Wei Wang

With the rapid expansion of parameters in large language models (LLMs) and the application of Reinforcement Learning with Human Feedback (RLHF), there has been a noticeable growth in the moral competence of LLMs. However, several questions warrant further exploration: Is it really possible for LLMs to fully align with human values through RLHF? How can the current moral growth be philosophically contextualized? We identify similarities between LLMs’ moral growth and Deweyan ethics in terms of the discourse of human moral development. We then attempt to use Dewey’s theory on an experimental basis to examine and further explain the extent to which the current alignment pathway enables the development of LLMs. A beating experiment serves as the foundational case for analyzing LLMs’ moral competence across various parameters and stages, including basic moral cognition, moral dilemma judgment, and moral behavior. The results demonstrate that the moral competence of the GPT series has seen a significant improvement, and Dewey’s Impulse-Habit-Character theory of moral development can be used to explain this: the moral competence of LLMs has been enhanced through experience-based learning, supported by human feedback. Nevertheless, LLMs’ moral development through RLHF remains constrained and does not reach the character stage described by Dewey, possibly due to their lack of self-consciousness. This fundamental difference between humans and LLMs underscores both the limitations of LLMs’ moral growth and the challenges of applying RLHF for AI alignment. It also emphasizes the need for external societal governance and legal regulation.

With the rapid expansion of parameters in large language models (LLMs) and the application of Reinforcement Learning with Human Feedback (RLHF), there has been a noticeable growth in the moral competence of LLMs. However, several questions warrant further exploration: Is it really possible for LLMs to fully align with human values through RLHF? How can the current moral growth be philosophically contextualized? We identify similarities between LLMs’ moral growth and Deweyan ethics in terms of the discourse of human moral development. We then attempt to use Dewey’s theory on an experimental basis to examine and further explain the extent to which the current alignment pathway enables the development of LLMs. A beating experiment serves as the foundational case for analyzing LLMs’ moral competence across various parameters and stages, including basic moral cognition, moral dilemma judgment, and moral behavior. The results demonstrate that the moral competence of the GPT series has seen a significant improvement, and Dewey’s Impulse-Habit-Character theory of moral development can be used to explain this: the moral competence of LLMs has been enhanced through experience-based learning, supported by human feedback. Nevertheless, LLMs’ moral development through RLHF remains constrained and does not reach the character stage described by Dewey, possibly due to their lack of self-consciousness. This fundamental difference between humans and LLMs underscores both the limitations of LLMs’ moral growth and the challenges of applying RLHF for AI alignment. It also emphasizes the need for external societal governance and legal regulation.

Previous studies revealed that the susceptibility, contacting preference, and recovery probability markedly alter the epidemic outbreak size and threshold. The recovery probability of an infected node is closely related to its obtained resources. How to allocate limited resources to infected neighbors is extremely important for containing the epidemic spreading on complex networks. In this paper, we proposed an epidemic spreading model on complex networks, in which we assume that the node has heterogeneous susceptibility and contacting preference, and susceptible nodes are willing to share their resources to neighbors. Through a developed heterogeneous mean-field theory and a large number of numerical simulations, we find that the recovered nodes provide resources uniformly to their infected neighbor nodes, and the epidemic spreading can be suppressed optimally on homogeneous and heterogeneous networks. Besides, altering the susceptibility and contacting preference does not qualitatively change the results. The susceptibility of the node decreases, which makes the outbreak threshold of epidemic spreading increase, and the outbreak size decreases. Our theory agrees well with the numerical simulations.

As the essence of livestream e-commerce is social commerce, building a livestream brand community and attracting brand followers are the key aspects to achieving sustained revenue. For many companies, inviting celebrities has become a shortcut to attract new followers. Considering the unsustainability and high cost of the celebrity host mode, some companies switched to using their own branded broadcasters to attract followers. However, as branded broadcasters lack a fan base, choosing the suitable broadcaster type has become a challenge in livestream e-commerce. The motivation of consumers to follow brand livestream accounts is mainly to obtain potential value by embedding them in social networks. Therefore, based on motive theory, this research explores how different broadcaster types affect consumer’s intention to follow a livestream brand community. Results from the analysis of secondary data from livestream platforms and two laboratory experiments reveal that celebrities contribute more to consumer’s intention to follow than branded broadcasters, and utilitarian products can strengthen the effect of branded broadcasters on attracting potential followers. Moreover, branded broadcasters can promote consumer’s intention to follow a livestream brand community by satisfying consumer’s need for informational value during utilitarian product evaluation. This research analyzes the differential effects of different types of broadcasters on livestream brand community building. The findings can deepen the understanding of the consumer’s behavior of following brand livestream communities and provide companies with suggestions on broadcaster selection in livestream e-commerce.

