Differentiable programming has emerged as an important tool for gradient-based inverse problems and end-to-end hybrid physics–machine learning in computational fluid dynamics. Existing differentiable CFD platforms predominantly rely on structured Cartesian grids, limiting their direct application to geometrically complex domains where body-conforming unstructured discretizations are essential. We present DiFVM, the first GPU-native, end-to-end differentiable finite-volume (FV) CFD solver for incompressible flow on unstructured polyhedral meshes. The central methodological contribution is a graph-based computational reformulation of the complete unstructured FV/PISO solution procedure: by expressing FV operators as static scatter/gather primitives on the mesh connectivity graph, DiFVM transforms irregular unstructured connectivity into a GPU-compatible computational representation. All operations are implemented in JAX/XLA, providing just-in-time compilation, operator fusion, and automatic differentiation through the complete simulation pipeline. Differentiable Windkessel outlet boundary conditions are provided for cardiovascular applications, and DiFVM accepts standard OpenFOAM case directories without modification for seamless adoption in existing workflows. Forward accuracy is assessed through analytical and canonical benchmark problems, comparisons with published reference data, and independent solver comparisons, while end-to-end differentiability is demonstrated through proof-of-concept inverse modeling examples. DiFVM provides a differentiable CFD framework for gradient-based optimization, inverse modeling and future physics-integrated machine learning on complex unstructured geometries.
We derive the irreversibility condition in fracture for the inverse-deformation approach using the second law of thermodynamics. We consider the problem of brittle failure in an elastic bar previously solved in (Rosakis et al 2021). Despite the presence of a non-zero interfacial/surface energy, the third derivative of the inverse-deformation map is discontinuous at the crack faces. This is due to the presence of the inequality constraint ensuring the inverse strain is nonnegative and the orientation of matter is preserved. A change in the material location of a crack results in negative entropy production, violating the second law. Consequently, such changes are disallowed giving the irreversibility condition. The inequality constraint and the irreversibility condition limit the space of admissible variations. We prove necessary and sufficient conditions for local stability that incorporate these restrictions. Their numerical implementation shows that all broken equilibria found in (Rosakis et al 2021) are locally stable.
Heterologous plant protein blending enhances nutritional and functional properties, yet the molecular mechanisms of amyloid fibril co-assembly remain underexplored. This study explores PSP-QP co-fibrillation to clarify the links between mixing ratio, structural evolution and physicochemical properties. Regarding nutritional properties, the composite proteins exhibited a significant synergistic effect on bioaccessibility. Notably, the PQ1:4 blend demonstrated superior performance, releasing 183.6% more free amino acids after in-vitro digestion compared to PSP alone. In terms of fibrillation potential, the co-assembly pathway was found to be highly ratiodependent. PSP-dominated systems successfully formed extensive and interconnected fibrillar networks, whereas QP-rich systems predominantly yielded compact, disordered aggregates. This morphological divergence was confirmed through multiscale characterization; after 24 h of acid-heat treatment, mature, well-defined fibrils were observed in PSP and PQ1:1 systems, while high QP content led to amorphous or fragmented structures. Spectroscopic analysis revealed that this fibrillation process is driven by a conformational transition from random coils to beta-sheets, a transformation significantly modulated by the rigid structure of QP. Furthermore, molecular dynamics simulations unveiled the atomistic basis for these interactions. Pumpkin 11S globulin forms a highly stable complex with quinoa 11SB globulin (Delta G = -132.5 kcal/mol), driven by hydrophobic and van der Waals forces. This stable heterodimer serves as a critical nucleus for ordered fibril growth. Conversely, its interaction with quinoa 2S albumin proved unstable, leading to dissociation and amorphous aggregation. This study elucidates the regulatory mechanism of PSP-QP co-fibrillation, supporting the controllable construction and high-value plant-based food applications of heterologous plant protein amyloid fibrils.
Bibliometric analysis is a powerful means of evaluating research. This article employs a sequential mixed-methods approach to enrich the understanding of bibliometric analysis and presents a structured guide to maximize the technique’s benefits. A thematically synthesized narrative commentary on bibliometric review papers, based on input from 53 senior editorial members of leading hospitality and tourism journals, constitutes the study’s foundation. Cutting-edge literature comprising 195 bibliometric investigations is also considered. Our evidence-based guidance provides scholars with a roadmap for conducting state-of-the-art bibliometric analyses. These suggestions are intended to improve the method’s quality and accessibility, affirming its utility. Ultimately, this paper seeks to enhance bibliometric studies’ clarity, rigor, and efficacy by presenting forward-looking, real-world insights. Its step-by-step guide will facilitate decision-making and advance the field’s theoretical and managerial understanding.
Carbon reporting is critical for assessing climate commitments, yet current methods are time-intensive and vulnerable to inconsistencies. While Large Language Models (LLMs) offer revolutionary potential in carbon reporting, LLMs struggle with hallucination issues, factual delay, and data security. This study proposes SMARAG, a Trustworthy Artificial Intelligence (TAI)-powered carbon reporting system that combines TAI design principles and Self-decisive Multi-Agentic Retrieval-Augmented Generation (RAG) to provide accurate, efficient, and trustworthy carbon reporting. Grounded in the EU Ethics Guidelines for Trustworthy AI, SMARAG replicates real-world carbon reporting workflows through a team of role-specialized autonomous agents, such as sustainability director, project manager, data analyst, report writer. The agents draw from a locally hosted knowledge repository that integrates public carbon standards, regulatory frameworks, and private organizational data, minimizing the risk of unauthorized data exposure to external LLMs. Moreover, SMARAG employs both a self-decisive validation loop and human-in-the-loop architecture to ensure outputs meet regulatory standards before finalization. Evaluated on real-life carbon data and standards, SMARAG outperforms multi-agentic RAG, single-agent (vanilla) RAG, and standalone LLMs, including GPT-4o, DeepSeek R1, Claude-3.5, Gemini-1.5, and Mistral-Instruct-v0.3, in terms of report quality, usefulness, faithfulness, and relevancy. It not only empowers resource-constrained organizations such as small and medium-sized enterprises to participate in climate governance, but also offers valuable insights for policymakers and practitioners committed to broader trustworthy sustainability disclosures and knowledge-intensive regulatory domains.
