Human Preferences Research Articles

A HUG taxonomy of humans with potential in human–robot hugs

Humans can easily perform various types of hugs in human contact and affection experience. With the prevalence of robots in social applications, they would be expected to possess the capability of hugs as humans do. However, it is still not an easy task for robots, considering the complex force and spatial constraints of robot hugs. In this work, we propose the HUG taxonomy, which distinguishes between different hugging patterns based on human demonstrations and prior knowledge. In this taxonomy, hugs are arranged according to (1) hugging tightness, (2) hugging style, and (3) bilateral coordination, resulting in 16 different hug types. We then further study the hug type preference of humans in different scenarios and roles. Furthermore, we propose a rule-based classification system to validate the potential of this taxonomy in human–robot hugs based on a humanoid robot with an E-skin of contact sensation. The HUG taxonomy could provide human hugging behavior information in advance, facilitating the action control of humanoid robots. We believe the results of our work can benefit future studies on human–robot hugging interactions.

Deep Learning Social Filtering Model for Event Recommendation Services

In the contemporary time, technology has made the determination and discovery of human preferences, priorities and personal inclinations possible through the use of recommender systems. Activities of users on the internet can be monitored, extracted, stored, analyzed and used by the recommender systems for suggesting future events to users on the web. This paper aims at developing and analyzing a model for event services recommendation for visitors to events. Event seekers, organizers and event service providers get notified, plan and book for upcoming events from their comfort zones without hassles of gallivanting nooks and crannies to enquire about prospective events. There is not any compelling need to interface with under-enthusiasts and intermediaries in the course of organizing, visiting and providing services for an event. However, it is obvious that massive amount of available information on the web exhibit frustrating attributes, hence it is increasingly a difficult task for users to find the content of interest; in other words, a huge chunk of information undiscovered on the network is left behind as “dark information”. In context, event service recommendation uses deep learning social filtering base techniques which adopt similarity computation measures with a bias for Pearson correlation coefficient, cosine similarity, and Euclidean similarity to recommend related and most relevant events/services to the targeted online audience. In this paper, the aim is to develop a deep learning model which integrates social filtering technique for enhancing the quality of event recommendation for users. A model based on the deep learning algorithm of multilayered perceptron and Neural Collaborative Filtering is proposed for event recommender services. The results from various simulations using meetup website dataset shows that the proposed model performs better than other techniques. The results yield 70% accuracy, 66% precision and 98% recall.

Alien introgression to wheat for food security: functional and nutritional quality for novel products under climate change.

Crop yield and quality has increased globally during recent decades due to plant breeding, resulting in improved food security. However, climate change and shifts in human dietary habits and preferences display novel pressure on crop production to deliver enough quantity and quality to secure food for future generations. This review paper describes the current state-of-the-art and presents innovative approaches related to alien introgressions into wheat, focusing on aspects related to quality, functional characteristics, nutritional attributes, and development of novel food products. The benefits and opportunities that the novel and traditional plant breeding methods contribute to using alien germplasm in plant breeding are also discussed. In principle, gene introgressions from rye have been the most widely utilized alien gene source for wheat. Furthermore, the incorporation of novel resistance genes toward diseases and pests have been the most transferred type of genes into the wheat genome. The incorporation of novel resistance genes toward diseases and pests into the wheat genome is important in breeding for increased food security. Alien introgressions to wheat from e.g. rye and Aegilops spp. have also contributed to improved nutritional and functional quality. Recent studies have shown that introgressions to wheat of genes from chromosome 3 in rye have an impact on both yield, nutritional and functional quality, and quality stability during drought treatment, another character of high importance for food security under climate change scenarios. Additionally, the introgression of alien genes into wheat has the potential to improve the nutritional profiles of future food products, by contributing higher minerals levels or lower levels of anti-nutritional compounds into e.g., plant-based products substituting animal-based food alternatives. To conclude, the present review paper highlights great opportunities and shows a few examples of how food security and functional-nutritional quality in traditional and novel wheat products can be improved by the use of genes from alien sources, such as rye and other relatives to wheat. Novel and upcoming plant breeding methods such as genome-wide association studies, gene editing, genomic selection and speed breeding, have the potential to complement traditional technologies to keep pace with climate change and consumer eating habits.

