Enterprise API design is often bottlenecked by the tension between rapid feature delivery and the rigorous maintenance of usability standards. We present an industrial case study evaluating an AI-assisted design workflow trained on API Improvement Proposals (AIPs). Through a controlled study with 16 industry experts, we compared AI-generated API specifications against human-authored ones. While quantitative results indicated AI superiority in 10 of 11 usability dimensions and an 87% reduction in authoring time, qualitative analysis revealed a paradox: experts frequently misidentified AI work as human (19% accuracy) yet described the designs as unsettlingly "perfect." We characterize this as a "Perfection Paradox" -- where hyper-consistency signals a lack of pragmatic human judgment. We discuss the implications of this perfection paradox, proposing a shift in the human designer's role from the "drafter" of specifications to the "curator" of AI-generated patterns.
Recent advances in large language models (LLMs) have sparked growing interest in building fully autonomous agents. However, fully autonomous LLM-based agents still face significant challenges, including limited reliability due to hallucinations, difficulty in handling complex tasks, and substantial safety and ethical risks, all of which limit their feasibility and trustworthiness in real-world applications. To overcome these limitations, LLM-based human-agent systems (LLM-HAS) incorporate human-provided information, feedback, or control into the agent system to enhance system performance, reliability and safety. These human-agent collaboration systems enable humans and LLM-based agents to collaborate effectively by leveraging their complementary strengths. This paper provides the first comprehensive and structured survey of LLM-HAS. It clarifies fundamental concepts, systematically presents core components shaping these systems, including environment & profiling, human feedback, interaction types, orchestration and communication, explores emerging applications, and discusses unique challenges and opportunities arising from human-AI collaboration. By consolidating current knowledge and offering a structured overview, we aim to foster further research and innovation in this rapidly evolving interdisciplinary field. Paper lists and resources are available at https://github.com/HenryPengZou/Awesome-Human-Agent-Collaboration-Interaction-Systems.
Chitralekha Gupta, Hanjun Wu, Praveen Sasikumar
et al.
Wearable devices are transforming human capabilities by seamlessly augmenting cognitive functions. In this position paper, we propose a voice-based, interactive learning companion designed to amplify and extend cognitive abilities through informal learning. Our vision is threefold: (1) to enable users to discover new knowledge on-the-go through contextual interactive quizzes, fostering critical thinking and mindfulness, (2) to proactively detect misinformation, empowering users to critically assess information in real time, and (3) to provide spoken language correction and prompting hints for second language learning and effective communication. As an initial step toward this vision, we present Factually - a proactive, wearable fact-checking system integrated into devices like smartwatches or rings. Factually discreetly alerts users to potential falsehoods via vibrotactile feedback, helping them assess information critically. We demonstrate its utility through three illustrative scenarios, highlighting its potential to extend cognitive abilities for real-time misinformation detection. Early qualitative feedback suggests that Factually can enhance users' fact-checking capabilities, offering both practical and experiential benefits.
ABSTRACT Background Atrophy of the masseter muscle can result in an aged facial appearance and diminished chewing function. Electrical stimulation (ES) is known for its ability to facilitate tissue healing and functional recovery, but its precise role in the repair of atrophic masseter muscles remains incompletely understood. Methods We induced masseter muscle atrophy in rats through botulinum toxin (BTX) injection and subsequently treated the animals with or without ES. Single‐nucleus sequencing (sn‐RNA seq) was conducted to analyse the changes in macrophages of masseter muscles between control, BTX and BTX + ES groups. The role and mechanism of macrophage phenotypic transformation in the process of ES promoting the recovery of atrophied masseter muscles were both verified through in vivo and in vitro experiments. Results Our results indicate that ES treatment within defined current parameters significantly ameliorated muscle condition by reducing atrophy‐related gene expression (MuRF1: BTX: 10.15 ± 1.69; BTX + ES: 1.05 ± 0.06; Fbxo32: BTX: 8.62 ± 1.19, BTX + ES: 1.19 ± 0.07, p < 0.0001) and enhancing vascularisation (Vegf positive area: BTX: 6.60 ± 2.87%, BTX + ES: 27.23 ± 1.70%, p < 0.001). Analysis conducted with sn‐RNA seq demonstrated increased infiltration of M1 macrophages during muscle atrophy, with a subsequent transition to M2 macrophages following ES treatment (M1 macrophage portion: Ctrl: 15.2%, BTX: 25.8%, BTX + ES: 14.7%; M2 macrophages: Ctrl: 67.9%, BTX: 46.9%, BTX + ES: 70.5%). Further investigations demonstrated that ES reduced M1 macrophage infiltration (five‐fold lower of CD86+ cell number, BTX: 30 ± 2; BTX + ES: 6 ± 2, p < 0.0001) while increasing M2 macrophage presence (3.3‐fold higher of CD163+cell, BTX: 10 ± 3; BTX + ES: 33 ± 8, p < 0.01), potentially via activation of the PI3K‐Akt pathway (p‐Akt/Akt ratio, BTX:0.58 ± 0.20%; BTX + ES:1.03 ± 0.07%, p < 0.05). Depletion of macrophages using clodronate liposomes reversed the beneficial effects of ES on induced masseter atrophy (MuRF1: BTX + ES: 2.20 ± 0.16; BTX + ES + CL: 12.93 ± 0.98, p < 0.0001), highlighting the involvement of macrophages in the therapeutic process. In vitro studies demonstrated that ES promoted the transition from M1 to M2 macrophages and enhanced proliferation and differentiation of myogenic cells. Conclusions Our findings suggest that ES can enhance masseter muscle tissue repair by modulating macrophage polarisation, offering valuable insights into the potential of ES in noninvasive tissue regeneration strategies for treating masseter muscle atrophy.
