Research

Byzantine Fault Tolerant Federated Learning for Machine-to-Machine Communication: A Smart City Vehicular Network

Abstract: A common conception of smart vehicular and transportation systems is to create a fault-tolerant pair-to-pair network among the vehicles in the communication range such that they can share information about traffic, road conditions, and many more. Although there are several hurdles to overcome, recently most vehicles are equipped with LIDAR, ultrasonic, and camera sensors for the Advance Driving Assistant System (ADAS). Two major problems in developing a machine-to-machine network with these sensors are the sheer amount of data (gigabytes per second) that needs to be sent over the communication channel and some of the nodes may not respond or can provide inaccurate Data. However, the end goal of this data is to train a certain ML or DL model for prediction and Automation. Thus using federated learning techniques in each vehicle's onboard computation unit, the main model can be trained, and each of the model parameters (A few KB) can be sent over the network instead of the data. Then this newly created model information can be validated using a specific consensus algorithm to remove the BFT problem. Central Cloud storage might also be helpful to accommodate the overall ML/DL model version history such that if any of the models proved to be falsified by misleading training data, a quick rollback can be possible

Taking All the Factors We Need: A Multimodal Depression Classification with Uncertainty Approximation

Depression and Anxiety are surprisingly common but generally overlooked mental illness, where more than 5% of total population is thought to suffer from the disease. It's also the major factor behind sucide which is fourth Common reason for death among youngsters. Although there has been several impactful research in these fields, most of the research encompasses one or two factors for detection purposes, whereas in a real world scenario, most patients go through rigorous steps of diagnosis systems. Another crucial insight about using ML/DL for any detection is that it might falsely give classification reports, which might be fatal for depression patients. To mitigate these issues, at first we have to take all possible symptoms related to depression, and create a multimodal diagnosis system that might take any number of factors for any particular patient. These factors might include visual information, behavioral patterns, voice pitch, hyper or hypo activity and medical data like heart rate, breathing rate, screen temperature etc. If any number of factors can be addressed within the same learning model it might also be helpful for data collection and further development. The uncertainty approximation will provide important information about the ML/DL based diagnosis processes, where results from each steps may be further evaluated.

Microchain: A Dicentralized Ledger Technology with distributed hash tree and proof of consistency

Each node keeps a record of all the events (transactions + votes) it has received It assigns a trust score to other nodes based on this history and the transitive gossip received

Machine Learning Classification Assumption by Adaptive Threshold and Multilabel

Hierarchical Neural Transfer Learning and Object Classification

Polycystic Ovary Syndrome: A case study for cause and cure analyses in modern lifestyle

Polycystic ovary syndrome is a common but hardly discussed disease in Bangladesh among womenfolk. From existing research, numerous literature reviews, and practical datasets of patients who are suffering from PCOS - we have found that particular medicines can cure this disease. But there are several parameters like lack of balanced diet, obesity, irregular sexual intercourse, etc. which can affect the causes of PCOS. Using data science, we will find the correlation of these parameters how they are interconnected with each other, and exactly to which extent this result matches the medical analysis.