Simultaneously learning intentions and preferences during physical human-robot cooperation

The advent of collaborative robots allows humans and robots to cooperate in a direct and physical way. While this leads to amazing new opportunities to create novel robotics applications, it is challenging to make the collaboration intuitive for the human. From a system’s perspective, understanding the human intentions seems to be one promising way to get there. However, human behavior exhibits large variations between individuals, such as for instance preferences or physical abilities. This paper presents a novel concept for simultaneously learning a model of the human intentions and preferences incrementally during collaboration with a robot. Starting out with a nominal model, the system acquires collaborative skills step-by-step within only very few trials. The concept is based on a combination of model-based reinforcement learning and inverse reinforcement learning, adapted to fit collaborations in which human and robot think and act independently. We test the method and compare it to two baselines: one that imitates the human and one that uses plain maximum entropy inverse reinforcement learning, both in simulation and in a user study with a Franka Emika Panda robot arm.

Designing Reward Functions Using Active Preference Learning for Reinforcement Learning in Autonomous Driving Navigation

This study presents a method based on active preference learning to overcome the challenges of designing reward functions for autonomous navigation. Results obtained from training with artificially designed reward functions may not accurately reflect human intentions. We focus on the limitations of traditional reward functions, which often fail to facilitate complex tasks in continuous state spaces. We propose the adoption of active preference learning to resolve these issues and to generate reward functions that align with human preferences. This approach leverages an individual’s subjective preferences to guide an agent’s learning process, enabling the creation of reward functions that reflect human desires. We utilize mutual information to generate informative queries and apply information gained to balance the agent’s uncertainty with the human’s response capacity, encouraging the agent to pose straightforward and informative questions. We further employ the No-U-Turn Sampler (NUTS) method to refine the belief model, which outperforms that constructed using the Metropolis algorithm. Subsequently, we retrain the agent using reward weights derived from active preference learning. As a result, our autonomous driving vehicle can navigate between random starting and ending points without dependence on high-precision maps or routing, relying solely on its forward vision. We validate our approach’s performance within the CARLA simulation environment. Our algorithm significantly improved the success rate of autonomous driving navigation tasks that originally failed due to artificially designed rewards, increasing it to approximately 60%. Experimental results show significant improvement over the baseline algorithm, providing a solid foundation for enhancing navigation capabilities in autonomous driving systems and advancing the field of autonomous driving intelligence.

Is the perception of odour pleasantness shared across cultures and ecological conditions? Evidence from Amazonia, East Africa, New Guinea, Malaysia and Poland.

What makes an odour pleasant or unpleasant? The inherent properties of the constituent chemical compounds, or the nose of the beholder, driven by idiosyncratic differences and culture-specific learning? Here, 582 individuals, including Tanzanian Hadza hunter-gatherers, Amazonian Tsimane' horticulturalists, Yali from the Papuan highlands and two industrialized populations (Poles, Malaysians), rated the pleasantness of 15 odour samples. We find considerable similarities in odour assessments across cultures, but our data do not fully support a claim regarding the universality of smell preferences. Despite cross-cultural similarities in olfactory assessments, probably driven by odour properties, we suggest that odour availability in ecological and cultural niches bears an undeniable effect on human odour preferences.

Modeling and Analyzing the Effect of Human Preferences on a Local Electricity Market

Modeling and Analyzing the Effect of Human Preferences on a Local Electricity Market

An Evolutionary Psychological Approach Toward BDSM Interest and Behavior

Bondage/discipline, Dominance/submission, and Sadism/Masochism (BDSM) have gained increased attention and discussion in recent years. This prevalence is accompanied by a shift in perceptions of BDSM, including the declassification of sadomasochism as a paraphilic disorder. Evolutionary psychology offers a unique perspective of why some individuals are interested in BDSM and why some prefer certain elements of BDSM over others (e.g., dominance versus submission). In this paper, we examine BDSM from an evolutionary standpoint, examining biopsychosocial factors that underlie the BDSM interests and practice. We articulate this perspective via an exploration of: proximate processes, such as the role of childhood experiences, sexual conditioning, and physiological factors; as well as ultimate explanations for power play and pain play dimensions of BDSM, highlighting the potential adaptive advantages of each. While BDSM may not be adaptive in itself, we examine the literature of sex differences in BDSM role preferences and argue that these preferences may stem from the extreme forms of behaviors which enhance reproductive success. In the realm of pain play, we explore the intersection of pain and pleasure from both physiological and psychological perspectives, highlighting the crucial role of psychological and play partner factors in modulating the experience of pain. Finally, we encourage future research in social sciences to utilize evolutionary frameworks to further explore the subject and help alleviate the mystification surrounding BDSM. This multifaceted exploration of BDSM provides valuable insights for clinicians, kink-identified individuals, and scholars seeking to understand the evolutionary perspectives of human sexual behavior and preferences.

