As the number of vehicles increases, road accidents are on the rise every day. According to the World Health Organization (WHO) survey, 1.4 million people have died, and 50 million people have been injured worldwide every year. The key cause of death is the unavailability of medical care at the accident site or the high response time in the rescue operation. A cognitive agent-based collision detection smart accident alert and rescue system will help us to minimize delays in a rescue operation that could save many lives. With the growing popularity of smart cities, intelligent transportation systems (ITS) are drawing major interest in academia and business, and are considered as a means to improve road safety in smart cities. This article proposed an intelligent accident detection and rescue system which mimics the cognitive functions of the human mind using the Internet of Things (IoTs) and the Artificial Intelligence system (AI). An IoT kit is developed that detects the accident and collects all accident-related information, such as position, pressure, gravitational force, speed, etc., and sends it to the cloud. In the cloud, once the accident is detected, a deep learning (DL) model is used to validate the output of the IoT module and activate the rescue module. Once the accident is detected by the DL module, all the closest emergency services such as the hospital, police station, mechanics, etc., are notified. Ensemble transfer learning with dynamic weights is used to minimize the false detection rate. Due to the dataset’s unavailability, a personalized dataset is generated from the various videos available on the Internet. The proposed method is validated by a comparative analysis of ResNet and InceptionResnetV2. The experiment results show that InceptionResnetV2 provides a better performance compared to ResNet with training, validation, and a test accuracy of 98%, respectively. To measure the performance of the proposed approach in the real world, it is validated on the toy car.

In the modern era, posting sarcastic comments on social media became the common trend. Sarcasm is often used by people to taunt or pester others. It is frequently expressed through inflexion, tonal stress in speech or in the form of lexical, pragmatic, and hyperbolic features present in the text. Most of the existing work has been focused on either detecting sarcasm in textual data using text features or audio data using audio features. This article proposed a novel approach by combining textual and audio features together to detecting sarcasm in conversational data. This hybrid method takes a combined vector of extracted audio and text features from their respective models as the input. This combined features will compensated the shortcomings of only text features and vice‐versa. The obtained result of hybrid model outperforms both the individual model significantly.