Search

Health-related social control refers to intentional attempts to influence people's health behaviors, often seen in personal relationships. Social robots hold promise in influencing people's health by exerting health-related social control, but it is unclear which social control strategies used by robots are appropriate and potentially effective. This study investigates the effects of positive versus negative, and relationship-oriented versus target-oriented social control strategies from a social robot on people's psychological reactions. In an online video prototype study, participants viewed scenarios of a social robot attempting to change their sedentary behaviors by using different strategies. We found that positive (versus negative) strategies elicited stronger positive affect, enjoyment, and perceived social appropriateness, reduced perceived threats to freedom, and strengthened behavioral intention. Meanwhile, the relationship-oriented (versus target-oriented) strategies elevated people's negative affect, reduced enjoyment and perceived appropriateness, elevated perceived threats to freedom, and weakened behavioral intentions. Given these findings, we give recommendations for designing health influence strategies in social robots.

Yunpeng Feng,1 Shifeng Qiao,1 Jie Chen,1 Xin Wen,2 Yanlei Chen,3 Xiaoyu Song,1 Jiaxin Xu,1 Xiucheng Qiao,1 Jing Yang,2 Shenshen Zhang,1 Yang Feng,1 Yu Gao4 1The Second Ward of Colorectal Surgery, The First Affiliated Hospital of Jinzhou Medical University, Jinzhou, Liaoning, 12100, People’s Republic of China; 2Department of Pathology, The First Affiliated Hospital of Jinzhou Medical University, Jinzhou, Liaoning, 12100, People’s Republic of China; 3The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, People’s Republic of China; 4School of Basic Medical Sciences, Jinzhou Medical University, Jinzhou, Liaoning, 12100, People’s Republic of ChinaCorrespondence: Yu Gao, School of Basic Medical Sciences, Jinzhou Medical University, No. 40, Section 3 Songpo Road, Jinzhou, Liaoning, 121000, People’s Republic of China, Tel +86 15904161717, Email jzykdxgy@163.comPurpose: This research explored the association between CD163-labeled M2-type macrophages and cancer-associated fibroblasts (CAFs) in the tumor microenvironment (TME) of 38 colorectal cancer (CRC) liver metastases. In addition, we investigated the correlation differences between M2-type macrophages and CAFs in the tumor microenvironments of 38 primary colorectal cancer patients with confirmed liver metastases and 946 colorectal cancer patients, as well as possible mechanisms of action between the two cells.Methods: The Immunohistochemistry (IHC) method was applied to detect the expression levels of M2-type macrophages and CAFs in the tissues of 984 cases of CRC and to analyze the correlation between M2-type macrophages and CAFs in colorectal cancer tissues. The IHC method was also applied to detect the expression levels of M2-type macrophages and CAFs in the liver metastases of 38 cases of CRC in the experimental group and to analyze the correlation between the two cells in liver metastases.Results: 1. M2-type macrophages and CAFs expression were significantly higher in 38 primary co

Inverse molecular design with diffusion models holds great potential for advancements in material and drug discovery. Despite success in unconditional molecule generation, integrating multiple properties such as synthetic score and gas permeability as condition constraints into diffusion models remains unexplored. We present the Graph Diffusion Transformer (Graph DiT) for multi-conditional molecular generation. Graph DiT has a condition encoder to learn the representation of numerical and categorical properties and utilizes a Transformer-based graph denoiser to achieve molecular graph denoising under conditions. Unlike previous graph diffusion models that add noise separately on the atoms and bonds in the forward diffusion process, we propose a graph-dependent noise model for training Graph DiT, designed to accurately estimate graph-related noise in molecules. We extensively validate the Graph DiT for multi-conditional polymer and small molecule generation. Results demonstrate our superiority across metrics from distribution learning to condition control for molecular properties. A polymer inverse design task for gas separation with feedback from domain experts further demonstrates its practical utility., Comment: 21 pages, 9 figures, 7 tables

The application of social robots in persuading people to change health behaviors is an increasing research topic. However, little is known in what ways, and under what conditions, effective health persuasion can be achieved in human-robot interaction (HRI). This position paper presents a conceptual model that integrates interpersonal relationship theories to postulate a mechanism through which social robots can change people’s health behaviors. In this paper, we first briefly describe the two interpersonal relationship theories we selectively focus on, namely social control and interdependence theory, and we discuss the possibility of people forming relationships with social robots. Then, we propose the conceptual model depicting the potential positive and negative influence of social robots’ health persuasion on people’s psychological and behavioral reactions and the modulating role of human-robot relationships. Finally, we discuss the implications of this model for future research.

Artificial intelligence (AI) has been applied to health behavior change research for over a decade. Current research programs include machine learning for delivering just-in-time adaptive interventions, computational modeling of behavior change processes, and the use of social AI for communication and persuasion. With new advances in AI, we propose an international workshop to bring together experts from all related disciplines to discuss and explore the potentials of AI for behavior change research. We discuss in this proposal the aims, planned activities, expected outcomes, and a promotion strategy for the workshop.

