阳狮集团

Deadline-driven digital tools typically assume a linear, schedule-bound interaction model, implicitly privileging monochronic workstyles while offering limited support for polychronic, interleaved task management strategies. This paper introduces Time by Design, a dual-mode scheduling interface that adapts task granularity, reminder elasticity, progress metaphors, and interruption handling to an individual’s Workstyle Orientation Index (WOI)—a short-form measure grounded in monochronic–polychronic time-orientation theory. The WOI captures preferences for sequencing, tolerance for interruption, and comfort with parallelism, enabling onboarding-time personalization. We present the design and implementation of two functional prototypes: Mono-mode, which emphasizes fixed-duration blocks, critical-path visibility, and focus-protective reminders; and Poly-mode, which supports parallel work cards, elastic reminder windows (±Δ), and mosaic-style progress views. A three-phase mixed-methods evaluation includes a survey and diary study (N = 120), a counterbalanced lab study (N = 64), and a 10-day field deployment (N = 40) comparing orientation-matched and mismatched modes. Preliminary pilot simulations and early qualitative probes suggest that orientation–mode matching may improve on-time completion by 12–18

The rapid deployment of generative AI (GenAI) chatbots across the financial, healthcare, and e-commerce sectors faces a critical obstacle, as users lack verifiable trust in these systems. The current solutions for explainability and safety filtering do not provide sufficient protection because they rely on untrustworthy logs that can be modified or hidden from disclosure. The research presents a trust-based conversational AI system that combines blockchain anchoring with GenAI reasoning systems to create dialogues that are transparent, tamper-proof, and fully auditable. The Verifiable Trust Score (VTS) serves as the main component that evaluates the auditability, transparency, and privacy protection levels of each dialogue segment. The system uses a minimal blockchain anchoring method to store conversation summaries and trust information on the blockchain while keeping confidential data stored outside the blockchain network. The simulation tests conducted in customer service and healthcare environments demonstrate that our system maintains a 95

The rapid scaling of enterprise chatbots across cloud and edge environments has amplified the carbon footprint and operational cost of inference workloads. This study introduces AnyScale Inference, a novel framework that optimizes large-language-model (LLM) deployment through carbon- and cost-aware geo-shifting combined with heterogeneous accelerator scheduling. Unlike prior solutions that statically bind inference to fixed regions or hardware tiers, AnyScale Inference dynamically orchestrates model execution across multi-region data centers and diverse accelerators—CPUs, GPUs, and TPUs—based on real-time telemetry of carbon intensity, energy price, and latency budgets. A hierarchical scheduler leverages reinforcement learning to balance three competing objectives: sustainability, service-level objectives (SLOs), and total cost of ownership (TCO). Experimental results on enterprise-scale chatbot benchmarks demonstrate up to 42

Retrieval-Augmented Generation (RAG) has emerged as the dominant architecture for grounding large language models with factual and context-aware information. Yet, little is known about its end-to-end performance behavior under load—particularly how retriever latency, token generation rate, and concurrency interact to determine overall throughput and user-perceived responsiveness. This study presents the first unified framework to our knowledge for analytical and empirical modeling of RAG pipelines under variable load conditions. This study develops a queue-coupled (queuing analysis–based) performance model that captures cross-tier dependencies between retrieval and generation, integrating parameters such as retrieval depth, document encoding cost, and decoder token rate. Complementary large-scale experiments on cloud-deployed RAG instances validate the model across multiple workloads and retriever–generator pairings. Results reveal previously unreported latency coupling effects, where minor retrieval slowdowns propagate non-linearly to generation throughput, leading to underutilized GPU compute. This work further identifies scaling laws that predict degradation thresholds and proposes lightweight scheduler adaptations that improve throughput by up to 27

According to Dutch philosopher Spinoza (1631-1677), the body is a power of acting. This force of existence can be affected (checked or stimulated) by the mechanisms of subjection that discipline the body, with obvious consequences for our development and well-being. These were the questions for inquiry: what importance should be attributed to Spinoza in the ecology of existential knowledge? What kind of body experience does he advocate? What lessons for education can be drawn from his thinking? In this sense, we highlight as an axial objective: To discuss Spinoza’s importance in the idea of body as ecology of knowledge. As sources we used the published works of Spinoza. To achieve our goals, we adopted a methodological strategy that is enacted in an approximation between phenomenology and historical hermeneutics (in the wake of Husserl, Heidegger and Gadamer). It is a question of inquiring how the body is revealed to consciousness (disconcealment), because it plays an important role in the production of truth. As a conclusion we will say that Spinoza knew how to distance himself from the legacy of the process of body mortification and he also knew how to position himself critically vis-à-vis the ideas of modernity. He helps us find a more ecological type of education so we can establish healthier relationships.