Transformer transfer learning emotion detection model: synchronizing socially agreed and self-reported emotions in big data.

Tactics to determine the emotions of authors of texts such as Twitter messages often rely on multiple annotators who label relatively small data sets of text passages. An alternative method gathers large text databases that contain the authors' self-reported emotions, to which artificial intelligence, machine learning, and natural language processing tools can be applied. Both approaches have strength and weaknesses. Emotions evaluated by a few human annotators are susceptible to idiosyncratic biases that reflect the characteristics of the annotators. But models based on large, self-reported emotion data sets may overlook subtle, social emotions that human annotators can recognize. In seeking to establish a means to train emotion detection models so that they can achieve good performance in different contexts, the current study proposes a novel transformer transfer learning approach that parallels human development stages: (1) detect emotions reported by the texts' authors and (2) synchronize the model with social emotions identified in annotator-rated emotion data sets. The analysis, based on a large, novel, self-reported emotion data set (n = 3,654,544) and applied to 10 previously published data sets, shows that the transfer learning emotion model achieves relatively strong performance.

deepNIR: Datasets for Generating Synthetic NIR Images and Improved Fruit Detection System Using Deep Learning Techniques.

This paper presents datasets utilised for synthetic near-infrared (NIR) image generation and bounding-box level fruit detection systems. A high-quality dataset is one of the essential building blocks that can lead to success in model generalisation and the deployment of data-driven deep neural networks. In particular, synthetic data generation tasks often require more training samples than other supervised approaches. Therefore, in this paper, we share the NIR+RGB datasets that are re-processed from two public datasets (i.e., nirscene and SEN12MS), expanded our previous study, deepFruits, and our novel NIR+RGB sweet pepper (capsicum) dataset. We oversampled from the original nirscene dataset at 10, 100, 200, and 400 ratios that yielded a total of 127 k pairs of images. From the SEN12MS satellite multispectral dataset, we selected Summer (45 k) and All seasons (180k) subsets and applied a simple yet important conversion: digital number (DN) to pixel value conversion followed by image standardisation. Our sweet pepper dataset consists of 1615 pairs of NIR+RGB images that were collected from commercial farms. We quantitatively and qualitatively demonstrate that these NIR+RGB datasets are sufficient to be used for synthetic NIR image generation. We achieved Frechet inception distances (FIDs) of 11.36, 26.53, and 40.15 for nirscene1, SEN12MS, and sweet pepper datasets, respectively. In addition, we release manual annotations of 11 fruit bounding boxes that can be exported in various formats using cloud service. Four newly added fruits (blueberry, cherry, kiwi and wheat) compound 11 novel bounding box datasets on top of our previous work presented in the deepFruits project (apple, avocado, capsicum, mango, orange, rockmelon and strawberry). The total number of bounding box instances of the dataset is 162 k and it is ready to use from a cloud service. For the evaluation of the dataset, Yolov5 single stage detector is exploited and reported impressive mean-average-precision, mAP[0.5:0.95] results of min:0.49, max:0.812. We hope these datasets are useful and serve as a baseline for future studies.

Older adults' experiences and perceptions of living with Bomy, an assistive dailycare robot: a qualitative study.

An aging global population and preference for aging-in-place pose the opportunity for home-based robots to assist older adults with their daily routines. However, there is limited research into the experiences of older adults using robots in their own homes. In this descriptive qualitative feasibility study, older self-supporting and community-dwelling adults with various age-related health needs used Bomy, a dailycare robot in their homes for up to one week. The study explored the usefulness of the robot and participants' perceptions and experiences of using it. Bomy reminded them of daily activities and delivered cognitive stimulation games. Semi-structured in-person interviews were conducted, and data were analyzed thematically. Findings revealed an acceptance toward robots and the value of assistive dailycare robots. Participants perceived Bomy as a companion and made suggestions for improvement, including resolving technical issues associated with long-term use. Future functions should be personalizable, to accommodate each user's health needs and could also include smoke detection and reading aloud functions. Dailycare robots show promising potential in elderly care, especially in providing reminders for medication, health and wellbeing. This study highlights the importance of co-design and testing robotics in the environments for which they have been developed. Widespread implementation of Bomy might be feasible in the future, with some further adjustments.