Philosophical regress arguments typically arise when an explanation is taken to demand another explanation of the same form, generating the appearance of an unbounded sequence. Standard responses—foundational, coherentist, infinitist, or deflationary—attempt to manage such sequences, but they all presuppose that the regress is in principle formulable. This paper challenges that assumption. I argue that many classical regresses arise only when a relation defined for items within a domain is illicitly applied to the domain as a whole. The mistake is structural rather than epistemic: the whole is not an element of the domain governed by the relevant explanatory relation. Once this level-shift and domain-closure constraint is made explicit, certain well-known regresses turn out not to require a stopping point; they fail to start. A sequence of the form X \rightarrow R(X) \rightarrow R^2(X) cannot extend to the whole, because R is not defined at that level. Attempts to initiate such regresses rest on a category error. The resulting framework dissolves a wide range of regress formulations in metaphysics, epistemology, philosophy of science, and philosophy of mind. Rather than providing another principle for terminating a chain, the paper identifies a structural condition under which the chain is not merely vicious but impossible to formulate.

Individuals with high risk of internet gaming disorder showed abnormal psychological performances in response inhibition, impulse control, and emotion regulation, and are considered the high-risk stage of internet gaming disorder. The identification of this population mainly relies on clinical scales, which are less accurate. This study aimed to explore whether these performances have highly accurate for discriminating HIGD from low-risk ones. Eye tracking based anti-saccade task, Barratt impulsiveness scale, and Wong and Law emotional intelligence scale were used to evaluate psychological performances in 57 individuals with HIGD and 52 matched low risk of internet gaming disorder. HIGD group showed significantly increased BIS total, attention, motor, and non-planning scores, but significantly decreased WLEIS emotion regulation score and correct rate of eye tracking anti-saccade task compared with LIGD group. BIS total score was negatively correlated with the WLEIS total and WLEIS emotion regulation scores. A combination of the WLEIS emotion regulation score and the correct rate of anti-saccade task could discriminate HIGD from LIGD with 91.23% sensitivity, 82.69% specificity, and 87.16% accuracy. Participants with higher gaming hours daily were 40 times more likely to be high risk than their counterparts. Hence, psychological performances were worse in HIGD. A combination of abnormal emotion regulation and response inhibition might be a potential marker to identify HIGD individuals.

This LNCS book is part of the FOLLI book series and constitutes the proceedings of the 10th International Conference on Logic, Rationality, and Interaction, LORI 2025, held in Xi’an, China, in October 2025. The 14 full papers presented in this book were carefully reviewed and selected from 47 submissions.The topics addressed in this program effectively showcase the breadth and depth characteristic of the LORI conference series, including contributions on dynamic epistemic logic, nonmonotonic reasoning, belief revision theory, decision theory, causal inference, social epistemology and so on.

A scientific paper’s citation represents its influence, which is the most intuitive indicator to access the quality of papers. This paper mainly adopts the social network analysis method, using the authors and the keywords of sports industry papers in China to constitute the networks of collaboration and knowledge, to explore effects of the degree centrality of authors and keywords and the structural hole of authors and keywords on the citation of papers in the collaboration and knowledge networks and draw the following conclusions: as for collaboration networks, the degree centrality at the paper level is positively correlated with citations; in the collaboration network, the positive correlation between the structural hole at the paper level and citations does not exist; within knowledge networks, an inverted- U shape was found between degree centrality and paper’s citation; and within knowledge networks, a positive correlation is in existence between the structural hole of papers and their citation. This study synthesizes the already widely used collaboration network with the knowledge network constructed through keywords, distinguishes from the previous network features focusing at the author level, and explores research projects of Chinese Sports Industry from the paper level, providing a new perspective for the research of sports industry in China, complementing the methods and ideas of sports industry research, as well as providing a reference for the research in other disciplinary fields.

Due to the complexity of the network structure, log analysis is usually necessary for the maintenance of network-based distributed systems since logs record rich information about the system behaviors. In recent years, numerous works have been proposed for log analysis; however, they ignore temporal relationships between logs. In this paper, we target on the problem of mining informative patterns from temporal log data. We propose an approach to discover sequential patterns from event sequences with temporal regularities. Discovered patterns are useful for engineers to understand the behaviors of a network-based distributed system. To solve the well-known problem of pattern explosion, we resort to the minimum description length principle and take a step forward in summarizing the temporal relationships between adjacent events of a pattern. Experiments on real log datasets prove the efficiency and effectiveness of our method.

