17837 research outputs found
Sort by
The Influence of Human-inspired Agentic Sophistication in LLM-driven Strategic Reasoners
Full text linkThe rapid rise of large language models (LLMs) has shifted artificial intelligence (AI) research toward agentic systems, motivating the use of weaker and more flexible notions of agency. However, this shift raises key questions about the extent to which LLM-based agents replicate human strategic reasoning, particularly in game-theoretic settings. In this context, we examine the role of agentic sophistication in shaping artificial reasoners' performance by evaluating three agent designs: a simple game-theoretic model, an unstructured LLM-as-agent model, and an LLM integrated into a traditional agentic framework. Using guessing games as a testbed, we benchmarked these agents against human participants across general reasoning patterns and individual role-based objectives. Furthermore, we introduced obfuscated game scenarios to assess agents' ability to generalise beyond training distributions. Our analysis, covering over 2000 reasoning samples across 25 agent configurations, shows that human-inspired cognitive structures can enhance LLM agents' alignment with human strategic behaviour. Still, the relationship between agentic design complexity and human-likeness is non-linear, highlighting a critical dependence on underlying LLM capabilities and suggesting limits to simple architectural augmentation
Protecting Public Interest in Sovereign and Municipal Restructurings
Full text linkThis paper explores the intersection of public interest and debt restructuring frameworks for sovereign and municipal debtors. The analysis underscores the need for legal systems to prioritize transparency, service continuity, and equitable burden-sharing among stakeholders, especially in the context of rising fiscal pressures and governance deficits
From Well-Being to Values or From Values to Well-Being?:Three Longitudinal Studies in the East and West on the Same Life Transition
No full textTheory and research on the relationships between values and well-being have predominantly focused on how values may affect well-being, with less attention to the reverse direction—how well-being may shape values. We present three alternative theoretical perspectives on this latter direction: well-being as an activator, reinforcer, and operator of value change. Using diverse measures of well-being and personal values, we conducted three longitudinal studies centered on the same life transition—entering university—across distinct samples from Mainland China (N = 218), Hong Kong (N = 252), and the United Kingdom (N = 196). Results from parallel process latent growth models (LGMs) showed more substantial evidence for well-being predicting later value change than the reverse. Specifically, well-being predicted later value changes in three cases: Initial life satisfaction predicted increases in the openness to change value dimension in Study 1, and initial self-esteem predicted increases in both openness to change and self-enhancement value dimensions in Study 3. Values predicted later well-being changes only once: Initial self-enhancement value dimension predicted increases in positive affect in Study 2. These results illuminate the intricate nature of the relationships between well-being and values, highlighting well-being’s role as a catalyst for value development and the importance of contextual factors during life transitions
Contrastive Translation With Dynamical Temperature for Sequential Recommendation
Full text linkContrastive learning is a promising solution to the problem of data sparsity in the field of recommendation system since it is able to extract self-supervised signals from raw data. The traditional contrastive learning-based sequential recommendation algorithms generate augmentations of original item sequences by utilizing crop, mask and reorder operations. However, those augmentation schemes destroy the underlying semantics of item sequences, resulting in difficulty in accurately defining positive and negative samples. To address this issue, we propose a contrastive translation based sequential recommendation algorithm, namely, CT4Rec. Specifically, CT4Rec generates augmented views of item sequences by injecting noises into embeddings of users and items, which is able to guarantee that the underlying semantics of augmented views are consistent with those of original item sequence. Hence, CT4Rec is able to effectively learn the invariances among the augmented views. In addition, the personalized translation operations are utilized to model the third-order relationships among entities. Moreover, it is difficult for contrastive learning-based recommendation algorithms with static temperature to simultaneously capture the differences among individual users/items and among the clusters of users/items. Hence, we utilize a dynamic temperature strategy to enhance CT4Rec, which endows CT4Rec with the capabilities of group-wise discrimination and instance discrimination. Our validation on five benchmark datasets shows that CT4Rec outperforms SOTA sequential recommendation methods. Our code is released at https://github.com/zar123123/CT4Rec
Too Many or Too Few? Information Cues in Recommender Systems and Consequences for Search and Purchase Behavior
Full text linkAn activities expansion of the transition polynomial of a multimatroid
Full text linkThe weighted transition polynomial of a multimatroid is a generalization of the Tutte polynomial. By defining the activity of a skew class with respect to a basis in a multimatroid, we obtain an activities expansion for the weighted transition polynomial. We also decompose the set of all transversals of a multimatroid as a union of subsets of transversals. Each term in the decomposition has the structure of a boolean lattice, and each transversal belongs to a number of terms depending only on the sizes of some of its skew classes. Further expressions for the transition polynomial of a multimatroid are obtained via an equivalence relation on its bases and by extending Kochol's theory of compatible sets.We apply our multimatroid results to obtain a result of Morse about the transition polynomial of a delta-matroid and get a partition of the boolean lattice of subsets of elements of a delta-matroid determined by the feasible sets. Finally, we describe how multimatroids arise from graphs embedded in surfaces and apply our results to obtain an activities expansion for the topological transition polynomial. Our work extends results for the Tutte polynomial of a matroid
Subsets of free groups with distinct differences
Full text linkLet F_n be a free group of rank n, with free generating set X. A subset D of F_n is a Distinct Difference Configuration if the differences g^{-1}h are distinct, where g and h range over all (ordered) pairs of distinct elements of D. The subset D has diameter at most d if these differences all have word length at most d. When n is fixed and d is large, the paper shows that the largest distinct difference configuration in F_n of diameter at most d has size approximately (2n-1)^{d/3}.<br/
Data sources:
Access Repository Dashboard
Do you manage Royal Holloway - Pure? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!