Attitudes towards Technology: Insights on Rarely Discussed Influences on Older Adults’ Willingness to Adopt Active Assisted Living (AAL)

Background: European research policy promotes active assisted living (AAL) to alleviate costs and reach new markets. The main argument for massive investments in AAL is its potential to raise older adults’ Quality of Life and enhance their freedom, autonomy, mobility, social integration, and communication. However, AAL is less widely spread in older adults’ households than expected. Research Aim: We investigate how the variable “technology acceptance” is connected to socio-economic-, social, health, “personal attitude towards ageing”, and “Quality of life” variables. Method: We conducted a study in Vienna between 2018 and 2020, questioning 245 older adults (M = 74, SD = 6.654) living in private homes. We calculated multivariate models regressing technology acceptance on the various exploratory and confounding variables. Results: Experiencing an event that made the person perceive their age differently changed the attitude towards using an assistive technological system. Participants perceived technology that is directly associated with another human being (e.g., the use of technology to communicate with a physician) more positively. Conclusion: Older adults’ attitudes towards technology may change throughout their lives. Using major events in life as potential entry points for technology requires awareness to avoid reducing the lives of older adults to these events. Secondly, a certain human preference for “human technology” may facilitate abuse if technology is given a white coat, two eyes, a nose, and a mouth that may falsely be associated with a natural person. This aspect raises the ethical issue of accurate information as a significant precondition for informed consent.

Foraging in a non-foraging task: Fitness maximization explains human risk preference dynamics under changing environment.

Changes in risk preference have been reported when making a series of independent risky choices or non-foraging economic decisions. Behavioral economics has put forward various explanations for specific changes in risk preference in non-foraging tasks, but a consensus regarding the general principle underlying these effects has not been reached. In contrast, recent studies have investigated human economic risky choices using tasks adapted from foraging theory, which require consideration of past choices and future opportunities to make optimal decisions. In these foraging tasks, human economic risky choices are explained by the ethological principle of fitness maximization, which naturally leads to dynamic risk preference. Here, we conducted two online experiments to investigate whether the principle of fitness maximization can explain risk preference dynamics in a non-foraging task. Participants were asked to make a series of independent risky economic decisions while the environmental richness changed. We found that participants' risk preferences were influenced by the current and past environments, making them more risk-averse during and after the rich environment compared to the poor environment. These changes in risk preference align with fitness maximization. Our findings suggest that the ethological principle of fitness maximization might serve as a generalizable principle for explaining dynamic preferences, including risk preference, in human economic decision-making.

Neural Diversity and Decisions

PurposeDifferences in choices provide a window into the variety of human preferences and behaviors. Most non-trivial decisions recruit multiple regions of the brain with activity that shows substantial variation across individuals and also depends on personality traits and attitudes.MethodsThe present study measured nine physiologic factors and collected 15 types of psychological data while participants made decisions in standard tasks measuring risk aversion, patience, altruism, cooperation, generosity, trust, and trustworthiness. Robust predictors were defined as those that were statistically significant in both forward and backward stepwise regressions using all the collected independent variables.ResultsThe analysis showed that multiple measures of physiologic stress decrease cooperative behaviors while testosterone consistently increased antisocial behaviors. In addition, cognitive abilities robustly increased patience but decreased cooperation, while those more satisfied with their lives were more trustworthy. Participants scoring high on the personality trait of imagination were shown to be less altruistic and generous, while neuroticism increased prosociality.ConclusionThis study identified key drivers of behaviors that should be measured in decision-making experiments in order to capture the multiple factors that affect choices. We conclude with suggestions on how to establish causal relationships between the identified factors and decisions.

Reinforcement Learning from AI Feedback A Review

Reinforcement Learning from AI Feedback (RLAIF) is a big step forward compared to Reinforcement Learning from Human Feedback (RLHF). It's especially useful for large language models like GPT-4. RLAIF is better because it can handle more data at scale and is more efficient. It uses AI-generated feedback instead of human feedback. This shift to AI-generated feedback enhances the efficiency and speed of training AI systems. Additionally, RLAIF optimizes the AI's ability to align with desired outcomes, although it may not directly improve understanding human preferences. RLAIF uses a Preference Model (PM) that follows constitutional principles. This ensures that AI responses are ethical, safe, and high -quality. The constitution sets rules for AI decision-making. It makes sure AI follows ethical and social standards. This is important as AI keeps evolving. RLAIF is moving towards an automated, moral feedback system focusing on responsible AI governance and ethical guidelines.

