INTRODUCTION: In today's modern world of networking and intelligent devices, there is an imperative need to upgrade everyday things and make them intelligent; additionally, this is not the era to unquestioningly trust old and traditional security measures, particularly when it comes to smart door lock systems. Mostly, every smart door lock system has a security access code or fingerprint access outside the door that makes it vulnerable. The password in a classic security system can be readily hacked with advanced technology and, therefore, is no longer suitable for today's real-time environment. OBJECTIVES: As a result, this paper intends to provide enhanced security for the user through facial recognition using a machine-learning approach with high accuracy and remote access via an Android application. Automated solutions leveraging machine learning have shown to be quite effective in security. METHODS: Machine Learning (ML) algorithms are used to train different sets of images to identify and classify various sorts of faces. The specific algorithm that is to be used is Dlib and Support Vector Machine (SVM); Dlib is utilized for face recognition along with HOG (Histogram of Oriented Gradients), whereas SVM is used for image classification, which is used to authorize the personnel. RESULTS: When compared to other cutting-edge methodologies, empirical results reveal that the proposed approach achieves 96% accuracy rate, with a recognition speed of 0.5 seconds per face in facia recognition which is more effective, reliable, and utilizes fewer resources. CONCLUSION: Real-time face recognition enables quick and secure identification, supporting multi-factor authentication and monitoring. The proposed smart lock system uses the HOG+SVM approach, achieving higher accuracy and faster face detection.
Reinforcement learning (RL) has emerged as a promising approach to upgrade recommender systems, particularly in dynamic and evolving user environments. Conventional approaches such as collaborative filtering and matrix factorization are hampered by static modeling, data sparsity, and short-term thinking. RL, however, provides an approach that allows systems to learn and evolve in response to user interactions and optimize for immediate and long-term user engagement. This review examines the use of RL in recommender systems through the examination of over 60 works of literature spanning three main approaches: model-free, model-based, and hybrid approaches. This work also highlights several gaps in previous reviews, namely their lack of coverage of deep RL techniques and lack of systematic taxonomies for each of the three approaches. An important discovery in this area is that traditional metrics for offline evaluation such as Precision@K and NDCG cannot account for the long-term performance of policies, showing an inherent disconnect between offline evaluation and actual implementation. Multi-agent reinforcement learning (MARL) and offline RL are two of the most significant emergent areas of study, with the former allowing for complex multi-stage recommendation processes and the latter ensuring data efficiency without needing live interaction.