PINNs for Blood Flow Simulation|| ZENN for Heterogeneous Data-Driven Modeling || March 20, 2026

Value of the Information & Strength of the Argument

The seminar provides substantial value by presenting two novel computational frameworks with clear theoretical foundations and practical applications. The first talk offers a detailed methodology for using PINNs in complex fluid dynamics, addressing both forward and inverse problems with innovative approaches like ALE formulation and quasi-conformal mapping. The argumentation is solid, supported by quantitative results comparing with finite element methods and ablation studies. The second talk introduces ZENN, a conceptually rich framework that bridges thermodynamics and machine learning, with demonstrations across diverse domains. The argumentation is well-structured, explaining the theoretical basis and showing empirical evidence. Both talks are technically rigorous and contribute original ideas to their respective fields.

Scientific Rigor, Source Quality, Title Accuracy

The seminar demonstrates high scientific rigor, with both talks presenting mathematically detailed formulations and experimental validations. The sources cited are primarily the speakers’ own research, which is appropriate for a seminar format. The title accurately reflects the content, which is a dual presentation. The adequacy between title and content is excellent. No external sources are explicitly cited in the video description, but the talks reference prior work such as Phoenix and SimVascular, which are well-known in the field. The absence of external citations is typical for seminar presentations and does not detract from the overall quality.

The title accurately reflects the content, which consists of two distinct talks on PINNs for blood flow and ZENN for heterogeneous data-driven modeling.

8/10

The seminar presents two research talks with detailed technical content, including mathematical formulations and experimental results. The speakers are from reputable institutions (City University of Hong Kong, Pennsylvania State University). The content is peer-reviewed research presented in a seminar format, indicating high reliability. However, the video has low viewership and no external verification, so a slight deduction is applied.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:00Introduction to the seminar and first speaker Han Zhang.
• 02:00Motivation for blood flow simulation: FFR and QFR.
• 10:00Introduction to PINNs and ALE formulation.
• 30:00Forward problem: PINN for blood flow in elastic vessels.
• 1:00:00Inverse problem: reconstruction of blood flow from noisy data.
• 1:30:00Introduction to ZENN and zentropy theory.
• 2:00:00ZENN applications: image classification, materials science, geospatial.
• 2:20:00Q&A and discussion.

• Phoenix — Mentioned as a related work for solving Navier-Stokes equations with finite element methods.
• SimVascular — Mentioned as an open-source software for blood flow simulation.

• Physics-Informed Neural Networks — General reference for PINNs.