Diseases of the musculoskeletal system, Human anatomy
Abstract Research on articular cartilage has primarily focused on athletes, particularly regarding increased metatarsal cartilage thickness in high-impact sports, with limited studies in military settings. This study examined Air Force Academy cadets, who experience unique mechanical demands on their hands. A total of 30 senior male cadets (age: 23.14 ± 0.51 years; height: 174.07 ± 3.14 cm; weight: 70.20 ± 6.63 kg) were evaluated based on their hand/wrist injury history, categorizing them into those with an injury history (IH; n,17) and non-injury history (NIH; n,13). Various assessments included body composition, hand grip strength, hand length, and ultrasonography, applying a validated tool (Michigan Hand Outcomes Questionnaire). Ultrasound measurements showed that the IH group had thicker cartilage than the NIH group in the index, ring, and little fingers, indicating the need for improved ergonomics in cockpit design and enhanced training protocols to mitigate injury risks among pilots.
AbstractThis study aims to cultivate students' independent learning capacity, promote the interdisciplinary integration of “nursing + anatomy,” and establish a curriculum system to enhance applied nursing abilities based on project‐based teaching reform of everyday clinical nursing operations. A total of 151 second‐year (class of 2021) undergraduate nursing students at a Chinese university were selected for this study. By adjusting the curriculum, reconstructing the teaching contents, employing the “hybrid + flip” teaching method based on BOPPPS (bridge‐in, outcomes, preassessment, participatory learning, post‐evaluation, summary), and implementing a teaching system based on the “three re‐three linkage,” a Human Anatomy curriculum with a focus on basic anatomical knowledge was developed and connected with nursing clinical operation practice. The restructuring of the course content received unanimous recognition from both the teaching staff and the students. Notably, students in the class of 2021 achieved significantly higher grades than did students in the class of 2020, who received traditional face‐to‐face instruction (p < 0.01). These results indicate enhanced clinical application skills among the former group of students. following the implementation of instructional reforms during one semester, students exhibited notable improvements in motivation, program implementation, self‐management, and interpersonal communication. A statistically significant increase in overall scores for self‐directed learning capacities over the preinstructional period was observed (p < 0.05). Furthermore, the findings of the student satisfaction surveys reflected highly favorable perceptions of the enriched instructional format, high levels of course engagement, frequent faculty–student interactions, and augmented overall competence. The practical implementation of the reform in the context of a Human Anatomy course for undergraduate nursing students led to significant positive outcomes, thereby enhancing the effectiveness of teaching and learning. Students' clinical application abilities and self‐directed learning capacities notably improved, while overall satisfaction with the course remained high.
Morten Roed Frederiksen, Katrin Fischer, Maja Matarić
This paper describes a between-subjects Amazon Mechanical Turk study (n = 220) that investigated how a robot's affective narrative influences its ability to elicit empathy in human observers. We first conducted a pilot study to develop and validate the robot's affective narratives. Then, in the full study, the robot used one of three different affective narrative strategies (funny, sad, neutral) while becoming less functional at its shopping task over the course of the interaction. As the functionality of the robot degraded, participants were repeatedly asked if they were willing to help the robot. The results showed that conveying a sad narrative significantly influenced the participants' willingness to help the robot throughout the interaction and determined whether participants felt empathetic toward the robot throughout the interaction. Furthermore, a higher amount of past experience with robots also increased the participants' willingness to help the robot. This work suggests that affective narratives can be useful in short-term interactions that benefit from emotional connections between humans and robots.
Mohamad Shaaban, Simone Macci`o, Alessandro Carf`ı
et al.
