controls and 65 diagnosed with (or are pre-manifest) spinocerebellar ataxias (SCAs), performing the gait task of the Scale for the Assessment and Rating of Ataxia (SARA) from 11 medical sites located in 8 different states in the United States. We developed a method to separate the participants from their surroundings and constructed several features to capture gait characteristics like step width, step length, swing, stability, speed, etc. Our risk-prediction model achieves 83.06% accuracy and an 80.23% F1 score. Similarly, our severity-assessment model achieves a mean absolute error (MAE) score of 0.6225 and a Pearson's correlation coefficient score of 0.7268. Our models still performed competitively when evaluated on data from sites not used during training. Furthermore, through feature importance analysis, we found that our models associate wider steps, decreased walking speed, and increased instability with greater ataxia severity, which is consistent with previously established clinical knowledge. Our models create possibilities for remote ataxia assessment in non-clinical settings in the future, which could significantly improve accessibility of ataxia care. Furthermore, our * Equal contribution † Equal contribution
In this paper, we introduce “Embedding Barrier”, a phenomenon that limits the monolingual perform... more In this paper, we introduce “Embedding Barrier”, a phenomenon that limits the monolingual performance of multilingual models on low-resource languages having unique typologies. We build `BanglaBERT', a Bangla language model pretrained on 18.6 GB Internet-crawled data and benchmark on five standard NLU tasks. We discover a significant drop in the performance of the state-of-the-art multilingual model (XLM-R) from BanglaBERT and attribute this to the Embedding Barrier through comprehensive experiments. We identify that a multilingual model's performance on a low-resource language is hurt when its writing script is not similar to any of the high-resource languages. To tackle the barrier, we propose a straightforward solution by transcribing languages to a common script, which can effectively improve the performance of a multilingual model for the Bangla language. As a bi-product of the standard NLU benchmarks, we introduce a new downstream dataset on natural language inference ...
International Journal of Applied Research on Information Technology and Computing, 2015
Speech feature extraction is the mathematical representation of the speech file, which converts t... more Speech feature extraction is the mathematical representation of the speech file, which converts the speech waveform to some type of parametric representation for further analysis and processing in speech recognition. A good feature may produce a good result for any recognition system. This paper presents a simple and novel feature extraction approach for extracting binary features of Bangla speech words. This technique is based on frequency-domain signal features and dynamic thresholding method. First the frequency-domain signal feature, i.e., spectrogram feature is computed from the original speech words and then the binary features of these speech words are computed by using the dynamic thresholding technique. The developed system has been justified with several Bangla speech words. To test and analyse theses binary features, a speech recognition module has been developed. The speech recognition is done by using the neural network with back-propagation training algorithm. All the algorithms used in this research are implemented in MATLAB and the implemented speech recognition system achieved recognition accuracy of 96%.
Uploads
Papers by Md. Saiful Islam