Search

Emotion has long been considered to distinguish humans from Artificial Intelligence (AI). Previously, AI's ability to interpret and express emotions was seen as mere text interpretation. In humans, emotions co-ordinate a suite of behavioral actions, e.g., under negative emotion being risk averse or under positive emotion being generous. So, we investigated such coordination to emotional cues in AI chatbots. We treated AI chatbots like human participants, prompting them with scenarios that prime positive emotions, negative emotions, or no emotions. Multiple OpenAI ChatGPT Plus accounts answered questions on investment decisions and prosocial tendencies. We found that ChatGPT-4 bots primed with positive emotions, negative emotions, and no emotions exhibited different risk-taking and prosocial actions. These effects were weaker among ChatGPT-3.5 bots. The ability to coordinate responses with emotional cues may have become stronger in large language models as they evolved. This highlights the potential of influencing AI using emotion and it suggests that complex AI possesses a necessary capacity for “having” emotion.

Supplemental material, sj-docx-1-pac-10.1177_18344909231166758 for Growth mindset of meaning in Life: Viewing meaning in life as malleable matters by Zhen Huang, Yiwen Wu, Yukun Zhao and Kaiping Peng in Journal of Pacific Rim Psychology

Supplemental material, sj-docx-1-pac-10.1177_18344909231157466 for Autonomy matters: Influences of causality orientations on Chinese adolescents’ growth mindset by Yukun Zhao, Zhen Huang, Yiwen Wu and Kaiping Peng in Journal of Pacific Rim Psychology

Supplement 1

Supplemental material, sj-docx-1-pac-10.1177_18344909231157466 for Autonomy matters: Influences of causality orientations on Chinese adolescents’ growth mindset by Yukun Zhao, Zhen Huang, Yiwen Wu and Kaiping Peng in Journal of Pacific Rim Psychology

Supplement 1

Natural language understanding (NLU) models often rely on dataset biases rather than intended task-relevant features to achieve high performance on specific datasets. As a result, these models perform poorly on datasets outside the training distribution. Some recent studies address this issue by reducing the weights of biased samples during the training process. However, these methods still encode biased latent features in representations and neglect the dynamic nature of bias, which hinders model prediction. We propose an NLU debiasing method, named debiasing contrastive learning (DCT), to simultaneously alleviate the above problems based on contrastive learning. We devise a debiasing, positive sampling strategy to mitigate biased latent features by selecting the least similar biased positive samples. We also propose a dynamic negative sampling strategy to capture the dynamic influence of biases by employing a bias-only model to dynamically select the most similar biased negative samples. We conduct experiments on three NLU benchmark datasets. Experimental results show that DCT outperforms state-of-the-art baselines on out-of-distribution datasets while maintaining in-distribution performance. We also verify that DCT can reduce biased latent features from the model's representation., Accepted by AAAI2023

Additional file 1: Table S1. Comparisons of baseline characteristics between the original cohort and matched cohort.

Supporting content for the paper

Supporting information for manuscript

In path tracking control systems of unmanned helicopters, it is essential to have high-performance controllers. Currently many design methods for tracking controllers have general control effects, but it is difficult to provide theoretical and practical basis for real unmanned helicopter tracking control systems. In this paper, the external PID controller of an unmanned helicopter is designed, in which an improved particle swarm optimization (PSO) algorithm is proposed to adjust controller parameters. To test and verify the performance of the controller, an improved artificial potential field (APF) method of tracking with combination of unmanned helicopter model. Firstly, safe distance of the unmanned helicopter is reserved in the course of flight. Secondly, the path planning produced by artificial potential field method is taken as the expected trajectory and the input into the unmanned helicopter model in real time to verify the control effect of the controller under complex and irregular conditions. Finally, the simulation is conducted to verify the feasibility of the proposed method and the rapidity of the responses. The simulation results show that the flight trajectory of the unmanned helicopter is basically consistent with the desired trajectory, and the tracking error trajectory has a high degree of convergence with reaching the target point smoothly

Acetamide is a promising phase change materials (PCMs) for thermal storage,but the large supercooling during the freezing process has limited its application. In this study, we prepared acetamide-SiO2composites by adding nano-SiO2into acetamide. This modified PCM was then impregnated into the porous graphite foam forming acetamide-SiO2-graphite foam form-stable composites. These composites were subjected to melting-solidification cycles 50 times; the time-temperature curves were tracked and recorded during these cycles. The time-temperature curves showed that, for the acetamide containing 2 wt. % SiO2, the supercooling phenomenon was eliminated and the material’s performance was stable for 50 cycles. The solidification temperature of the acetamide-SiO2-graphite foam samples was 65°C and the melting temperature was lowered to 65°C. The samples exhibited almost no supercooling and the presence of SiO2had no significant effect on the melting-solidification temperature. The microscopic supercooling of the acetamide-SiO2composite was measured using differential scanning calorimetry (DSC). The results indicated that when the content of SiO2was 1 wt. to 2 wt. %, the supercooling could be reduced to less than 10°C and heat was sufficiently released during solidification. Finally, a set of algorithms was derived using MATLAB software for simulating the crystallization of samples based on the classical nucleation theory. The results of the simulation agreed with the experiment results.

China abounds in hydropower resources. Due to precipitation and topographical features, these resources concentrate principally in the Southwest, Central South and Northwest Regions. In Eastern China (including East, North and Northeast Regions), hydropower potential is comparatively less and is mainly used for taking up the peak load of thermal-based electric systems. While in Western China, hydropower resource is so abundant that it forms the bulk of electric energy supply, and the long distance transmission of electric power from West to Eastern China will play an important role in the near future. During the past 32 years, much hydropower development has been accomplished in China. Up to the present time, the total installed generating capacity of water power throughout China is approximately 22 GW with annual output of 65 TWh, which is only 3% of the exploitable potential. Thus, the prospect of hydroelectric power development in China has a very brilliant future.