This article investigates mixed reality (MR) to enhance human-robot collaboration (HRC). The proposed solution adopts MR as a communication layer to convey a mobile manipulator's intentions and upcoming actions to the humans with whom it interacts, thus improving their collaboration. A user study involving 20 participants demonstrated the effectiveness of this MR-focused approach in facilitating collaborative tasks, with a positive effect on overall collaboration performances and human satisfaction.
With the rise of human-machine communication, machines are increasingly designed with humanlike characteristics, such as gender, which can inadvertently trigger cognitive biases. Many conversational agents (CAs), such as voice assistants and chatbots, default to female personas, leading to concerns about perpetuating gender stereotypes and inequality. Critiques have emerged regarding the potential objectification of females and reinforcement of gender stereotypes by these technologies. This research, situated in conversational AI design, aims to delve deeper into the impacts of gender biases in human-CA interactions. From a behavioral and communication research standpoint, this program focuses not only on perceptions but also the linguistic styles of users when interacting with CAs, as previous research has rarely explored. It aims to understand how pre-existing gender biases might be triggered by CAs' gender designs. It further investigates how CAs' gender designs may reinforce gender biases and extend them to human-human communication. The findings aim to inform ethical design of conversational agents, addressing whether gender assignment in CAs is appropriate and how to promote gender equality in design.
My research centers on the development of context-adaptive AI systems to improve end-user adoption through the integration of technical methods. I deploy these AI systems across various interaction modalities, including user interfaces and embodied agents like robots, to expand their practical applicability. My research unfolds in three key stages: design, development, and deployment. In the design phase, user-centered approaches were used to understand user experiences with AI systems and create design tools for user participation in crafting AI explanations. In the ongoing development stage, a safety-guaranteed AI system for a robot agent was created to automatically provide adaptive solutions and explanations for unforeseen scenarios. The next steps will involve the implementation and evaluation of context-adaptive AI systems in various interaction forms. I seek to prioritize human needs in technology development, creating AI systems that tangibly benefit end-users in real-world applications and enhance interaction experiences.
BackgroundPreeclampsia (PE), a pregnancy specific syndrome, is defined as new-onset hypertension (≥140/90 mmHg) and proteinuria diagnosed after gestational week 20 or new-onset pre-eclampsia associated signs in the absence of proteinuria, and it may tend to present as late as 4–6 weeks’ postpartum period. It is a leading cause of maternal mortality in both developed and developing countries. In order to prevent PE, the disease must be diagnosed at its earliest stage, however, the triads of high blood pressure, edema and albuminuria is neither specific nor sensitive enough for diagnosing the disease. Lactate dehydrogenase (LDH) is useful biochemical marker reflecting the occurrence of complications associated with preeclampsia. Besides, it has been suggested as potential biomarker to predict the severity of preeclampsia and as indicator of multi-organ involvement. The aim of this study was to investigate the diagnostic accuracy of LDH, which is affordable and easy to test, as a potential clinical biomarker to predict onset of preeclampsia.MethodsA hospital based cross-sectional study was conducted as of September 9 to December 24, 2022 at Debre Birhan Comprehensive Specialized Hospital (DBCSH). A total of 132 study subjects (66 preeclamptic and 66 normotensive controls) were enrolled in the study. A receiver operating characteristics (ROC) curve was used to calculate the area under the curve (AUC) and determine diagnostic accuracy of LDH. Youden’s index was used to identify an optimal cut-off point for LDH in detecting preeclampsia associated complications.ResultAUC for LDH was found to be 0.963 (95% CI, 0.91, 1.0; p = 0.000) from ROC curve analysis. An optimal cut-off point for LDH was 376.5 U/L having a sensitivity and specificity of 87.5 and 90.8%, respectively.ConclusionSerum LDH had an AUC of greater than 0.8 and showed good diagnostic accuracy in predicting development of preeclampsia. Disease duration, gestational age, systolic and diastolic blood pressure among enormous number of predictor variables had association with serum level of LDH.
Proprioceptive training is crucial for enhancing coordination, balance, and injury prevention in athletes, especially in their developmental stages. This study investigates the application of the Gyko sensor for testing proprioceptive abilities in junior athletes. The Gyko sensor, a portable motion analysis device, provides real-time data on movement and postural control. The study assesses the effectiveness of proprioceptive training by analyzing parameters such as balance, stability, and joint positioning accuracy in young athletes. Results highlight the sensor's potential in monitoring improvements and tailoring individualized training programs to optimize athletic performance and reduce injury risk.
