Mehedi Hasan

Mehedi Hasan, Moloud Abdar, Abbas Khosravi et al.

Although neural networks (especially deep neural networks) have achieved \textit{better-than-human} performance in many fields, their real-world deployment is still questionable due to the lack of awareness about the limitation in their knowledge. To incorporate such awareness in the machine learning model, prediction with reject option (also known as selective classification or classification with abstention) has been proposed in literature. In this paper, we present a systematic review of the prediction with the reject option in the context of various neural networks. To the best of our knowledge, this is the first study focusing on this aspect of neural networks. Moreover, we discuss different novel loss functions related to the reject option and post-training processing (if any) of network output for generating suitable measurements for knowledge awareness of the model. Finally, we address the application of the rejection option in reducing the prediction time for the real-time problems and present a comprehensive summary of the techniques related to the reject option in the context of extensive variety of neural networks. Our code is available on GitHub: \url{https://github.com/MehediHasanTutul/Reject_option}

Mehedi Hasan, Abbas Khosravi, Ibrahim Hossain et al.

Uncertainty quantification in a neural network is one of the most discussed topics for safety-critical applications. Though Neural Networks (NNs) have achieved state-of-the-art performance for many applications, they still provide unreliable point predictions, which lack information about uncertainty estimates. Among various methods to enable neural networks to estimate uncertainty, Monte Carlo (MC) dropout has gained much popularity in a short period due to its simplicity. In this study, we present a new version of the traditional dropout layer where we are able to fix the number of dropout configurations. As such, each layer can take and apply the new dropout layer in the MC method to quantify the uncertainty associated with NN predictions. We conduct experiments on both toy and realistic datasets and compare the results with the MC method using the traditional dropout layer. Performance analysis utilizing uncertainty evaluation metrics corroborates that our dropout layer offers better performance in most cases.

Md. Motahar Mahtab, Monirul Haque, Mehedi Hasan et al.

Intentionally luring readers to click on a particular content by exploiting their curiosity defines a title as clickbait. Although several studies focused on detecting clickbait titles in English articles, low resource language like Bangla has not been given adequate attention. To tackle clickbait titles in Bangla, we have constructed the first Bangla clickbait detection dataset containing 15,056 labeled news articles and 65,406 unlabelled news articles extracted from clickbait dense news sites. Each article has been labeled by three expert linguists and includes an article's title, body, and other metadata. By incorporating labeled and unlabelled data, we finetune a pretrained Bangla transformer model in an adversarial fashion using Semi Supervised Generative Adversarial Networks (SS GANs). The proposed model acts as a good baseline for this dataset, outperforming traditional neural network models (LSTM, GRU, CNN) and linguistic feature based models. We expect that this dataset and the detailed analysis and comparison of these clickbait detection models will provide a fundamental basis for future research into detecting clickbait titles in Bengali articles. We have released the corresponding code and dataset.

Ahnaf Mozib Samin, M. Humayon Kobir, Md. Mushtaq Shahriyar Rafee et al.

Despite huge improvements in automatic speech recognition (ASR) employing neural networks, ASR systems still suffer from a lack of robustness and generalizability issues due to domain shifting. This is mainly because principal corpus design criteria are often not identified and examined adequately while compiling ASR datasets. In this study, we investigate the robustness of the state-of-the-art transfer learning approaches such as self-supervised wav2vec 2.0 and weakly supervised Whisper as well as fully supervised convolutional neural networks (CNNs) for multi-domain ASR. We also demonstrate the significance of domain selection while building a corpus by assessing these models on a novel multi-domain Bangladeshi Bangla ASR evaluation benchmark - BanSpeech, which contains approximately 6.52 hours of human-annotated speech and 8085 utterances from 13 distinct domains. SUBAK.KO, a mostly read speech corpus for the morphologically rich language Bangla, has been used to train the ASR systems. Experimental evaluation reveals that self-supervised cross-lingual pre-training is the best strategy compared to weak supervision and full supervision to tackle the multi-domain ASR task. Moreover, the ASR models trained on SUBAK.KO face difficulty recognizing speech from domains with mostly spontaneous speech. The BanSpeech will be publicly available to meet the need for a challenging evaluation benchmark for Bangla ASR.

Mehedi Hasan, Mohammad Jahid Ibna Basher, Md. Tanvir Rouf Shawon

Intent classification is a fundamental task in natural language understanding, aiming to categorize user queries or sentences into predefined classes to understand user intent. The most challenging aspect of this particular task lies in effectively incorporating all possible classes of intent into a dataset while ensuring adequate linguistic variation. Plenty of research has been conducted in the related domains in rich-resource languages like English. In this study, we introduce BNIntent30, a comprehensive Bengali intent classification dataset containing 30 intent classes. The dataset is excerpted and translated from the CLINIC150 dataset containing a diverse range of user intents categorized over 150 classes. Furthermore, we propose a novel approach for Bengali intent classification using Generative Adversarial BERT to evaluate the proposed dataset, which we call GAN-BnBERT. Our approach leverages the power of BERT-based contextual embeddings to capture salient linguistic features and contextual information from the text data, while the generative adversarial network (GAN) component complements the model's ability to learn diverse representations of existing intent classes through generative modeling. Our experimental results demonstrate that the GAN-BnBERT model achieves superior performance on the newly introduced BNIntent30 dataset, surpassing the existing Bi-LSTM and the stand-alone BERT-based classification model.

Ayaan Haque, Ipsita Sutradhar, Mahziba Rahman et al.

Diabetes is one of the most prevalent chronic diseases in Bangladesh, and as a result, Diabetic Retinopathy (DR) is widespread in the population. DR, an eye illness caused by diabetes, can lead to blindness if it is not identified and treated in its early stages. Unfortunately, diagnosis of DR requires medically trained professionals, but Bangladesh has limited specialists in comparison to its population. Moreover, the screening process is often expensive, prohibiting many from receiving timely and proper diagnosis. To address the problem, we introduce a deep learning algorithm which screens for different stages of DR. We use a state-of-the-art CNN architecture to diagnose patients based on retinal fundus imagery. This paper is an experimental evaluation of the algorithm we developed for DR diagnosis and screening specifically for Bangladeshi patients. We perform this validation study using separate pools of retinal image data of real patients from a hospital and field studies in Bangladesh. Our results show that the algorithm is effective at screening Bangladeshi eyes even when trained on a public dataset which is out of domain, and can accurately determine the stage of DR as well, achieving an overall accuracy of 92.27\% and 93.02\% on two validation sets of Bangladeshi eyes. The results confirm the ability of the algorithm to be used in real clinical settings and applications due to its high accuracy and classwise metrics. Our algorithm is implemented in the application Drishti, which is used to screen for DR in patients living in rural areas in Bangladesh, where access to professional screening is limited.