Indoor thermal comfort immensely impacts the health and performance of occupants. Therefore, researchers and engineers have proposed numerous computational models to estimate thermal comfort (TC). Given the impetus toward energy efficiency, the current focus is on data-driven TC prediction solutions that leverage state-of-the-art machine learning (ML) algorithms. However, an occupant’s perception of indoor thermal comfort (TC) is subjective and multi-dimensional. Different aspects of TC are represented by various standard metrics/scales viz., thermal sensation (TSV), thermal comfort (TCV), and thermal preference (TPV). The current ML-based TC prediction solutions adopt the Single-task Learning approach, i.e., one prediction model per metric. Consequently, solutions often focus on only one TC metric. Moreover, when several metrics are considered, multiple ML models for a single indoor space lead to conflicting predictions, rendering real-world deployment infeasible. This work addresses these problems by leveraging Multi-task Learning for TC prediction in naturally ventilated buildings. First, a survey-and-measurement study is conducted in the composite climatic region of north India, in 14 naturally ventilated classrooms of 5 schools, involving 512 primary school students. Next, the dataset is analyzed for important environmental, physiological, and psycho-social factors that influence thermal comfort of children. Further, “DeepComfort”, a deep neural network based Multi-task Learning model is proposed. DeepComfort predicts multiple TC output metrics viz., TSV, TPV, and TCV, simultaneously through a single model. It is validated on ASHRAE-II database and the primary student dataset created in this study. It demonstrates high F1-scores, Accuracy (≈90%), and generalization capability, despite the challenges of illogical responses and data imbalance. DeepComfort is also shown to outperform 6 popular metric-specific single-task machine learning algorithms.

Classroom thermal comfort has a direct effect on student health and educational outcomes. However, measuring thermal comfort (TC) is a non-trivial task. It is represented by several subjective metrics e.g., Thermal Sensation Vote, Thermal Comfort Vote, Thermal Preference Vote, etc. Since machine learning (ML) is being increasingly used to predict occupant comfort, multiple TC metrics for the same indoor space may yield contradictory results. This poses the challenge of selecting the most suitable single TC metric or the minimal TC metric combination for a given indoor space. Ideally, it will be a metric that can be used to predict all other TC metrics and occupant behavior with high accuracy. This work addresses this problem by using a primary student thermal comfort dataset gathered from 11 schools and over 500 unique students. A comprehensive evaluation is carried out through hundreds of TC prediction models using several ML algorithms. It evaluates the ability of TC metrics to predict (a) other TC metrics, and (b) the adaptive behavior of primary students. An algorithm is proposed to select the most suitable single TC metric or the minimal TC metric input combination. Results show that ML models can accurately predict all TC metrics and occupant-adaptive behavior using a small subset of TC metrics with an average accuracy as high as 79%. This work also found Thermal Sensation Vote to be the most significant single TC predictor, followed by Thermal Satisfaction Level. Interestingly, satisfaction with clothing was found to be as equally relevant as thermal preference. Furthermore, the impact of seasons and choice of ML algorithms on TC metric and occupant behavior prediction is shown.

Metaverse technologies are transforming how distant individuals interact in immersive virtual environments. These technologies, in combination with the latest developments in 3D sensing, will create a hybrid metaverse blending virtual and physical spaces. However, high-resolution 3D sensing poses privacy risks from unrestricted spatial data sharing. It may inadvertently expose the human visual appearance and private objects and spaces in the users’ physical environments to the shared metaverse. To mitigate these challenges, this article surveys the visual privacy issues in the era of the metaverse. It then introduces a visual privacy control method enabling privacy protection in the current metaverse and beyond. Our goal is to empower users to control the visual privacy of spatial objects by comprehending their inherent semantics. The design allows users to define their privacy protection levels or, if preferred, delegate this responsibility to an automated control system. We present two use-case scenarios to demonstrate this concept.