Architectural design, visual attention, and human cognition: exploring responses to federal building styles

ABSTRACT This article examines the ongoing architectural debate surrounding styles for American government buildings. It tests the results of the 2020 National Civic Art Society survey that found 72% of Americans prefer traditional federal buildings over modern designs. To gain a neuro-biological perspective on these results, eye-tracking emulation software is used to evaluate the relationship between architectural design, visual attention, and human preferences. The hypothesis that traditional buildings would capture and retain pre-attentive visual processing more than modern styles was confirmed in several experiments. The research offers valuable insights for architects, urban planners, and policymakers considering design decisions about federal buildings.

Advancing cross-subject olfactory EEG recognition: A novel framework for collaborative multimodal learning between human-machine

Odor sensory evaluation is broadly applied in food, clothing, cosmetics, and other fields. Traditional artificial sensory evaluation has poor repeatability, and the machine olfaction represented by the electronic nose (E-nose) is difficult to reflect human feelings. Olfactory electroencephalogram (EEG) contains features associated with human olfactory preference, which has unique advantages in odor sensory evaluation. However, the difficulty of cross-subject olfactory EEG recognition dramatically limits its application. In this paper, a human–machine collaborative multimodal learning method is proposed for cross-subject olfactory preference recognition. Firstly, the olfactory EEG and E-nose multimodal data acquisition and preprocessing paradigms are established. Secondly, a complementary multimodal data mining strategy is proposed to effectively mine the individual and common features representing subjects' olfactory preferences from multimodal data. Finally, the cross-subject olfactory preference recognition is achieved in 24 subjects by fusing the extracted common and individual features, and the recognition effect is superior to the state-of-the-art recognition methods. Furthermore, the advantages of the proposed method in cross-subject olfactory preference recognition indicate its potential for practical odor evaluation applications.

Fast and Comfortable Robot-to-Human Handover for Mobile Cooperation Robot System.

In this work, we present a method that enables a mobile robot to hand over objects to humans efficiently and safely by combining mobile navigation with visual perception. Our robotic system can map its environment in real time and locate objects to pick up. It uses advanced algorithms to grasp objects in a way that suits human preference and employs path planning and obstacle avoidance to navigate back to the human user. The robot adjusts its movements during handover by analyzing the human's posture and movements through visual sensors, ensuring a smooth and collision-free handover. Tests of our system show that it can successfully hand over various objects to humans and adapt to changes in the human's hand position, highlighting improvements in safety and versatility for robotic handovers.

Implementation on Large-Scale Mobile Fitness App Usage Analysis for Smart Health

Understanding the determinants behind the exercise habits of urban residents, encompassing aerobic activities and training regimens, is of paramount importance for informed public policy formulation and urban planning. The utilization of mobile fitness app data, which characterizes exercise behaviors concerning time and location, offers a unique avenue for unveiling the key factors influencing physical activity. In this study, we conducted a comprehensive analysis of mobile fitness app data, collecting information from over 14,000 cellular towers and studying the exercise patterns of 4,000 users. Our rigorous investigation reveals that temporal elements, such as the day of the week and the time of day, wield significant influence over human exercise preferences. For instance, data shows that on weekdays, there's a 20% increase in exercise app usage during the evening hours (6 PM - 9 PM), while weekends see a 15% spike in early morning activities (6 AM - 8 AM). Geographical location also emerges as a prominent influencer. Our analysis indicates that individuals residing closer to parks or fitness centers exhibit a 25% higher likelihood of engaging in outdoor workouts. Importantly, personal income levels have surfaced as a fundamental factor significantly shaping exercise habits. Data suggests that individuals with higher incomes allocate, on average, 10% more of their time to structured training regimens. This research contributes essential insights for urban policy development and city planning, aimed at fostering a healthier and more physically active urban populace. Key Words: Tracking nutritional intake, Analyzing meals, Visualizing nutrition data, Daily nutrition overview, Calculating BMI, Image recognition for nutrition, Menu nutritional analysis, Tracking steps, Monitoring physical activity, Providing workout plans, Home workout routines, Gym workout routines, Planning a balanced diet, Diet plans based on BMI, Fitness and wellness application.

Changes in prehistoric wood procurement strategies in Northern China from 6500 – 2000 BP. Evidence for human preferential harvesting in the face of climatic change