The book is a translation of the second edition of a much-used and research-based Chinese textbook. As a succinct and issue-based introduction to the Western philosophy of science, the book brings eight focal issues in the field to the fore and augments each topic by incorporating Chinese perspectives. Followed by an overview of the historical framework and logical underpinnings of the philosophy of science, the book thoroughly discusses eight issues in the discipline: (1) the criteria of cognitive meaning, (2) induction and confirmation, (3) scientific explanation, (4) theories of scientific growth, (5) the demarcation between science and pseudoscience, (6) scientific realism and empiricism; (7) the philosophy of scientific experimentation, (8) science and value. Not confined to Western mainstream discourse in this field, the book also introduces voices of Chinese philosophers of note and adopts a stance that productively combines logical empiricism and Kuhnianism, both of which tend to be covered in less detail by many English language textbooks. In the final chapter the author offers a prognosis regarding the future of the discipline based on recent trends. This book will be of value to students who study philosophy of science and hope to gain a better understanding of science and technology.

Scientific explanation, laws of nature and causation are crucial and frontier issues in the philosophy of science. This book studies the complex relationship between the three concepts, aiming to achieve a holistic synthesis about explanation–laws–causation. By reviewing Hempel's Scientific Explanation models and Salmon's three conceptions – the epistemic, modal and ontic conception – the book suggests that laws are essential to explanation and our understanding of laws will help solve the problems of the latter. Concerning the nature of laws, this book tackles both the problems of regularity approach and necessitarian approach. It also proposes that the ontological order of explanation should be from events (or processes) to causation, then to regularity (laws), and finally to science system, but the epistemological order should be from science system to laws to explanation and causation. In addition, this book examines the legitimacy of Ceteris Paribus laws, the connection between explanation and reduction, the relation between explanation and interpretation, and some other issues closely related to explanation–laws–causation. This book will attract scholars and students of philosophy of science, philosophy of nature, cognitive science, etc.

Scientific explanation, laws of nature and causation are crucial and frontier issues in the philosophy of science. This book studies the complex relationship between the three concepts, aiming to achieve a holistic synthesis about explanation–laws–causation. By reviewing Hempel's Scientific Explanation models and Salmon's three conceptions – the epistemic, modal and ontic conception – the book suggests that laws are essential to explanation and our understanding of laws will help solve the problems of the latter. Concerning the nature of laws, this book tackles both the problems of regularity approach and necessitarian approach. It also proposes that the ontological order of explanation should be from events (or processes) to causation, then to regularity (laws), and finally to science system, but the epistemological order should be from science system to laws to explanation and causation. In addition, this book examines the legitimacy of Ceteris Paribus laws, the connection between explanation and reduction, the relation between explanation and interpretation, and some other issues closely related to explanation–laws–causation. This book will attract scholars and students of philosophy of science, philosophy of nature, cognitive science, etc.

S'appuyant sur des témoignages de professionnels de santé, cet ouvrage permet de mettre en exergue le glissement de la médecine du thérapeutique vers l'amélioration de l'Homme. Il a pour objectif d'engager des réflexions pluridisciplinaires et internationales sur les évolutions biologiques et technoscientifiques susceptibles de bouleverser la condition humaine. Il s'agit notamment de déterminer si la médecine améliorative doit être autorisée, limitée ou prohibée, l'enjeu étant de parvenir à préserver l'humanité de l'Homme."--Page 4 of cover.

Profound sensorineural hearing loss is an auditory disability associated with auditory and cognitive dysfunction. Due to distinct pathogenesis, some associated structural and functional changes within the brain have been investigated in previous studies, but whole-brain structural alterations are incompletely understood. We extended the exploration of neuroanatomic differences in whole-brain structure in children with profound SNHL who are primarily users of Chinese sign language. We employed surface-based morphometry and subcortical analyses. T1-weighted magnetic resonance images of 26 children with profound SNHL and 27 age- and sex-matched children with normal hearing were analyzed. Compared with the normal control group, children with profound SNHL showed diverse structural changes in surface-based and subcortical analyses, including decreased cortical thickness in the left postcentral gyrus, superior parietal lobule, paracentral lobule, precuneus, the right transverse temporal gyri, and the middle temporal gyrus; a noticeable increase in the Local Gyrification Index in the left precuneus and superior parietal lobule; and diverse changes in gray-matter volume in different brain regions. Surface-based vertex analyses revealed regional contractions in the right thalamus, putamen, pallidum, and the brainstem of children with profound SNHL when compared with those in the NC group. Volumetric analyses showed decreased volumes of the right thalamus and pallidum in children with profound SNHL. Our data suggest that children with profound SNHL are associated with diffuse cerebral dysfunction to cortical and subcortical nuclei, and revealed neuroplastic reorganization in the precuneus, superior parietal lobule, and temporal gyrus. Our study provides robust evidence for changes in connectivity and structure in the brain associated with hearing loss.