Graph property prediction tasks are important and numerous. While each task offers a small size of labeled examples, unlabeled graphs have been collected from various sources and at a large scale. A conventional approach is training a model with the unlabeled graphs on self-supervised tasks and then fine-tuning the model on the prediction tasks. However, the self-supervised task knowledge could not be aligned or sometimes conflicted with what the predictions needed. In this paper, we propose to extract the knowledge underlying the large set of unlabeled graphs as a specific set of useful data points to augment each property prediction model. We use a diffusion model to fully utilize the unlabeled graphs and design two new objectives to guide the model's denoising process with each task's labeled data to generate task-specific graph examples and their labels. Experiments demonstrate that our data-centric approach performs significantly better than fifteen existing various methods on fifteen tasks. The performance improvement brought by unlabeled data is visible as the generated labeled examples unlike the self-supervised learning., Comment: Accepted by NeurIPS 2023

This thesis investigates how a speech alert that uses a driver’s own name affects the speed, accuracy, and subjective experience of responding to transitions of control in the context of semi-automated driving. Preceding studies have shown that a personal name can be an effective stimulus to capture attention. However, studies that used names for alerts in automotive settings found mixed results when comparing the effectiveness between name and non-name conditions. In this thesis, an experiment with a simulated semi-automated car was conducted, in which participants were watching driving videos while playing a mobile game. Occasionally they heard speech alerts that started with either their name or a warning sound. They were required to immediately disengage from the game and respond to the alerts by pressing a button on the steering wheel or by pressing the brake. Results showed that participants’ response times were significantly faster when hearing the alerts containing their names. There was no difference in subjective experience between conditions found. The results suggest that one’s name is a useful addition to an in-car alert and warrants deeper investigation for future research.

Rationale is defined as a subset of input features that best explains or supports the prediction by machine learning models. Rationale identification has improved the generalizability and interpretability of neural networks on vision and language data. In graph applications such as molecule and polymer property prediction, identifying representative subgraph structures named as graph rationales plays an essential role in the performance of graph neural networks. Existing graph pooling and/or distribution intervention methods suffer from lack of examples to learn to identify optimal graph rationales. In this work, we introduce a new augmentation operation called environment replacement that automatically creates virtual data examples to improve rationale identification. We propose an efficient framework that performs rationale-environment separation and representation learning on the real and augmented examples in latent spaces to avoid the high complexity of explicit graph decoding and encoding. Comparing against recent techniques, experiments on seven molecular and four polymer real datasets demonstrate the effectiveness and efficiency of the proposed augmentation-based graph rationalization framework., Comment: Accepted by KDD 2022

This thesis investigates how a speech alert that uses a driver’s own name affects the speed, accuracy, and subjective experience of responding to transitions of control in the context of semi-automated driving. Preceding studies have shown that a personal name can be an effective stimulus to capture attention. However, studies that used names for alerts in automotive settings found mixed results when comparing the effectiveness between name and non-name conditions. In this thesis, an experiment with a simulated semi-automated car was conducted, in which participants were watching driving videos while playing a mobile game. Occasionally they heard speech alerts that started with either their name or a warning sound. They were required to immediately disengage from the game and respond to the alerts by pressing a button on the steering wheel or by pressing the brake. Results showed that participants’ response times were significantly faster when hearing the alerts containing their names. There was no difference in subjective experience between conditions found. The results suggest that one’s name is a useful addition to an in-car alert and warrants deeper investigation for future research.

Physiological computing uses human physiological data as system inputs in real time. It includes, or significantly overlaps with, brain-computer interfaces, affective computing, adaptive automation, health informatics, and physiological signal based biometrics. Physiological computing increases the communication bandwidth from the user to the computer, but is also subject to various types of adversarial attacks, in which the attacker deliberately manipulates the training and/or test examples to hijack the machine learning algorithm output, leading to possible user confusion, frustration, injury, or even death. However, the vulnerability of physiological computing systems has not been paid enough attention to, and there does not exist a comprehensive review on adversarial attacks to them. This paper fills this gap, by providing a systematic review on the main research areas of physiological computing, different types of adversarial attacks and their applications to physiological computing, and the corresponding defense strategies. We hope this review will attract more research interests on the vulnerability of physiological computing systems, and more importantly, defense strategies to make them more secure., Comment: National Science Open, 2022

Our research topic is Human segmentation with static camera. This topic can be divided into three sub-tasks, which are object detection, instance identification and segmentation. These sub-tasks are three closely related subjects. The development of each subject has great impact on the other two fields. In this literature review, we will first introduce the background of human segmentation and then talk about issues related to the above three fields as well as how they interact with each other.

In this paper, we use the house price data ranging from January 2004 to October 2016 to predict the average house price of November and December in 2016 for each district in Beijing, Shanghai, Guangzhou and Shenzhen. We apply Autoregressive Integrated Moving Average model to generate the baseline while LSTM networks to build prediction model. These algorithms are compared in terms of Mean Squared Error. The result shows that the LSTM model has excellent properties with respect to predict time series. Also, stateful LSTM networks and stack LSTM networks are employed to further study the improvement of accuracy of the house prediction model.