This paper aims to estimate indoor occupancy given the real-time observed signals from the existing sensors; next, as a practical application, we build a dynamic control schedule for energy saving based on the estimated indoor occupancy. However, several issues need to be addressed. First, it is impossible to train the model with rich labels due to the expensive labeling cost. Second, manual annotation of the continuous occupancy rate is complex. Third, the mapping relationship between sensor data and occupancy will change in the long run. In this paper, we proposed a new algorithm named Self-Supervised Indoor Occupancy Estimation (SSIOE) to overcome the challenges. Specifically, our training scheme aims to (I) generate a set of pseudo labels in a simple way to mark the time periods believed to be either in a high or low occupancy state and (II) utilize these sparse labels for training a network to infer the continuous occupancy ratio. By reformulating the problem as a Wasserstein-distance-like estimation, SSIOE is a novel learning-based method that can rely only on the weak/sparse labels of either “high-occupancy” or “low-occupancy” and learn to estimate the continuous occupancy ratio. Furthermore, to deal with the scarce annotation problem, we proposed a novel physical constraint loss to model the physical prior. (III) Last but not least, to strengthen the adaptability, we integrated Model-Agnostic Meta-Learning (MAML) to train the model for dynamic model updating. Experimental results show that SSIOE can provide reliable occupancy estimation and flexibly adapt to various control modes without retraining the model.Note to Practitioners—This paper is motivated by the real-time occupancy estimation problem, which is crucial for building management systems to arrange lighting and air conditioning. A dynamic schedule adapted to the occupancy would help balance the user’s comfort and energy saving; however, installing the occupancy sensors raises additional costs. Therefore, this paper proposes SSIOE, a label-free machine learning algorithm that can learn to estimate the continuous occupancy ratio given the sensor status in real time. Prior works trained upon supervised machine learning methods suffer from heavy manual annotation workload, while the unsupervised machine learning approaches fail to provide continuous occupancy estimation values. This paper addresses the above challenges by proposing a novel self-supervised training scheme. Preliminary experiments suggest that this approach is feasible to work in real-world buildings. However, it should be noted that the occupancy estimated by SSIOE is a ratio (0.0-1.0), which is relative to the maximum and minimum occupancy of the training data (or fine-tuning data). For example, if the maximum number of occupants in the training data is 200 and the minimum number of occupants is 50, then the estimated occupancy value of 1.0 is about 200 occupants, and the occupancy value of 0.0 is about 50 occupants. When the maximum and minimum occupants of the space change, the model needs to be updated to capture the new mapping of the estimated occupancy ratio to the number of occupants. In future research, we plan to further improve and simplify the adaptation ability of SSIOE by continuously and automatically updating the system with the latest collected as the system runs.

This paper proposes a novel task-consistency learning method that enables us to train a vacant space detection network (target task) based on the logic consistency with the semantic outcomes from a flow-based motion behavior classifier (source task) in a parking lot. Note that the source task can introduce false detection during task-consistency learning, which implies noisy rewards or supervision. The target network can be trained in a reinforcement learning setting by appropriately designing the reward mechanism upon semantic consistency. We also introduce a novel symmetric constraint to detect corrupted samples and reduce the effect of noisy rewards. Unlike conventional corrupted learning methods that use only training losses to identify corrupted samples, our symmetric constraint also explores the relationship among training samples to improve performance. Compared with conventional supervised detection methods, the main contribution of our work is the ability to learn a vacant space detector via semantic consistency rather than supervised labels. The dynamic learning property allows the proposed detector to be easily deployed and updated in various lots without heavy human loads. Experiments demonstrate that our noisy task consistency mechanism can be successfully applied to train a vacant space detector from scratch.

Most of the existing algorithms for zero-shot classification problems typically rely on the attribute-based semantic relations among categories to realize the classification of novel categories without observing any of their instances. However, training the zero-shot classification models still requires attribute labeling for each class (or even instance) in the training dataset, which is also expensive. To this end, in this paper, we bring up a new problem scenario: ''Can we derive zero-shot learning for novel attribute detectors/classifiers and use them to automatically annotate the dataset for labeling efficiency?'' Basically, given only a small set of detectors that are learned to recognize some manually annotated attributes (i.e., the seen attributes), we aim to synthesize the detectors of novel attributes in a zero-shot learning manner. Our proposed method, Zero-Shot Learning for Attributes (ZSLA), which is the first of its kind to the best of our knowledge, tackles this new research problem by applying the set operations to first decompose the seen attributes into their basic attributes and then recombine these basic attributes into the novel ones. Extensive experiments are conducted to verify the capacity of our synthesized detectors for accurately capturing the semantics of the novel attributes and show their superior performance in terms of detection and localization compared to other baseline approaches. Moreover, we demonstrate the application of automatic annotation using our synthesized detectors on Caltech-UCSD Birds-200-2011 dataset. Various generalized zero-shot classification algorithms trained upon the dataset re-annotated by ZSLA shows comparable performance with those trained with the manual ground-truth annotations.

In this paper, we proposed a novel task-consistency learning method that allows training a vacant space detection network (target task) based on the logic consistency with the semantic outcomes from a flow-based motion behavior classifier (source task) in a parking lot. By well designing the reward mechanism upon semantic consistency, we show the possibility to train the target network in a reinforcement learning setting. Compared with conventional supervised detection methods, this work's main contribution is to learn a vacant space detector via semantic consistency rather than supervised labels. The dynamic learning property may make the proposed detector been deployed and updated in different lots easily without heavy human loads. The experiments show that based on the task consistency rewards from the motion behavior classifier, the vacant space detector can be trained successfully.

In this paper, we introduce an online-learning method to model the property of an office building. Unlike conventional control methods where the building property is modeled via a simulator or through offline learning, our building model is adaptively updated according to the dynamic response of a real environment. Upon the building model for environment prediction, the proposed action agent can control the heating, ventilation, and air conditioning (HVAC) system in a smarter way by scheduling the temperature reference point. To online learn the model and improve the agent, two practical and seldom discussed issues are addressed. The first challenge is data bias where the collected initial training dataset can only partially reveal the statistical mapping between the control input and the environment response. Hence, the trained model may lack generalization. To overcome the data bias issue, a data augmentation method is proposed to embed physical logic in order to train a proper initial model. Next, an online learning process is introduced to update the model generality during the system operation phase. The second practical issue is the constraints on agent exploration for discovering unknown data samples. During the business hours, to comfort employees, a control agent is not allowed to explore the possible controlling space randomly. To balance data collection and control stability, we introduce a hybrid control strategy that considers both the human control rule and the agent action. A confidence score of the agent model is also automatically estimated to determine a suitable control strategy finally. Our experiments have realized in an office building. The results outperform conventional methods and show its superior in terms of control stability.