In the past, humans have adapted their strategies of wood collection to shifts in the surrounding environment. Shifts in wood procurement strategies have been assumed to have been heavily influenced by changes in biomes due to shifts in climate. In these models, wood found on archaeological sites is interpreted as being representative of the surrounding environment. Around the world there is an increasing recognition that human preferences and management strategies impacted forest composition and species harvesting. We evaluated charcoal data gathered from 92 archaeological sites in northern China, with chronologies spanning the period from 6500 BP to 2000 BP. By combining the results of the analysis with published paleoenvironment records and archaeological evidence, we aimed to determine when and how there was an anthropogenic impact on wood procurement strategies. Our results indicated that a distinct spatiotemporal pattern of wood utilisation existed in northern China during the period 6500–2000 BP that could not be fully explained by changes in climate. Spatially, anthropogenically-collected woody assemblages showed remarkable differences in northern China. Northern China's climate gradually became colder and drier during the mid to late Holocene, especially during the period 4000–2000 BP. This appears to have led to a decreased use of Quercus L. in the Loess Plateau and the Guanzhong Basin. However, in the Central Plains and Shandong Peninsula, Quercus L. usage continued and increased in the late Holocene. During the same period, a striking shift in wood procurement occurred in the western part of China. In the northeastern Tibetan Plateau, wood procurement shifted from predominantly coniferous to broadleaved trees, and forest succession caused by anthropogenic collection played the significant role. Paleoclimatic information in Hexi Corridor suggests that some arbour trees in archaeological sites were not locally available, suggesting that human preference for certain wood types led them to engage in long distance transport. This finding indicates that starting after 4000 BP, climatically driven shifts in biomes were no longer a primarily influencing variable in human procurement of trees in north-western and eastern China. Key regional differences in strategy emerged. In Eastern China, wood species which humans preferred were still harvested even when climatic change reduced their availability. In North-western China, however, changes in biomes following 4000 BP meant that people needed to change the species on which they had previously relied and began to engage in long distance transport of wood.

Towards practical robotic chef: Review of relevant work and future challenges

AbstractRobotic chefs are a promising technology that can improve the availability of quality food by reducing the time required for cooking, therefore decreasing food's overall cost. This paper clarifies and structures design and benchmarking rules in this new area of research, and provides a comprehensive review of technologies suitable for the construction of cooking robots. The diner is an ultimate judge of the cooking outcome, therefore we put focus on explaining human food preferences and perception of taste and ways to use them for control. Mechanical design of robotic chefs at a practically low cost remains the challenge, but some recently published gripper designs as well as whole robotic systems show the use of cheap materials or off‐the‐shelf components. Moreover, technologies like taste sensing, machine learning, and computer vision are making their way into robotic cooking enabling smart sensing and therefore improving controllability and autonomy. Furthermore, objective assessment of taste and food palatability is a challenge even for trained humans, therefore the paper provides a list of procedures for benchmarking the robot's tasting and cooking abilities. The paper is written from the point of view of a researcher or engineer building a practical robotic system, therefore there is a strong priority for solutions and technologies that are proven, robust and self‐contained enough to be a part of a larger system.

Seeing Like a Model Fish: How Digital Extractions Mediate Metabolic Relations

Digital models have become key sites of biological practice and science policy. This paper examines efforts to craft a digital salmon model for metabolic research. It traces the data configurations of feed, food, and health in Norway’s bioeconomy aspirations—from cell culture studies in the lab to an integrated digital repository and public health nutrition trials in schools. A response to policy calls for more sustainable aquaculture, digital models become lynchpins for intensified data sourcing. I describe how datafication and digital model integration enact a particular mode of management, focused on profitability and human preferences, when optimizing fish feed to sustainability goals. Integrating molecular biology and mathematical functions, digital modeling promotes the idea of a prediction machine for preemptive optimization of feed and food across settings. Yet, unforeseen disruptions to aquaculture, for instance, by algae and lice, expose the complexity of marine food webs and the limitations of digital models. Even while in the making, the digital fish model becomes performative and shapes knowledge practices much beyond the lab. As knowledge infrastructures, models participate in the remaking of metabolic relations, recalibrating decision-making, while feeding back into and co-shaping the very entities and environments they were crafted to investigate.

Learning Optimal Advantage from Preferences and Mistaking It for Reward

We consider algorithms for learning reward functions from human preferences over pairs of trajectory segments, as used in reinforcement learning from human feedback (RLHF). Most recent work assumes that human preferences are generated based only upon the reward accrued within those segments, or their partial return. Recent work casts doubt on the validity of this assumption, proposing an alternative preference model based upon regret. We investigate the consequences of assuming preferences are based upon partial return when they actually arise from regret. We argue that the learned function is an approximation of the optimal advantage function, not a reward function. We find that if a specific pitfall is addressed, this incorrect assumption is not particularly harmful, resulting in a highly shaped reward function. Nonetheless, this incorrect usage of the approximation of the optimal advantage function is less desirable than the appropriate and simpler approach of greedy maximization of it. From the perspective of the regret preference model, we also provide a clearer interpretation of fine tuning contemporary large language models with RLHF. This paper overall provides insight regarding why learning under the partial return preference model tends to work so well in practice, despite it conforming poorly to how humans give preferences.