Predicting human mobility holds significant practical value, with applications ranging from enhancing disaster risk planning to simulating epidemic spread. In this paper, we present the GeoFormer, a decoder-only transformer model adapted from the GPT architecture to forecast human mobility. Our proposed model is rigorously tested in the context of the HuMob Challenge 2023 -- a competition designed to evaluate the performance of prediction models on standardized datasets to predict human mobility. The challenge leverages two datasets encompassing urban-scale data of 25,000 and 100,000 individuals over a longitudinal period of 75 days. GeoFormer stands out as a top performer in the competition, securing a place in the top-3 ranking. Its success is underscored by performing well on both performance metrics chosen for the competition -- the GEO-BLEU and the Dynamic Time Warping (DTW) measures. The performance of the GeoFormer on the HuMob Challenge 2023 underscores its potential to make substantial contributions to the field of human mobility prediction, with far-reaching implications for disaster preparedness, epidemic control, and beyond.
Reinforcement Learning from Human Feedback (RLHF) facilitates the alignment of large language models with human preferences, significantly enhancing the quality of interactions between humans and models. InstructGPT implements RLHF through several stages, including Supervised Fine-Tuning (SFT), reward model training, and Proximal Policy Optimization (PPO). However, PPO is sensitive to hyperparameters and requires multiple models in its standard implementation, making it hard to train and scale up to larger parameter counts. In contrast, we propose a novel learning paradigm called RRHF, which scores sampled responses from different sources via a logarithm of conditional probabilities and learns to align these probabilities with human preferences through ranking loss. RRHF can leverage sampled responses from various sources including the model responses from itself, other large language model responses, and human expert responses to learn to rank them. RRHF only needs 1 to 2 models during tuning and can efficiently align language models with human preferences robustly without complex hyperparameter tuning. Additionally, RRHF can be considered an extension of SFT and reward model training while being simpler than PPO in terms of coding, model counts, and hyperparameters. We evaluate RRHF on the Helpful and Harmless dataset, demonstrating comparable alignment performance with PPO by reward model score and human labeling. Extensive experiments show that the performance of RRHF is highly related to sampling quality which suggests RRHF is a best-of-n learner. Codes available at https://github.com/GanjinZero/RRHF.
Meriam Moujahid, David A. Robb, Christian Dondrup
et al.
Social Robots in human environments need to be able to reason about their physical surroundings while interacting with people. Furthermore, human proxemics behaviours around robots can indicate how people perceive the robots and can inform robot personality and interaction design. Here, we introduce Charlie, a situated robot receptionist that can interact with people using verbal and non-verbal communication in a dynamic environment, where users might enter or leave the scene at any time. The robot receptionist is stationary and cannot navigate. Therefore, people have full control over their personal space as they are the ones approaching the robot. We investigated the influence of different apparent robot personalities on the proxemics behaviours of the humans. The results indicate that different types of robot personalities, specifically introversion and extroversion, can influence human proxemics behaviours. Participants maintained shorter distances with the introvert robot receptionist, compared to the extrovert robot. Interestingly, we observed that human-robot proxemics were not the same as typical human-human interpersonal distances, as defined in the literature. We therefore propose new proxemics zones for human-robot interaction.
Merle M. Reimann, Florian A. Kunneman, Catharine Oertel
et al.
As social robots see increasing deployment within the general public, improving the interaction with those robots is essential. Spoken language offers an intuitive interface for the human-robot interaction (HRI), with dialogue management (DM) being a key component in those interactive systems. Yet, to overcome current challenges and manage smooth, informative and engaging interaction a more structural approach to combining HRI and DM is needed. In this systematic review, we analyse the current use of DM in HRI and focus on the type of dialogue manager used, its capabilities, evaluation methods and the challenges specific to DM in HRI. We identify the challenges and current scientific frontier related to the DM approach, interaction domain, robot appearance, physical situatedness and multimodality.
Fadil Elamin, Hassan Yahya Hassan Mohamed, Nihal Abdelazeem
et al.
Background Maturation of bones in the hand–wrist region varies among individuals of the same age and among world groups. Although some studies from Africa report differences to other ethnic groups, the lack of detailed bone-specific maturity data prevents meaningful comparisons. Aim The aim of this study was to describe bone-specific maturity for developing hand–wrist bones in individuals in Khartoum, Sudan. Subjects and methods The sample was selected from healthy patients attending a dental hospital in Khartoum with known age and ancestry (males = 280, females = 330; aged between 3 and 25 years). Bones were assessed from radiographs of the left hand and wrist after the Greulich and Pyle Atlas (1959). Median ages of attainment for bone stages were calculated using probit analysis for each stage in males and females separately. Results Maturity data for stages of the phalanges, metacarpals, carpals and radius and ulna in males and females are presented. Median ages in females were earlier compared to males for all stages. These results are largely earlier than previously published findings or where these could be calculated. Conclusion These results of individual maturity stages of phalanges, metacarpals, carpals and the distal epiphyses of the radius and ulna are useful to assess maturity in growing individuals from Sudan.