Educational Workshop

 

(Pre-Conference Workshop)
Workshop #1: 15:15 – 18:00, Sunday 11 October, 2026

Mark Iadicola: (Bio)

MEASUREMENT UNCERTAINTY FOR DIGITAL IMAGE CORRELATION

A long-standing issue with the use of DIC measurements is the need to understand and quantify the uncertainty of measurements. The International DIC Society Standardization Committee has a working group developing a measurement uncertainty guide for DIC. The guide is a collection of existing methods to quantify the uncertainty of typical DIC measurements. In this course, we will discuss the purpose of measurement uncertainty and walk through some of the methods of measurement uncertainty quantification for DIC. We will also consider which sources of uncertainty each method does and does not assess, through descriptions and examples.  

 

(Pre-Conference Workshop)
Workshop #2: 15:15 – 18:00, Sunday 11 October, 2026

Seungchul Lee: (Bio)

PHYSICS-INFORMED MACHINE LEARNING FOR PLASTICITY AND MATERIALS PROCESSING

With recent advances in artificial intelligence, there is growing interest in Physics-informed Machine Learning, which combines data with physical knowledge across many domains, including plasticity and materials processing. Purely data-driven AI models often require large datasets and may produce physically inconsistent predictions. Incorporating physical knowledge into AI model training addresses both issues. This course covers three connected topics. First, how to discover mathematical relationships between variables from experimental data using SINDy and symbolic regression methods such as SISSO. Second, how to embed known physical properties and equations into AI model training through hard and soft constraints, so that predictions satisfy conditions such as periodicity and positivity. Third, how to incorporate governing equations into AI model training by evaluating physics-based residuals through automatic differentiation or finite difference methods, with PINNs as a representative framework. PI-DeepONet is then introduced as a physics-informed operator learning framework for fast surrogate modeling of simulations. Finally, the course introduces agentic AI, demonstrating how the entire workflow, from knowledge discovery to physics-informed model training, can be carried out autonomously. Each topic is accompanied by a demonstration session using Colab to introduce the practical implementation of the methods covered in the course.

Part 1: Data-Driven Knowledge Discovery

  • SINDy, symbolic regression: discovering mathematical relationships between variables from data
  • with Colab: SISSO, a representative symbolic regression method, implementation to recover a known physical relationship from experimental data

Part 2: Embedding Prior Knowledge into AI Training

  • Hard and soft constraints: embedding known physical properties and equations into AI model training
  • with Colab: embedding known physical properties of simulation data, such as periodicity and positivity, into AI model training to improve prediction accuracy

Part 3: Physics-Informed AI

  • PINN: embedding governing equations into AI model training using automatic differentiation or finite difference method
  • PI-DeepONet: physics-informed operator learning for fast surrogate modeling of simulations
  • with Colab: building a PINN as a PDE solver to predict plastic deformation of a circular cylinder

Part 4: Agentic AI for Knowledge Discovery to Physics-Informed AI

  • Agentic AI: LLM-based agents for planning and executing workflows discussed in previous topics
  • with Colab: Agentic AI autonomously performing the entire workflow, from knowledge discovery to physics-informed AI model training

 

(After-Conference Workshop)
Workshop #3: 09:15 – 12:00, Friday 16 October, 2026

Shi-Hoon Choi: (Bio)

TEXTURE AND CRYSTAL PLASTICITY IN METAL FORMING: FROM MICROSTRUCTURE CHARACTERIZATION TO DEFORMATION PREDICTION

Part I — Fundamentals of Crystallographic Texture in Metal Forming
Part II — Texture Characterization and Deformation Mechanisms
Part III — Crystal Plasticity Modeling and Applications to Metal Forming

 

(After-Conference Workshop)
Workshop #4: 09:15 – 12:00, Friday 16 October, 2026

Eun-Ho Lee: (Bio)

THERMO-MECHANICAL BEHAVIOR AND THE EFFECTS OF PLASTIC DEFORMATION IN SEMICONDUCTOR/ADVANCED PACKAGING MANUFACTURING AND RELIABILITY

Artificial Intelligence (AI) technology is evolving beyond the digital realm into the physical world as Physical AI, expanding its application into autonomous driving, defense, and manufacturing. Implementing these advanced AI systems increasingly relies on modern semiconductor and packaging technologies.
In particular, because AI workloads demand massive data transfer and processing capacity, the complexity of inter-chip connectivity has escalated dramatically, giving rise to thermal challenges that degrade overall performance. In this context, the thermo-mechanical behavior of semiconductors and packaging plays a crucial role in determining the performance of AI infrastructure. Moreover, the plastic deformation of copper during inter-chip interconnects is a key factor affecting manufacturing quality.
Based on the fundamental theories of thermo-mechanical and plastic deformation, this lecture addresses practical issues in their application to heterogeneous integration in advanced semiconductor packaging. It also covers semiconductor performance degradation caused by thermo-mechanical deformation and discusses potential solutions. Through this session, attendees will understand the impact of plastic deformation on advanced semiconductor packaging design and manufacturing, as well as its contribution to realizing robust AI infrastructure.