The debate between the dynamical and the statistical interpretations of natural selection is centred on the question of whether all explanations that employ the concepts of natural selection and drift are reducible to causal explanations. The proponents of the statistical interpretation answer negatively, but insist on the fact that selection/drift arguments are explanatory. However, they remain unclear on where the explanatory power comes from. The proponents of the dynamical interpretation answer positively and try to reduce selection/drift arguments to some of the most prominent accounts of causal explanation. In turn, they face the criticism raised by statisticalists that current accounts of causation have to be violated in some of their core conditions or otherwise used in a very loose manner in order to account for selection/drift explanations. We propose a reconciliation of both interpretations by conveying evolutionary explanations within the unificationist model of scientific explanation. Therefore, we argue that the explanatory power in natural selection arguments is a result of successful unification of individual- and population-level facts. A short case study based on research on sympatric speciation will be presented as an example of how population- and individual-level facts are unified to explain the morphological mosaic of bill shape in island scrub jays.

Governing greenwashing in environmental, social, and governance (ESG) disclosure is an important issue, but relevant literature is scant. Based on the data on Chinese A-share listed firms from 2012 to 2021, we investigate the governance role of corporate digital transformation (DT) in ESG greenwashing and its influencing mechanism. We find that DT significantly inhibits ESG greenwashing. Moreover, DT mitigates ESG greenwashing by enhancing corporate green technology innovation (i.e., innovation channel), reducing information asymmetry (i.e., information channel), and increasing trade credit (i.e., resource channel). From the perspective of investor attention, we find that both retail investors' online opinions and institutional investors' site visits strengthen the inhibitory effect of DT on ESG greenwashing. Furthermore, such an inhibitory effect is more pronounced in large firms, industries with high competition, and regions with strong intellectual property protection. This research provides new insights and policy suggestions for governing corporate ESG greenwashing behavior.

Literature from the United Kingdom and the United States represents the cultural expression of those peoples' lived experiences. Reading British and American literature may also aid in our understanding of the values, worldview, and ideological underpinnings of western civilization. Therefore, this thesis examines the mythological and religious themes in British and American literature using literary works from both countries. Greek Myth is the source and soil of ancient Greek literature. Ancient Greek and Roman literature is a rich treasure for later generations to understand myths, providing creative materials for later British and American writers. It can be said that myths and religious ideas gave birth to great literary works and some immortal poets. Among them, "_Wuthering Heights_" is the main representative work of 19 th -century literature and occupies an important position in British and American literary history. Based on this, this paper discusses British and American literature's religious and cultural characteristics in "_Wuthering Heights_", mainly from the perspective of environment description and plot shaping. I hope this research can play a role in attracting jade.

Research into innovative behavior is not new, but its importance for organizational effectiveness has become even more evident in recent years. However, the psychological processes and underlying mechanism concerning how and why innovative behavior occurs within an organization still invite more investigation. The present study considers ambidextrous organizational culture as a pro-innovation culture and proposes that it can be perceived by employees, which leads to their innovative behavior. This study adds clarity by exploring the impact of perceived ambidextrous organizational culture on employees’ reactions related to innovation via the intermediate mechanism of psychological empowerment and the moderating condition of transformational leadership. Hypotheses are derived from a motivational perspective based on self-determination theory. Results are based on data collected from 647 Chinese employee-supervisor dyads. This study finds that employees’ perceptions of ambidextrous organizational culture have an indirect effect on innovative behavior through psychological empowerment. Specifically, the positive indirect relationship is amplified when transformational leadership is at a higher level. Our findings show how the mediating mechanism of psychological empowerment and the moderating condition of transformational leadership work together to improve innovation by individuals. The findings reveal several ways in which organizations can strategically focus on their cultural and supervisory training, such as applying this model to improve employees’ outcome related to innovation.