Subsentence Extraction from Text Using Coverage-Based Deep Learning Language Models.

Sentiment prediction remains a challenging and unresolved task in various research fields, including psychology, neuroscience, and computer science. This stems from its high degree of subjectivity and limited input sources that can effectively capture the actual sentiment. This can be even more challenging with only text-based input. Meanwhile, the rise of deep learning and an unprecedented large volume of data have paved the way for artificial intelligence to perform impressively accurate predictions or even human-level reasoning. Drawing inspiration from this, we propose a coverage-based sentiment and subsentence extraction system that estimates a span of input text and recursively feeds this information back to the networks. The predicted subsentence consists of auxiliary information expressing a sentiment. This is an important building block for enabling vivid and epic sentiment delivery (within the scope of this paper) and for other natural language processing tasks such as text summarisation and Q&A. Our approach outperforms the state-of-the-art approaches by a large margin in subsentence prediction (i.e., Average Jaccard scores from 0.72 to 0.89). For the evaluation, we designed rigorous experiments consisting of 24 ablation studies. Finally, our learned lessons are returned to the community by sharing software packages and a public dataset that can reproduce the results presented in this paper.

Correction: The Effect of Robot Attentional Behaviors on User Perceptions and Behaviors in a Simulated Health Care Interaction: Randomized Controlled Trial.

[This corrects the article DOI: 10.2196/13667.].

The Effect of Robot Attentional Behaviors on User Perceptions and Behaviors in a Simulated Health Care Interaction: Randomized Controlled Trial.

BACKGROUND: For robots to be effectively used in health applications, they need to display appropriate social behaviors. A fundamental requirement in all social interactions is the ability to engage, maintain, and demonstrate attention. Attentional behaviors include leaning forward, self-disclosure, and changes in voice pitch. OBJECTIVE: This study aimed to examine the effect of robot attentional behaviors on user perceptions and behaviors in a simulated health care interaction. METHODS: A parallel randomized controlled trial with a 1:1:1:1 allocation ratio was conducted. We randomized participants to 1 of 4 experimental conditions before engaging in a scripted face-to-face interaction with a fully automated medical receptionist robot. Experimental conditions included a self-disclosure condition, voice pitch change condition, forward lean condition, and neutral condition. Participants completed paper-based postinteraction measures relating to engagement, perceived robot attention, and perceived robot empathy. We video recorded interactions and coded for participant attentional behaviors. RESULTS: A total of 181 participants were recruited from the University of Auckland. Participants who interacted with the robot in the forward lean and self-disclosure conditions found the robot to be significantly more stimulating than those who interacted with the robot in the voice pitch or neutral conditions (P=.03). Participants in the forward lean, self-disclosure, and neutral conditions found the robot to be significantly more interesting than those in the voice pitch condition (P<.001). Participants in the forward lean and self-disclosure conditions spent significantly more time looking at the robot than participants in the neutral condition (P<.001). Significantly, more participants in the self-disclosure condition laughed during the interaction (P=.01), whereas significantly more participants in the forward lean condition leant toward the robot during the interaction (P<.001). CONCLUSIONS: The use of self-disclosure and forward lean by a health care robot can increase human engagement and attentional behaviors. Voice pitch changes did not increase attention or engagement. The small effects with regard to participant perceptions are potentially because of the limitations in self-report measures or a lack of comparison for most participants who had never interacted with a robot before. Further research could explore the use of self-disclosure and forward lean using a within-subjects design and in real health care settings.

Developing assistive robots for people with mild cognitive impairment and mild dementia: a qualitative study with older adults and experts in aged care.

OBJECTIVES: This research is part of an international project to design and test a home-based healthcare robot to help older adults with mild cognitive impairment (MCI) or early dementia. The aim was to investigate the perceived usefulness of different daily-care activities for the robot, developed from previous research on needs. DESIGN: Qualitative descriptive analysis using semistructured interviews. Two studies were conducted. In the first study, participants watched videos of a prototype robot performing daily-care activities; in the second study, participants interacted with the robot itself. SETTING: Interviews were conducted at a university and a retirement village. PARTICIPANTS: In study 1, participants were nine experts in aged care and nine older adults living in an aged care facility. In study 2, participants were 10 experts in aged care. RESULTS: The themes that emerged included aspects of the robot's interactions, potential benefits, the appearance, actions and humanness of the robot, ways to improve its functionality and technical issues. Overall, the activities were perceived as useful, especially the reminders and safety checks, with possible benefits of companionship, reassurance and reduced caregiver burden. Suggestions included personalising the robot to each individual, simplifying the language and adding more activities. Technical issues still need to be fixed. CONCLUSION: This study adds to knowledge about healthcare robots for people with MCI by developing and testing a new robot with daily-care activities including safety checks. The robot was seen to be potentially useful but needs to be tested with people with MCI.

Using Robots at Home to Support Patients With Chronic Obstructive Pulmonary Disease: Pilot Randomized Controlled Trial.

BACKGROUND: Socially assistive robots are being developed for patients to help manage chronic health conditions such as chronic obstructive pulmonary disease (COPD). Adherence to medication and availability of rehabilitation are suboptimal in this patient group, which increases the risk of hospitalization. OBJECTIVE: This pilot study aimed to investigate the effectiveness of a robot delivering telehealth care to increase adherence to medication and home rehabilitation, improve quality of life, and reduce hospital readmission compared with a standard care control group. METHODS: At discharge from hospital for a COPD admission, 60 patients were randomized to receive a robot at home for 4 months or to a control group. Number of hospitalization days for respiratory admissions over the 4-month study period was the primary outcome. Medication adherence, frequency of rehabilitation exercise, and quality of life were also assessed. Implementation interviews as well as benefit-cost analysis were conducted. RESULTS: Intention-to-treat and per protocol analyses showed no significant differences in the number of respiratory-related hospitalizations between groups. The intervention group was more adherent to their long-acting inhalers (mean number of prescribed puffs taken per day=48.5%) than the control group (mean 29.5%, P=.03, d=0.68) assessed via electronic recording. Self-reported adherence was also higher in the intervention group after controlling for covariates (P=.04). The intervention group increased their rehabilitation exercise frequency compared with the control group (mean difference -4.53, 95% CI -7.16 to -1.92). There were no significant differences in quality of life. Of the 25 patients who had the robot, 19 had favorable attitudes. CONCLUSIONS: This pilot study suggests that a homecare robot can improve adherence to medication and increase exercise. Further research is needed with a larger sample size to further investigate effects on hospitalizations after improvements are made to the robots. The robots could be especially useful for patients struggling with adherence. TRIAL REGISTRATION: Australian New Zealand Clinical Trials Registry ACTRN12615000259549; http://www.anzctr.org.au (Archived by WebCite at  http://www.webcitation.org/6whIjptLS).

Homecare Robots to Improve Health and Well-Being in Mild Cognitive Impairment and Early Stage Dementia: Results From a Scoping Study.

OBJECTIVES: This scoping study is the first step of a multiphase, international project aimed at designing a homecare robot that can provide functional support, track physical and psychological well-being, and deliver therapeutic intervention specifically for individuals with mild cognitive impairment. DESIGN: Observational requirements gathering study. PARTICIPANTS AND SETTINGS: Semistructured interviews were conducted with 3 participant groups: (1) individuals with memory challenges, mild cognitive impairment (MCI), or mild dementia (patients; n = 9); (2) carers of those with MCI or dementia (carers; n = 8); and (3) those with expertise in MCI or dementia research, clinical care, or management (experts; n = 16). Interviews took place at the university, at dementia care facilities or other workplaces, at participant's homes, or via skype (experts only). MEASUREMENTS: Semistructured interviews were conducted, transcribed, and reviewed. RESULTS: Several key themes were identified within the 4 topics of: (1) daily challenges, (2) safety and security, (3) monitoring health and well-being, and (4) therapeutic intervention. CONCLUSIONS: A homecare robot could provide both practical and therapeutic benefit for the mildly cognitively impaired with 2 broad programs providing routine and reassurance; and tracking health and well-being. The next phase of the project aims to program homecare robots with scenarios developed from these results, integrate components from project partners, and then test the feasibility, utility, and acceptability of the homecare robot.